During prenatal visits, a woman, pregnant with her fourth child, discussed undergoing labor and delivery in any position other than on her back; the ObGyn agreed. When she arrived at the hospital in labor, the patient told the nurse that she preferred to labor on her hands and knees. The nurse disagreed because of the fetal heart-rate monitor.
When the patient began hard labor, she turned herself over onto her hands and knees and again informed the nurse that she could not labor on her back. The nurse flipped the patient onto her back by taking her wrists and pulling the patient’s hands out from under her. The nurse then delayed delivery until the ObGyn arrived by putting pressure on the baby’s head. During delivery, a second nurse forcibly pressed the patient’s left knee back toward her chest, leaving her legs in an asymmetric position.
Two months later, the patient reported chronic severe pelvic pain and was found to have pudendal neuralgia. She underwent nerve blocks and takes medication for chronic pain.
PATIENT’S CLAIM:
The ObGyn did not assume responsibility when he arrived for the delivery. The nurses did not follow the standard of care. The patient’s injury was the result of tension and compression due to malpositioning of the patient’s legs during delivery.
DEFENDANTS’ DEFENSE:
There was no breach in the standard of care. The patient’s injury, if any, had not been caused by the delivery.
Although genetic testing was scheduled for a 37-year-old woman’s 15-week prenatal visit, the ObGyn’s staff failed to draw blood. At 19 weeks’ gestation (April 24), blood was drawn. The ObGyn signed off on test results that showed a high risk for fetal anomaly on May 2, but the patient was not informed until May 22. The ObGyn scheduled amniocentesis for June 3. On May 30, the hospital, based in Illinois, cancelled the test, telling the ObGyn that it was because the patient was over 24 weeks’ pregnant and there was no labor and delivery unit to respond if complications arose. Instead of notifying the patient, the ObGyn arranged for amniocentesis to be performed elsewhere on June 3. The ObGyn saw the amniocentesis results on June 13, but did not tell the patient until July 3, when he advised her to terminate the pregnancy because the baby had severe cardiac defects and Down syndrome; he felt the child would not survive or have very poor quality of life. The ObGyn arranged for the patient to undergo a third-trimester abortion in Kansas and paid all expenses. On July 14, the patient began the 5-day abortion process at 30+ weeks’ gestation.
PATIENT’S CLAIM:
She was never offered additional genetic testing or expedited amniocentesis. She was not told that abortion is illegal in Illinois after 23 6/7 weeks’ gestation. The ObGyn had a motive for paying for her abortion. He never counseled her about options to keep the child. She endured extreme pain and emotional trauma during the abortion and was later found to have posttraumatic stress disorder, multidepressive disorder, and anxiety as a result of the experience. She countered the ObGyn’s contact information claim by saying that her phone number had not changed.
PHYSICIAN’S DEFENSE:
The ObGyn admitted negligence in failing to timely communicate test results but contended that the patient was more than 50% responsible for any delay by failing to update her contact information when she moved. The ObGyn denied causation of any injuries or damage.
A woman presented to the hospital in labor. During delivery, the patient’s ObGyn encountered shoulder dystocia. The infant died shortly after birth.
PARENTS’ CLAIM:
The ObGyn and hospital nurses were negligent. The nurses failed to monitor labor and properly communicate with the ObGyn. The ObGyn failed to appreciate the baby’s large size and order a cesarean delivery. The infant’s death was due to a hypoxic event during delivery.
DEFENDANTS’ DEFENSE:
The baby gained an unexpected amount of weight between the last prenatal visit and labor. There was no reason to expect a complication to vaginal delivery. The nurses denied negligence. The child’s sudden death was caused by a genetic cardiac condition.
These cases were selected by the editors of OBG Management from Medical Malpractice Verdicts, Settlements & Experts, with permission of the editor, Lewis Laska (www.verdictslaska.com). The information available to the editors about the cases presented here is sometimes incomplete. Moreover, the cases may or may not have merit. Nevertheless, these cases represent the types of clinical situations that typically result in litigation and are meant to illustrate nationwide variation in jury verdicts and awards.
Share your thoughts! Send your Letter to the Editor to rbarbieri@frontlinemedcom.com. Please include your name and the city and state in which you practice.
During prenatal visits, a woman, pregnant with her fourth child, discussed undergoing labor and delivery in any position other than on her back; the ObGyn agreed. When she arrived at the hospital in labor, the patient told the nurse that she preferred to labor on her hands and knees. The nurse disagreed because of the fetal heart-rate monitor.
When the patient began hard labor, she turned herself over onto her hands and knees and again informed the nurse that she could not labor on her back. The nurse flipped the patient onto her back by taking her wrists and pulling the patient’s hands out from under her. The nurse then delayed delivery until the ObGyn arrived by putting pressure on the baby’s head. During delivery, a second nurse forcibly pressed the patient’s left knee back toward her chest, leaving her legs in an asymmetric position.
Two months later, the patient reported chronic severe pelvic pain and was found to have pudendal neuralgia. She underwent nerve blocks and takes medication for chronic pain.
PATIENT’S CLAIM:
The ObGyn did not assume responsibility when he arrived for the delivery. The nurses did not follow the standard of care. The patient’s injury was the result of tension and compression due to malpositioning of the patient’s legs during delivery.
DEFENDANTS’ DEFENSE:
There was no breach in the standard of care. The patient’s injury, if any, had not been caused by the delivery.
Although genetic testing was scheduled for a 37-year-old woman’s 15-week prenatal visit, the ObGyn’s staff failed to draw blood. At 19 weeks’ gestation (April 24), blood was drawn. The ObGyn signed off on test results that showed a high risk for fetal anomaly on May 2, but the patient was not informed until May 22. The ObGyn scheduled amniocentesis for June 3. On May 30, the hospital, based in Illinois, cancelled the test, telling the ObGyn that it was because the patient was over 24 weeks’ pregnant and there was no labor and delivery unit to respond if complications arose. Instead of notifying the patient, the ObGyn arranged for amniocentesis to be performed elsewhere on June 3. The ObGyn saw the amniocentesis results on June 13, but did not tell the patient until July 3, when he advised her to terminate the pregnancy because the baby had severe cardiac defects and Down syndrome; he felt the child would not survive or have very poor quality of life. The ObGyn arranged for the patient to undergo a third-trimester abortion in Kansas and paid all expenses. On July 14, the patient began the 5-day abortion process at 30+ weeks’ gestation.
PATIENT’S CLAIM:
She was never offered additional genetic testing or expedited amniocentesis. She was not told that abortion is illegal in Illinois after 23 6/7 weeks’ gestation. The ObGyn had a motive for paying for her abortion. He never counseled her about options to keep the child. She endured extreme pain and emotional trauma during the abortion and was later found to have posttraumatic stress disorder, multidepressive disorder, and anxiety as a result of the experience. She countered the ObGyn’s contact information claim by saying that her phone number had not changed.
PHYSICIAN’S DEFENSE:
The ObGyn admitted negligence in failing to timely communicate test results but contended that the patient was more than 50% responsible for any delay by failing to update her contact information when she moved. The ObGyn denied causation of any injuries or damage.
A woman presented to the hospital in labor. During delivery, the patient’s ObGyn encountered shoulder dystocia. The infant died shortly after birth.
PARENTS’ CLAIM:
The ObGyn and hospital nurses were negligent. The nurses failed to monitor labor and properly communicate with the ObGyn. The ObGyn failed to appreciate the baby’s large size and order a cesarean delivery. The infant’s death was due to a hypoxic event during delivery.
DEFENDANTS’ DEFENSE:
The baby gained an unexpected amount of weight between the last prenatal visit and labor. There was no reason to expect a complication to vaginal delivery. The nurses denied negligence. The child’s sudden death was caused by a genetic cardiac condition.
These cases were selected by the editors of OBG Management from Medical Malpractice Verdicts, Settlements & Experts, with permission of the editor, Lewis Laska (www.verdictslaska.com). The information available to the editors about the cases presented here is sometimes incomplete. Moreover, the cases may or may not have merit. Nevertheless, these cases represent the types of clinical situations that typically result in litigation and are meant to illustrate nationwide variation in jury verdicts and awards.
Share your thoughts! Send your Letter to the Editor to rbarbieri@frontlinemedcom.com. Please include your name and the city and state in which you practice.
She wanted to labor on hands and knees
During prenatal visits, a woman, pregnant with her fourth child, discussed undergoing labor and delivery in any position other than on her back; the ObGyn agreed. When she arrived at the hospital in labor, the patient told the nurse that she preferred to labor on her hands and knees. The nurse disagreed because of the fetal heart-rate monitor.
When the patient began hard labor, she turned herself over onto her hands and knees and again informed the nurse that she could not labor on her back. The nurse flipped the patient onto her back by taking her wrists and pulling the patient’s hands out from under her. The nurse then delayed delivery until the ObGyn arrived by putting pressure on the baby’s head. During delivery, a second nurse forcibly pressed the patient’s left knee back toward her chest, leaving her legs in an asymmetric position.
Two months later, the patient reported chronic severe pelvic pain and was found to have pudendal neuralgia. She underwent nerve blocks and takes medication for chronic pain.
PATIENT’S CLAIM:
The ObGyn did not assume responsibility when he arrived for the delivery. The nurses did not follow the standard of care. The patient’s injury was the result of tension and compression due to malpositioning of the patient’s legs during delivery.
DEFENDANTS’ DEFENSE:
There was no breach in the standard of care. The patient’s injury, if any, had not been caused by the delivery.
Although genetic testing was scheduled for a 37-year-old woman’s 15-week prenatal visit, the ObGyn’s staff failed to draw blood. At 19 weeks’ gestation (April 24), blood was drawn. The ObGyn signed off on test results that showed a high risk for fetal anomaly on May 2, but the patient was not informed until May 22. The ObGyn scheduled amniocentesis for June 3. On May 30, the hospital, based in Illinois, cancelled the test, telling the ObGyn that it was because the patient was over 24 weeks’ pregnant and there was no labor and delivery unit to respond if complications arose. Instead of notifying the patient, the ObGyn arranged for amniocentesis to be performed elsewhere on June 3. The ObGyn saw the amniocentesis results on June 13, but did not tell the patient until July 3, when he advised her to terminate the pregnancy because the baby had severe cardiac defects and Down syndrome; he felt the child would not survive or have very poor quality of life. The ObGyn arranged for the patient to undergo a third-trimester abortion in Kansas and paid all expenses. On July 14, the patient began the 5-day abortion process at 30+ weeks’ gestation.
PATIENT’S CLAIM:
She was never offered additional genetic testing or expedited amniocentesis. She was not told that abortion is illegal in Illinois after 23 6/7 weeks’ gestation. The ObGyn had a motive for paying for her abortion. He never counseled her about options to keep the child. She endured extreme pain and emotional trauma during the abortion and was later found to have posttraumatic stress disorder, multidepressive disorder, and anxiety as a result of the experience. She countered the ObGyn’s contact information claim by saying that her phone number had not changed.
PHYSICIAN’S DEFENSE:
The ObGyn admitted negligence in failing to timely communicate test results but contended that the patient was more than 50% responsible for any delay by failing to update her contact information when she moved. The ObGyn denied causation of any injuries or damage.
A woman presented to the hospital in labor. During delivery, the patient’s ObGyn encountered shoulder dystocia. The infant died shortly after birth.
PARENTS’ CLAIM:
The ObGyn and hospital nurses were negligent. The nurses failed to monitor labor and properly communicate with the ObGyn. The ObGyn failed to appreciate the baby’s large size and order a cesarean delivery. The infant’s death was due to a hypoxic event during delivery.
DEFENDANTS’ DEFENSE:
The baby gained an unexpected amount of weight between the last prenatal visit and labor. There was no reason to expect a complication to vaginal delivery. The nurses denied negligence. The child’s sudden death was caused by a genetic cardiac condition.
These cases were selected by the editors of OBG Management from Medical Malpractice Verdicts, Settlements & Experts, with permission of the editor, Lewis Laska (www.verdictslaska.com). The information available to the editors about the cases presented here is sometimes incomplete. Moreover, the cases may or may not have merit. Nevertheless, these cases represent the types of clinical situations that typically result in litigation and are meant to illustrate nationwide variation in jury verdicts and awards.
Share your thoughts! Send your Letter to the Editor to rbarbieri@frontlinemedcom.com. Please include your name and the city and state in which you practice.
A randomized controlled multicenter study published by Carson et al. in JAMA concluded that, for patients with “chronic critical illness” (defined as requiring 7 days of mechanical ventilation), palliative care team-led informational and emotional support meetings did not reduce anxiety or depression for families and may have increased posttraumatic stress disorder symptoms (2016:316[1]:51-62. doi: 10.1001/jama.2016.8474).
This report may surprise surgeons, as well as practitioners in other specialties, as the disconnect between palliative care and critical care services has been previously perceived as an education and access issue, not an outcome problem.
Dr. Emily RivetWhere can we look in the fields of surgery and palliative care to explain and improve these outcomes? We could start with our openness to cross-pollination of these fields. Just as the field of surgery is evolving through the assimilation of palliative care principles, the field of palliative care may also evolve through the perspectives of surgery, including the uniqueness of the surgeon. When describing techniques and outcomes, surgeons often employ the phrase, “in my hands,” to rationalize variable outcomes stemming from subtle differences in surgical technique, population, relationships, institutional culture, and processes which defy easy quantification. Although the field of surgery is shifting from a cult of personality to protocol-based approaches in its undertakings, there is still a place for “surgeon preference” for equipment and other elements of surgical care. Palliative care is comparably dependent on individual approaches, relationships, and culture.
Carson and colleagues point out that fidelity to some components of the meeting “templates” was low, suggesting that there was some flexibility baked into the study design. However, as Russ and Kaufman aptly described, patients and families vary greatly in their appetite for explicit information about prognosis (Cult Med Psychiatry 2005;29[1]:103-23). Conversely, the hypothesis that direct communication about prognosis will be welcomed by families is a core element of the Carson study. The manuscript supplement reports that discussion of the patient’s condition and prognosis took place in 100% of initial meetings. If the same variability in family receptiveness to this information exists in this population as was described by Russ and Kaufman, it is not hard to see why some families experienced negative consequences because of these discussions.
Furthermore, the authors of the Carson study point out that it was not intended to replicate the components of specialist palliative care (JAMA. 2016;316[15]:1598-9).
Essential elements of specialist palliative care include symptom management, a multidisciplinary approach, and fairly close contact in the acute care setting. These features were lacking in the study protocol. Experienced providers of palliative care will often use symptom assessment and symptom management optimization as a conduit for building rapport and to avoid focusing on prognosis until trust has been established. A period of delay before broaching challenging subjects also allows the palliative care team to develop an understanding of the patient’s or surrogates’ preferences regarding the amount and type of information communicated. Palliative care providers benefit from the deepening of relationships with patients and families over time, as much as or possibly more so than providers of other specialties.
The necessity of the multidisciplinary approach to successful palliative care outcomes cannot be overstated. In many programs, patients seen for specialist palliative care consultation are seen by a physician or advanced practitioner, a chaplain, and a social worker within 24-48 hours of initial referral, and these providers have key roles in addressing the sequelae of anxiety, depression, and stress that were the key outcomes in the JAMA study. In the study, the “support and information team” included a palliative care physician and an advanced practice nurse but not a chaplain or social worker, despite the significance of existential/spiritual and social consequences of ventilator withdrawal or progression to tracheostomy for long-term vent support.
Palliative care providers consider the family meeting to be the “procedure” of their field, a belief that may seem incongruous with a surgical understanding of the nature of procedures but is informative as a framework for understanding the results of the Carson study. Just as surgical procedures carry risk of complications or adverse outcomes, family meetings have risk for worsening instead of improving the coping of families and surrogates. And, as surgical technique can be connected to complications, the family meeting technique applied by Carson et al. may be related to its results. Although there was formalized communication between the ICU team and the palliative care team regarding the patient’s condition, prognosis, and treatment plan, there was not a representative from the critical care team present during the majority of the support and information team led family meetings. This represents a marked deviation from common practice at our institution and many others. Our usual practice is to have a member of the ICU team present for discussions focused on patient prognosis, in order to make sure that there is alignment between the messages of the ICU and palliative care teams and also to prevent the crippling of palliative support that occurs when it becomes the sole repository of unwelcome news.
Because the relief of suffering is a core value of surgery and palliative care, there are countless ways these disciplines can inform one another. The outcome of the Carson study is a cautionary tale about the fallibility of the integration of surgical and palliative care teams, both of which would acknowledge the importance of the multidisciplinary approach, relationships developed over time, and symptom management. As surgeons intuitively understand from their operative experience, the “procedure” (the family meeting) has the potential for both risk and benefit, the outcome of which may be determined “in my hands.”
Dr. Rivet is a colon and rectal surgeon with training and board certification in hospice and palliative medicine. She is an assistant professor, departments of surgery and internal medicine, Virginia Commonwealth University, Richmond. She has no disclosures.
A randomized controlled multicenter study published by Carson et al. in JAMA concluded that, for patients with “chronic critical illness” (defined as requiring 7 days of mechanical ventilation), palliative care team-led informational and emotional support meetings did not reduce anxiety or depression for families and may have increased posttraumatic stress disorder symptoms (2016:316[1]:51-62. doi: 10.1001/jama.2016.8474).
This report may surprise surgeons, as well as practitioners in other specialties, as the disconnect between palliative care and critical care services has been previously perceived as an education and access issue, not an outcome problem.
Dr. Emily RivetWhere can we look in the fields of surgery and palliative care to explain and improve these outcomes? We could start with our openness to cross-pollination of these fields. Just as the field of surgery is evolving through the assimilation of palliative care principles, the field of palliative care may also evolve through the perspectives of surgery, including the uniqueness of the surgeon. When describing techniques and outcomes, surgeons often employ the phrase, “in my hands,” to rationalize variable outcomes stemming from subtle differences in surgical technique, population, relationships, institutional culture, and processes which defy easy quantification. Although the field of surgery is shifting from a cult of personality to protocol-based approaches in its undertakings, there is still a place for “surgeon preference” for equipment and other elements of surgical care. Palliative care is comparably dependent on individual approaches, relationships, and culture.
Carson and colleagues point out that fidelity to some components of the meeting “templates” was low, suggesting that there was some flexibility baked into the study design. However, as Russ and Kaufman aptly described, patients and families vary greatly in their appetite for explicit information about prognosis (Cult Med Psychiatry 2005;29[1]:103-23). Conversely, the hypothesis that direct communication about prognosis will be welcomed by families is a core element of the Carson study. The manuscript supplement reports that discussion of the patient’s condition and prognosis took place in 100% of initial meetings. If the same variability in family receptiveness to this information exists in this population as was described by Russ and Kaufman, it is not hard to see why some families experienced negative consequences because of these discussions.
Furthermore, the authors of the Carson study point out that it was not intended to replicate the components of specialist palliative care (JAMA. 2016;316[15]:1598-9).
Essential elements of specialist palliative care include symptom management, a multidisciplinary approach, and fairly close contact in the acute care setting. These features were lacking in the study protocol. Experienced providers of palliative care will often use symptom assessment and symptom management optimization as a conduit for building rapport and to avoid focusing on prognosis until trust has been established. A period of delay before broaching challenging subjects also allows the palliative care team to develop an understanding of the patient’s or surrogates’ preferences regarding the amount and type of information communicated. Palliative care providers benefit from the deepening of relationships with patients and families over time, as much as or possibly more so than providers of other specialties.
The necessity of the multidisciplinary approach to successful palliative care outcomes cannot be overstated. In many programs, patients seen for specialist palliative care consultation are seen by a physician or advanced practitioner, a chaplain, and a social worker within 24-48 hours of initial referral, and these providers have key roles in addressing the sequelae of anxiety, depression, and stress that were the key outcomes in the JAMA study. In the study, the “support and information team” included a palliative care physician and an advanced practice nurse but not a chaplain or social worker, despite the significance of existential/spiritual and social consequences of ventilator withdrawal or progression to tracheostomy for long-term vent support.
Palliative care providers consider the family meeting to be the “procedure” of their field, a belief that may seem incongruous with a surgical understanding of the nature of procedures but is informative as a framework for understanding the results of the Carson study. Just as surgical procedures carry risk of complications or adverse outcomes, family meetings have risk for worsening instead of improving the coping of families and surrogates. And, as surgical technique can be connected to complications, the family meeting technique applied by Carson et al. may be related to its results. Although there was formalized communication between the ICU team and the palliative care team regarding the patient’s condition, prognosis, and treatment plan, there was not a representative from the critical care team present during the majority of the support and information team led family meetings. This represents a marked deviation from common practice at our institution and many others. Our usual practice is to have a member of the ICU team present for discussions focused on patient prognosis, in order to make sure that there is alignment between the messages of the ICU and palliative care teams and also to prevent the crippling of palliative support that occurs when it becomes the sole repository of unwelcome news.
Because the relief of suffering is a core value of surgery and palliative care, there are countless ways these disciplines can inform one another. The outcome of the Carson study is a cautionary tale about the fallibility of the integration of surgical and palliative care teams, both of which would acknowledge the importance of the multidisciplinary approach, relationships developed over time, and symptom management. As surgeons intuitively understand from their operative experience, the “procedure” (the family meeting) has the potential for both risk and benefit, the outcome of which may be determined “in my hands.”
Dr. Rivet is a colon and rectal surgeon with training and board certification in hospice and palliative medicine. She is an assistant professor, departments of surgery and internal medicine, Virginia Commonwealth University, Richmond. She has no disclosures.
A randomized controlled multicenter study published by Carson et al. in JAMA concluded that, for patients with “chronic critical illness” (defined as requiring 7 days of mechanical ventilation), palliative care team-led informational and emotional support meetings did not reduce anxiety or depression for families and may have increased posttraumatic stress disorder symptoms (2016:316[1]:51-62. doi: 10.1001/jama.2016.8474).
This report may surprise surgeons, as well as practitioners in other specialties, as the disconnect between palliative care and critical care services has been previously perceived as an education and access issue, not an outcome problem.
Dr. Emily RivetWhere can we look in the fields of surgery and palliative care to explain and improve these outcomes? We could start with our openness to cross-pollination of these fields. Just as the field of surgery is evolving through the assimilation of palliative care principles, the field of palliative care may also evolve through the perspectives of surgery, including the uniqueness of the surgeon. When describing techniques and outcomes, surgeons often employ the phrase, “in my hands,” to rationalize variable outcomes stemming from subtle differences in surgical technique, population, relationships, institutional culture, and processes which defy easy quantification. Although the field of surgery is shifting from a cult of personality to protocol-based approaches in its undertakings, there is still a place for “surgeon preference” for equipment and other elements of surgical care. Palliative care is comparably dependent on individual approaches, relationships, and culture.
Carson and colleagues point out that fidelity to some components of the meeting “templates” was low, suggesting that there was some flexibility baked into the study design. However, as Russ and Kaufman aptly described, patients and families vary greatly in their appetite for explicit information about prognosis (Cult Med Psychiatry 2005;29[1]:103-23). Conversely, the hypothesis that direct communication about prognosis will be welcomed by families is a core element of the Carson study. The manuscript supplement reports that discussion of the patient’s condition and prognosis took place in 100% of initial meetings. If the same variability in family receptiveness to this information exists in this population as was described by Russ and Kaufman, it is not hard to see why some families experienced negative consequences because of these discussions.
Furthermore, the authors of the Carson study point out that it was not intended to replicate the components of specialist palliative care (JAMA. 2016;316[15]:1598-9).
Essential elements of specialist palliative care include symptom management, a multidisciplinary approach, and fairly close contact in the acute care setting. These features were lacking in the study protocol. Experienced providers of palliative care will often use symptom assessment and symptom management optimization as a conduit for building rapport and to avoid focusing on prognosis until trust has been established. A period of delay before broaching challenging subjects also allows the palliative care team to develop an understanding of the patient’s or surrogates’ preferences regarding the amount and type of information communicated. Palliative care providers benefit from the deepening of relationships with patients and families over time, as much as or possibly more so than providers of other specialties.
The necessity of the multidisciplinary approach to successful palliative care outcomes cannot be overstated. In many programs, patients seen for specialist palliative care consultation are seen by a physician or advanced practitioner, a chaplain, and a social worker within 24-48 hours of initial referral, and these providers have key roles in addressing the sequelae of anxiety, depression, and stress that were the key outcomes in the JAMA study. In the study, the “support and information team” included a palliative care physician and an advanced practice nurse but not a chaplain or social worker, despite the significance of existential/spiritual and social consequences of ventilator withdrawal or progression to tracheostomy for long-term vent support.
Palliative care providers consider the family meeting to be the “procedure” of their field, a belief that may seem incongruous with a surgical understanding of the nature of procedures but is informative as a framework for understanding the results of the Carson study. Just as surgical procedures carry risk of complications or adverse outcomes, family meetings have risk for worsening instead of improving the coping of families and surrogates. And, as surgical technique can be connected to complications, the family meeting technique applied by Carson et al. may be related to its results. Although there was formalized communication between the ICU team and the palliative care team regarding the patient’s condition, prognosis, and treatment plan, there was not a representative from the critical care team present during the majority of the support and information team led family meetings. This represents a marked deviation from common practice at our institution and many others. Our usual practice is to have a member of the ICU team present for discussions focused on patient prognosis, in order to make sure that there is alignment between the messages of the ICU and palliative care teams and also to prevent the crippling of palliative support that occurs when it becomes the sole repository of unwelcome news.
Because the relief of suffering is a core value of surgery and palliative care, there are countless ways these disciplines can inform one another. The outcome of the Carson study is a cautionary tale about the fallibility of the integration of surgical and palliative care teams, both of which would acknowledge the importance of the multidisciplinary approach, relationships developed over time, and symptom management. As surgeons intuitively understand from their operative experience, the “procedure” (the family meeting) has the potential for both risk and benefit, the outcome of which may be determined “in my hands.”
Dr. Rivet is a colon and rectal surgeon with training and board certification in hospice and palliative medicine. She is an assistant professor, departments of surgery and internal medicine, Virginia Commonwealth University, Richmond. She has no disclosures.
The Sureglide Cesarean Scalpel™ is designed to reduce the risk of fetal injury from nicks, cuts, or lacerations during hysterotomy and protect physicians, nurses, and other surgical staff from sharp injury. Sureglide manufacturer Ecomed Solutions points out that, as the number of cesarean deliveries increases so does the risk of fetal injury as often the face, cheek, and ear of the fetus are in line with the hysterotomy incision. Unlike typical scalpels, Sureglide’s design eliminates fetal exposure to the blade regardless of uterine wall thickness or number of passes, says Ecomed. The scalpel’s protected blade is designed to cut up and away from the fetus and eliminate exposure to the blade by surgical staff.
Viveve Medicalsays its GENEVEVE™ treatment offers improvement in vaginal laxity and sexual function using cryogen-cooled monopolar radiofrequency (CMRF) to uniformly deliver volumetric heating (90 J/cm2) while cooling delicate surface tissue. CMRF stimulates the body’s natural collagen formation process by penetrating 3–5 mm deep into connective tissue. The Geneveve single-session treatment is performed in an outpatient setting in 30 minutes.
According toViveve Medical, results of the recent Viveve I multicenter, blinded, randomized, sham controlled study showed no serious adverse effects plus improvement in arousal and/or orgasm self-reported by 9 of 10 women who noted vaginal laxity and sexual dysfunction following vaginal childbirth.
Hologic’s new NovaSure® ADVANCED global endometrial ablation system with a 6-mm sheath size requires less cervical dilation than the 8-mm NovaSure device. Hologic says the new device is designed to improve patient comfort and physician ease-of-use while maintaining clinical efficacy. NovaSure ADVANCED’s acorn-shaped cervical seal provides 13% more working length than the 8-mm NovaSure device. Smooth Access™ tips and blue handle simplify insertion. NovaSure endometrial ablation, a one-time procedure, can be performed in the office or operating room to reduce or stop abnormal uterine bleeding.
The Sureglide Cesarean Scalpel™ is designed to reduce the risk of fetal injury from nicks, cuts, or lacerations during hysterotomy and protect physicians, nurses, and other surgical staff from sharp injury. Sureglide manufacturer Ecomed Solutions points out that, as the number of cesarean deliveries increases so does the risk of fetal injury as often the face, cheek, and ear of the fetus are in line with the hysterotomy incision. Unlike typical scalpels, Sureglide’s design eliminates fetal exposure to the blade regardless of uterine wall thickness or number of passes, says Ecomed. The scalpel’s protected blade is designed to cut up and away from the fetus and eliminate exposure to the blade by surgical staff.
Viveve Medicalsays its GENEVEVE™ treatment offers improvement in vaginal laxity and sexual function using cryogen-cooled monopolar radiofrequency (CMRF) to uniformly deliver volumetric heating (90 J/cm2) while cooling delicate surface tissue. CMRF stimulates the body’s natural collagen formation process by penetrating 3–5 mm deep into connective tissue. The Geneveve single-session treatment is performed in an outpatient setting in 30 minutes.
According toViveve Medical, results of the recent Viveve I multicenter, blinded, randomized, sham controlled study showed no serious adverse effects plus improvement in arousal and/or orgasm self-reported by 9 of 10 women who noted vaginal laxity and sexual dysfunction following vaginal childbirth.
Hologic’s new NovaSure® ADVANCED global endometrial ablation system with a 6-mm sheath size requires less cervical dilation than the 8-mm NovaSure device. Hologic says the new device is designed to improve patient comfort and physician ease-of-use while maintaining clinical efficacy. NovaSure ADVANCED’s acorn-shaped cervical seal provides 13% more working length than the 8-mm NovaSure device. Smooth Access™ tips and blue handle simplify insertion. NovaSure endometrial ablation, a one-time procedure, can be performed in the office or operating room to reduce or stop abnormal uterine bleeding.
The Sureglide Cesarean Scalpel™ is designed to reduce the risk of fetal injury from nicks, cuts, or lacerations during hysterotomy and protect physicians, nurses, and other surgical staff from sharp injury. Sureglide manufacturer Ecomed Solutions points out that, as the number of cesarean deliveries increases so does the risk of fetal injury as often the face, cheek, and ear of the fetus are in line with the hysterotomy incision. Unlike typical scalpels, Sureglide’s design eliminates fetal exposure to the blade regardless of uterine wall thickness or number of passes, says Ecomed. The scalpel’s protected blade is designed to cut up and away from the fetus and eliminate exposure to the blade by surgical staff.
Viveve Medicalsays its GENEVEVE™ treatment offers improvement in vaginal laxity and sexual function using cryogen-cooled monopolar radiofrequency (CMRF) to uniformly deliver volumetric heating (90 J/cm2) while cooling delicate surface tissue. CMRF stimulates the body’s natural collagen formation process by penetrating 3–5 mm deep into connective tissue. The Geneveve single-session treatment is performed in an outpatient setting in 30 minutes.
According toViveve Medical, results of the recent Viveve I multicenter, blinded, randomized, sham controlled study showed no serious adverse effects plus improvement in arousal and/or orgasm self-reported by 9 of 10 women who noted vaginal laxity and sexual dysfunction following vaginal childbirth.
Hologic’s new NovaSure® ADVANCED global endometrial ablation system with a 6-mm sheath size requires less cervical dilation than the 8-mm NovaSure device. Hologic says the new device is designed to improve patient comfort and physician ease-of-use while maintaining clinical efficacy. NovaSure ADVANCED’s acorn-shaped cervical seal provides 13% more working length than the 8-mm NovaSure device. Smooth Access™ tips and blue handle simplify insertion. NovaSure endometrial ablation, a one-time procedure, can be performed in the office or operating room to reduce or stop abnormal uterine bleeding.
The care of hospitalized patients requires practitioners from multiple disciplines to assess and communicate the patient’s status in a dynamic manner during hospitalization. Although optimal teamwork is needed for patient care to be delivered reliably and efficiently, care within hospitals is typically delivered in a fragmented manner.1 A bedside model for daily interdisciplinary rounds (IDR) has been proposed as a method to provide a structured process and engage all team members in a patient-centered, system-of-care delivery.2 Specific advantages of convening rounds in the presence of the patient include the ability to directly assess care (eg, presence of a potentially unnecessary urinary catheter), patient engagement in key aspects of their care and disposition, and an increased opportunity for team members to develop a shared understanding of the patient’s views and needs.
Implementing dramatic changes to the workflow of multiple disciplines will require rigorous evidence to support a concerted effort from leadership and buy-in from stakeholders at the front line of patient care. Despite the urgency for evidence, there has been little investigation of this strategy. A systematic review3 identified 30 studies published between 1998 and 2013 addressing interdisciplinary interventions on medical wards, none of which examined a bedside IDR model. In a study performed after the period assessed by the systematic review, Stein et al4 described the restructuring of a medical ward as an accountable care unit (ACU), which included a bedside model for rounds by the interdisciplinary team. The change was associated with decreased mortality and length of stay (LOS), although the study did not isolate the impact of rounds or use a concurrent control group and presented aggregate rather than patient-level outcomes. The lack of convincing data may be a reason bedside rounds are not widely employed by hospitals. To provide high-quality evidence, we performed a large, prospective controlled trial comparing a structured bedside model (mobile interdisciplinary care rounds [MICRO]) with standard rounds.
METHODS
This study took place at the Mount Sinai Hospital, which is a 1171-bed tertiary care academic medical center in New York City, New York. A nonteaching unit offered the ability to use a prospective controlled design. Patients were assigned to the north and the south wings of the unit in a quasi-randomized manner, rather than based on diagnosis or acuity. We transformed IDR to a bedside model on the north side of the unit (MICRO group), while the south side of the unit continued using standard conference room-based IDR (control group). The north and south sides of the unit contain 17 and 14 beds, respectively. During the study period, nurses and hospitalists cared for patients on both sides of the study unit, although on any given day were assigned only patients on 1 side of the unit. The unit uses a clinical microsystem model, which has been defined as “a group of clinicians and staff working together with a shared clinical purpose to provide care for a population of patients,” and has a defined set of characteristics associated with high performance.5,6 Our microsystem model has incorporated features as described by Stein’s ACU model,4 including co-leadership by a hospitalist and a nurse manager, geographic assignment of patients to teams, and unit-level data reports. One hospitalist is assigned geographically to each area of the unit in a 2- to 4-week rotation. Coverage of the unit does not include house staff; patients are primarily assigned to hospitalists working with nurse practitioners. Patients were enrolled prospectively during their initial IDR by a research coordinator. Patient-level data and outcomes were collected prospectively by a research coordinator who attended IDR on the intervention and the control sides of the study unit daily.
Inclusion Criteria
All patients admitted to the medicine service on the study unit were eligible. Patients were greater than 18 years and admitted for an acute medical condition. Patients admitted to another unit and later transferred to the study unit were enrolled at the time of transfer. Patients could be included more than once if hospitalized on the study unit on more than 1 occasion. Most patients were covered by hospitalists, although patients covered by private physicians were included. Patients from other departments, including family medicine, are uncommonly admitted to the unit and were excluded. Patients were also excluded if they were admitted and discharged over the same weekend, because the MICRO rounds occur during weekdays and there was no opportunity to offer the intervention on Saturdays and Sundays.
MICRO Intervention
Interdisciplinary rounds occurred daily at 10:00 am for the control group and at 10:30 am for the MICRO group, and were attended by the hospitalist caring for the majority of patients on the unit, staff nurses, and the unit medical director, nurse manager, social worker, and case manager. Rounds on the control unit focused on the plan of care and disposition but did not follow any set structure and were typically 25 to 30 minutes in duration.
The MICRO rounds occurred at the bedside and followed a structured script (Appendix 1) that was designed to limit discussion of each patient to 3 minutes or less, and included speaking roles for the hospitalist, nurse, and social worker. For private physicians, the nurse practitioner assigned to the patient performed the role of the hospitalist. Rounds were expected to be approximately 50 minutes in duration. Patients were further engaged by asking for their main goal for the day. A patient safety checklist was reviewed. Initially, this task was performed by the nurse manager, who did not verbalize the items unless a deficiency was noted. After 6 months’ experience, this responsibility was given to the staff nurse, who reviewed the checklist verbally as part of the bedside script. Patients were seen daily, including those being discharged later that same day.
Staff and Clinician Education
We developed and implemented a curriculum based on a modified version of the Agency for Healthcare Research and Quality’s TeamStepps® program to ensure that all team members were provided with the basic principles of communication within the healthcare setting. The curriculum consisted of interactive didactics on essential elements of teamwork, including team structure, communication, situation monitoring, and mutual support, as well as the purpose and structure of the MICRO model. The curriculum was delivered to nurses at 3 monthly staff meetings on the study unit and to hospitalists during 3 hospital medicine grand rounds over a 3-month period. Nurses and physicians providing care on both geographic areas of the study unit received the education program because no group of practitioners was designated to only 1 geographic area.
Outcomes
Primary and Secondary Outcomes
The primary outcomes were clinical deterioration (CD) and length of stay. Clinical deterioration was a composite outcome defined a priori as death; escalation of care (ie, transfer to an intensive care unit, intermediate care unit, or teaching unit); or a hospital-acquired complication (ie, venous thromboembolism, fall, stage III-IV pressure ulcer, catheter-associated urinary tract infection, central-line associated bloodstream infection, or Clostridium difficile-associated diarrhea). The LOS was calculated as the mean LOS with outliers excluded (outliers defined as having a LOS 100 days or longer or 2.5 or more standard deviations from the expected LOS).
Process metrics on IDR, such as the duration of rounds, attendance by members of the interdisciplinary team, the percentage of patients discussed, or the effectiveness of communication, were not collected. We assessed patient satisfaction based on the Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS) survey.
Patient Safety Culture Survey
To assess the impact on the perceptions of patient safety, we administered the Agency for Healthcare Research and Quality (AHRQ) Hospital Survey on Patient Safety Culture toall staff and clinicians working on both sides of the study unit immediately before and 12 months after implementation of the MICRO model. Results are reported for the AHRQ dimensions that were most relevant to the MICRO intervention: “teamwork within units,” “overall perceptions of safety,” “communication,” “openness,” “overall patient safety grade,” and “handoffs and transitions.” The survey represents pre- and post-comparison. All nurses and hospitalists on both the MICRO and control sides of the study unit had received the TeamStepps curriculum and participated in MICRO rounds by the time of the postintervention survey. We added 3 questions specifically assessing the perception of the efficiency and effectiveness of IDR. Postintervention respondents reflected on their overall impression of IDR, which included their experiences on both sides of the unit, because no group of nurses or hospitalists was exposed only to the MICRO side or the control side of the unit. Responses to survey questions were recorded on a 5-point Likert scale (from “strongly disagree” to “strongly agree” for opinion questions; and “never,” rarely,” sometimes,” “most of the time,” and “always” for frequency questions) and given a score from 1 to 5. The question asking for an overall grade for patient safety was scored from 1 to 5 points corresponding to letter grade choices F, D, C, B, A.
Statistical Analysis
The sample size was based on the estimate of the baseline rate of the primary outcome of CD and the projected decrease by the MICRO intervention. A study using the Global Trigger Tool developed by the Institute for Healthcare Improvement provided a best estimate of 16% as the baseline rate for CD.7 A total of 2000 hospitalizations were planned to be included to have a power of at least 80% to detect a 25% reduction in the annual incidence of CD with a 2-tailed type I error rate of 0.05. Comparisons of dichotomous event rates were made using chi square tests at a 2-tailed level for significance of 0.05. The LOS was analyzed using the nonparametric median test and multivariable regression analysis. We used a generalized linear model with gamma distribution and log link for all analyses of LOS, where LOS was the outcome variable, and intervention vs. control unit type was the predictor variable. Age, sex, race, payer, case mix, and comorbidities defined with the Elixhauser algorithm were used as covariates.8 We used multivariable logistic regression for analysis of CD, where the dependent variable was CD. Predictor variables included intervention, patient age, sex, race, payer, case mix and comorbidities. Patient satisfaction data were compared using the chi square test. The Student t test for dependent means was used to analyze the patient safety culture survey data.
The study protocol was submitted to the Icahn Mount Sinai School of Medicine’s institutional review board and determined to be exempt from full review.
RESULTS
A total of 2005 hospitalizations were included over the 12-month study period, consisting of 1089 hospitalizations in the MICRO group and 916 in the control group. Bedside and standard IDR were completed daily, Monday through Friday without exception. The demographic characteristics and comorbidities were similar for the 2 groups (Table). Hospitalizations of patients who were initially admitted to another unit and subsequently transferred to the study unit accounted for 11.1% of hospitalizations.
Table
Risk-adjusted LOS was similar for the groups (6.6 vs 7.0 days, P = 0.17, for the MICRO and control groups, respectively). On subgroup analysis, a reduction in LOS was noted for patients transferred to the study unit (10.4 vs 14.0 days, P = 0.02, for the MICRO and control groups, respectively). The LOS was unchanged for patients admitted directly to the study unit (6.0 vs 5.8 days, P = 0.93). There was no difference in the incidence of clinical deterioration for the MICRO or control groups (7.7% vs 9.3%, odds ratio, 0.89; 95% confidence interval, 0.61-1.22, P = 0.46).
The finding of a LOS benefit for the MICRO group limited to patients transferred to the study unit prompted a comparison of patients transferred to the study unit and patients directly admitted to the study unit from the emergency department (Appendix 2). Compared to patients admitted directly to the study unit, patients transferred to the study unit were more likely to have Medicaid or no insurance, more likely to be discharged to a facility, had longer LOS, and were more likely to experience CD.
Patient Satisfaction
There were 175 and 140 responses to the HCAHPS survey for the MICRO and the control groups, respectively. Patients in the MICRO group were more likely to report that “doctors, nurses, or other hospital staff talk with you about whether you would have the help you needed when you left the hospital” (88% vs 78%, P = 0.01). Responses for all other HCAHPS items were similar for the 2 groups.
Figure 1
Clinician/Staff Survey
The response rate was 96% (30 nurses and 17 hospitalists) pre-intervention and 100% (30 nurses and 22 hospitalists) postintervention. Hospitalists and nurses gave significantly higher scores for the dimensions “teamwork within units,” “overall perception of patient safety,” and “patient safety grade” on the postintervention survey compared to the pre-intervention survey (Figure 1). Hospitalists and nurses rated the efficiency of IDR and the ability of IDR to identify safety issues higher on the postintervention survey compared to the pre-intervention survey (Figure 2).
Figure 2
DISCUSSION
We transformed daily IDR from a standard conference room model to a structured bedside model with scripted roles, and performed a rigorous comparison using patient-level data. Our finding that transforming daily IDR from a standard conference room model to a bedside model did not significantly reduce LOS suggests either that the model is ineffective or needs to be incorporated into more comprehensive efforts to improve clinical outcomes. Studies suggest that bedside rounding can improve outcomes when implemented in the context of comprehensive restructuring of patient care.4,9 Stein et al.4 have described the reorganization of a medical ward as an “accountable care unit.” The ACU model included daily IDRs at the bedside, as well as geographic-based teams, co-leadership by a hospitalist and nurse manager, and unit-level reporting. Although no definitive conclusions can be drawn based on their descriptive report, transformation of the unit was associated with reduced LOS and mortality. Similarly, Kara et al.9 found that the number of elements of an “accountable care team” model implemented by each unit was associated with greater reductions in LOS and cost. In contrast, our findings of a lack of an effect are consistent with a recent cluster-randomized trial by O’Leary et al,10 which found that implementation of patient-centered bedside rounds did not improve patient satisfaction or perceptions of shared decision-making compared to units using a model of structured IDRs in a conference room setting. It is notable that the control groups in both the O’Leary trial10 and this study did not represent usual care, because these groups featured localization of the clinical teams and high-quality IDR. In our trial, it is plausible that the control side of the unit was functioning at a high level, which would have decreased our ability to further improve outcomes. Whether restructuring unit processes, including implementation of bedside IDR, improves care compared to usual care without these processes is unknown.
We found that the MICRO intervention significantly decreased LOS compared to the control group for patients transferred to the study unit. This analysis was exploratory and the finding was unexpected. Patients were transferred to the study unit from units of higher acuity, and were more likely to have Medicaid or no insurance and be discharged to facilities rather than home, suggesting that these patients had substantial disposition challenges. It is plausible that this is the population for which bedside IDRs may have the greatest impact. This was a secondary analysis, however, and should be considered as hypothesis-generating for future investigations.
Although the impact on outcomes of bedside IDRs is uncertain, potential benefits and practical barriers have been examined. Gonzalo et al.11 surveyed inpatient physicians and nurses at a hospital employing bedside IDRs and found that the benefits ranked the highest were communication, coordination, and teamwork, and the lowest-ranked benefits were related to efficiency and outcomes. The 6 greatest barriers concerned the time required to complete bedside IDR. These results indicate that the time investiture by staff may be a barrier to widespread adoption. More modest changes, such as increasing the structure of standard conference room rounds, may improve care, although the data are mixed. O’Leary et al.12 assessed the value of a structured approach in a conference room setting, which primarily entailed implementing a checklist for newly admitted patients, and found no difference in LOS. Follow-up studies by these investigators found mixed results on the ability of structured IDR to decrease the incidence of adverse events.13,14
The results of our AHRQ survey of patient safety culture found that several important aspects of teamwork and safety were perceived as improved by the intervention, including the “overall grade on patient safety.” Other studies have similarly shown increases in teamwork and safety ratings through redesign of IDR. O’Leary et al.12 surveyed residents and nurses on a unit that implemented a structured, conference room-based IDR and found that providers on the intervention unit rated the teamwork climate higher than providers on the control unit. Our finding that hospitalists and nurses gave higher ratings for IDR being “efficient” and “a good use of my time” on the postintervention survey than the pre-intervention survey suggests that initial concerns about the additional time commitment may be offset by gains in overall efficiency and in development of an environment of enhanced communication, teamwork, and safety.
This study has several limitations. First, the trial may have been underpowered to find small differences between the groups. The trends for decreased LOS and clinical deterioration in the MICRO group may suggest that bedside IDR can provide a small but clinically significant benefit that would be statistically significant only in a larger trial. Second, patients were not randomized to the 2 groups. The impact is diminished, however, because the routine hospital process for assigning patients to the 2 areas in which the groups were located is random and based solely on bed availability. Third, nurses and hospitalists caring for patients in the control group likely experienced improved communication practices from the unit-wide TeamStepps education and from participating in the MICRO protocol when caring for patients on the intervention side of the unit. Fourth, we did not collect data on the effectiveness of communication and are unable to assess the fidelity with which the structured protocol was followed or whether interprofessional communication was fostered or hindered. Lastly, the study was implemented on a nonteaching unit at a single academic medical center. The protocol and the results may not be generalizable to other hospitals or units that include house staff.
In conclusion, transforming IDR from a conference room model to a bedside model did not reduce overall LOS or clinical deterioration on a unit using features of an ACU structure. Although several beneficial effects were noted, including a reduction in LOS for patients transferred to the study unit and higher ratings of the patient safety climate and efficiency of IDR, implementing bedside IDR in this setting has marginal benefit. Future studies should assess whether a comprehensive transformation of the inpatient model of care, including patient-centered bedside IDR, geographic cohorting of teams, and co-leadership, improves outcomes compared to models without these features.
Disclosures
This trial was funded by Medline’s Prevention Above All Discoveries Grant Program. The authors report no financial conflicts of interest.
1. O’Leary KJ, Sehgal NL, Terrell G, Williams MV; High Performance Teams and the Hospital of the Future Project Team. Interdisciplinary teamwork in hospitals: A review and practical recommendations for improvement. J Hosp Med. 2011;7(1):48-54. PubMed 2. Gonzalo JD, Wolpaw DR, Lehman E, Chuang CH. Patient-centered interprofessional collaborative care: factors associated with bedside interprofessional rounds. J Gen Intern Med. 2014;29(7):1040-1047. PubMed 3. Pannick S, Davis R, Ashrafian H, et al. Effects of interdisciplinary team care interventions on general medical wards. A systematic review. JAMA Intern Med. 2015;175(8):1288-1298. PubMed 4. Stein J, Payne C, Methvin A, et al. Reorganizing a hospital ward as an accountable care unit. J Hosp Med. 2015;10(1):36-40. PubMed 5. Mohr J, Batalden P, Barach P. Integrating patient safety into the clinical microsystem. Qual Saf Health Care. 2004;13(suppl 2):ii34-ii38. PubMed 6. Nelson EC, Batalden PB, Huber TP, et al. Microsystems in health care: Part 1. Learning from high-performing front-line clinical units. Jt Comm J Qual Improv. 2002;28:472-493. PubMed 7. Rutberg H, Borgstedt Risberg MB, Sjödahl R, Nordqvist P, Valter L, Nilsson L. Characterisations of adverse events detected in a university hospital: a 4-year study using the Global Trigger Tool method. BMJ Open. 2014;4(5):e004879. PubMed 8. Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Med Care. 1998;36(1):8-27. PubMed 9. Kara A, Johnson CS, Nicley A, Niemeier MR, Hui SL. Redesigning accountable care: testing the effectiveness of an accountable care team model. J Hosp Med. 2015;10(12):773-779. PubMed 10. O’Leary KJ, Killarney A, Hansen LO, Jones S, Malladi M, Marks K, et al. Effect of patient-centred bedside rounds on hospitalised patients’ decision control, activation and satisfaction with care. BMJ Qual Saf. 2016;25(12):921-928. PubMed 11. Gonzalo JD, Kuperman E, Lehman E, Haidet P. Bedside interprofessional rounds: perceptions of benefits and barriers by internal medicine nursing staff, attending physicians, and housestaff physicians. J Hosp Med. 2014;9(10):646-651. PubMed 12. O’Leary KJ, Wayne DB, Haviley C, Slade ME, Lee J, Williams MV. Improving teamwork: impact of structured interdisciplinary rounds on a medical teaching unit. J Gen Intern Med. 2010;25:826-832. PubMed 13. O’Leary KJ, Buck R, Fligiel HM, et al. Structured interdisciplinary rounds in a medical teaching unit: improving patient safety. Arch Intern Med. 2011;171(7):678-684. PubMed 14. O’Leary KJ, Creden AJ, Slade ME, et al. Implementation of unit-based interventions to improve teamwork and patient safety on a medical service. Am J Med Qual. 2015;30(5):409-416. PubMed
The care of hospitalized patients requires practitioners from multiple disciplines to assess and communicate the patient’s status in a dynamic manner during hospitalization. Although optimal teamwork is needed for patient care to be delivered reliably and efficiently, care within hospitals is typically delivered in a fragmented manner.1 A bedside model for daily interdisciplinary rounds (IDR) has been proposed as a method to provide a structured process and engage all team members in a patient-centered, system-of-care delivery.2 Specific advantages of convening rounds in the presence of the patient include the ability to directly assess care (eg, presence of a potentially unnecessary urinary catheter), patient engagement in key aspects of their care and disposition, and an increased opportunity for team members to develop a shared understanding of the patient’s views and needs.
Implementing dramatic changes to the workflow of multiple disciplines will require rigorous evidence to support a concerted effort from leadership and buy-in from stakeholders at the front line of patient care. Despite the urgency for evidence, there has been little investigation of this strategy. A systematic review3 identified 30 studies published between 1998 and 2013 addressing interdisciplinary interventions on medical wards, none of which examined a bedside IDR model. In a study performed after the period assessed by the systematic review, Stein et al4 described the restructuring of a medical ward as an accountable care unit (ACU), which included a bedside model for rounds by the interdisciplinary team. The change was associated with decreased mortality and length of stay (LOS), although the study did not isolate the impact of rounds or use a concurrent control group and presented aggregate rather than patient-level outcomes. The lack of convincing data may be a reason bedside rounds are not widely employed by hospitals. To provide high-quality evidence, we performed a large, prospective controlled trial comparing a structured bedside model (mobile interdisciplinary care rounds [MICRO]) with standard rounds.
METHODS
This study took place at the Mount Sinai Hospital, which is a 1171-bed tertiary care academic medical center in New York City, New York. A nonteaching unit offered the ability to use a prospective controlled design. Patients were assigned to the north and the south wings of the unit in a quasi-randomized manner, rather than based on diagnosis or acuity. We transformed IDR to a bedside model on the north side of the unit (MICRO group), while the south side of the unit continued using standard conference room-based IDR (control group). The north and south sides of the unit contain 17 and 14 beds, respectively. During the study period, nurses and hospitalists cared for patients on both sides of the study unit, although on any given day were assigned only patients on 1 side of the unit. The unit uses a clinical microsystem model, which has been defined as “a group of clinicians and staff working together with a shared clinical purpose to provide care for a population of patients,” and has a defined set of characteristics associated with high performance.5,6 Our microsystem model has incorporated features as described by Stein’s ACU model,4 including co-leadership by a hospitalist and a nurse manager, geographic assignment of patients to teams, and unit-level data reports. One hospitalist is assigned geographically to each area of the unit in a 2- to 4-week rotation. Coverage of the unit does not include house staff; patients are primarily assigned to hospitalists working with nurse practitioners. Patients were enrolled prospectively during their initial IDR by a research coordinator. Patient-level data and outcomes were collected prospectively by a research coordinator who attended IDR on the intervention and the control sides of the study unit daily.
Inclusion Criteria
All patients admitted to the medicine service on the study unit were eligible. Patients were greater than 18 years and admitted for an acute medical condition. Patients admitted to another unit and later transferred to the study unit were enrolled at the time of transfer. Patients could be included more than once if hospitalized on the study unit on more than 1 occasion. Most patients were covered by hospitalists, although patients covered by private physicians were included. Patients from other departments, including family medicine, are uncommonly admitted to the unit and were excluded. Patients were also excluded if they were admitted and discharged over the same weekend, because the MICRO rounds occur during weekdays and there was no opportunity to offer the intervention on Saturdays and Sundays.
MICRO Intervention
Interdisciplinary rounds occurred daily at 10:00 am for the control group and at 10:30 am for the MICRO group, and were attended by the hospitalist caring for the majority of patients on the unit, staff nurses, and the unit medical director, nurse manager, social worker, and case manager. Rounds on the control unit focused on the plan of care and disposition but did not follow any set structure and were typically 25 to 30 minutes in duration.
The MICRO rounds occurred at the bedside and followed a structured script (Appendix 1) that was designed to limit discussion of each patient to 3 minutes or less, and included speaking roles for the hospitalist, nurse, and social worker. For private physicians, the nurse practitioner assigned to the patient performed the role of the hospitalist. Rounds were expected to be approximately 50 minutes in duration. Patients were further engaged by asking for their main goal for the day. A patient safety checklist was reviewed. Initially, this task was performed by the nurse manager, who did not verbalize the items unless a deficiency was noted. After 6 months’ experience, this responsibility was given to the staff nurse, who reviewed the checklist verbally as part of the bedside script. Patients were seen daily, including those being discharged later that same day.
Staff and Clinician Education
We developed and implemented a curriculum based on a modified version of the Agency for Healthcare Research and Quality’s TeamStepps® program to ensure that all team members were provided with the basic principles of communication within the healthcare setting. The curriculum consisted of interactive didactics on essential elements of teamwork, including team structure, communication, situation monitoring, and mutual support, as well as the purpose and structure of the MICRO model. The curriculum was delivered to nurses at 3 monthly staff meetings on the study unit and to hospitalists during 3 hospital medicine grand rounds over a 3-month period. Nurses and physicians providing care on both geographic areas of the study unit received the education program because no group of practitioners was designated to only 1 geographic area.
Outcomes
Primary and Secondary Outcomes
The primary outcomes were clinical deterioration (CD) and length of stay. Clinical deterioration was a composite outcome defined a priori as death; escalation of care (ie, transfer to an intensive care unit, intermediate care unit, or teaching unit); or a hospital-acquired complication (ie, venous thromboembolism, fall, stage III-IV pressure ulcer, catheter-associated urinary tract infection, central-line associated bloodstream infection, or Clostridium difficile-associated diarrhea). The LOS was calculated as the mean LOS with outliers excluded (outliers defined as having a LOS 100 days or longer or 2.5 or more standard deviations from the expected LOS).
Process metrics on IDR, such as the duration of rounds, attendance by members of the interdisciplinary team, the percentage of patients discussed, or the effectiveness of communication, were not collected. We assessed patient satisfaction based on the Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS) survey.
Patient Safety Culture Survey
To assess the impact on the perceptions of patient safety, we administered the Agency for Healthcare Research and Quality (AHRQ) Hospital Survey on Patient Safety Culture toall staff and clinicians working on both sides of the study unit immediately before and 12 months after implementation of the MICRO model. Results are reported for the AHRQ dimensions that were most relevant to the MICRO intervention: “teamwork within units,” “overall perceptions of safety,” “communication,” “openness,” “overall patient safety grade,” and “handoffs and transitions.” The survey represents pre- and post-comparison. All nurses and hospitalists on both the MICRO and control sides of the study unit had received the TeamStepps curriculum and participated in MICRO rounds by the time of the postintervention survey. We added 3 questions specifically assessing the perception of the efficiency and effectiveness of IDR. Postintervention respondents reflected on their overall impression of IDR, which included their experiences on both sides of the unit, because no group of nurses or hospitalists was exposed only to the MICRO side or the control side of the unit. Responses to survey questions were recorded on a 5-point Likert scale (from “strongly disagree” to “strongly agree” for opinion questions; and “never,” rarely,” sometimes,” “most of the time,” and “always” for frequency questions) and given a score from 1 to 5. The question asking for an overall grade for patient safety was scored from 1 to 5 points corresponding to letter grade choices F, D, C, B, A.
Statistical Analysis
The sample size was based on the estimate of the baseline rate of the primary outcome of CD and the projected decrease by the MICRO intervention. A study using the Global Trigger Tool developed by the Institute for Healthcare Improvement provided a best estimate of 16% as the baseline rate for CD.7 A total of 2000 hospitalizations were planned to be included to have a power of at least 80% to detect a 25% reduction in the annual incidence of CD with a 2-tailed type I error rate of 0.05. Comparisons of dichotomous event rates were made using chi square tests at a 2-tailed level for significance of 0.05. The LOS was analyzed using the nonparametric median test and multivariable regression analysis. We used a generalized linear model with gamma distribution and log link for all analyses of LOS, where LOS was the outcome variable, and intervention vs. control unit type was the predictor variable. Age, sex, race, payer, case mix, and comorbidities defined with the Elixhauser algorithm were used as covariates.8 We used multivariable logistic regression for analysis of CD, where the dependent variable was CD. Predictor variables included intervention, patient age, sex, race, payer, case mix and comorbidities. Patient satisfaction data were compared using the chi square test. The Student t test for dependent means was used to analyze the patient safety culture survey data.
The study protocol was submitted to the Icahn Mount Sinai School of Medicine’s institutional review board and determined to be exempt from full review.
RESULTS
A total of 2005 hospitalizations were included over the 12-month study period, consisting of 1089 hospitalizations in the MICRO group and 916 in the control group. Bedside and standard IDR were completed daily, Monday through Friday without exception. The demographic characteristics and comorbidities were similar for the 2 groups (Table). Hospitalizations of patients who were initially admitted to another unit and subsequently transferred to the study unit accounted for 11.1% of hospitalizations.
Table
Risk-adjusted LOS was similar for the groups (6.6 vs 7.0 days, P = 0.17, for the MICRO and control groups, respectively). On subgroup analysis, a reduction in LOS was noted for patients transferred to the study unit (10.4 vs 14.0 days, P = 0.02, for the MICRO and control groups, respectively). The LOS was unchanged for patients admitted directly to the study unit (6.0 vs 5.8 days, P = 0.93). There was no difference in the incidence of clinical deterioration for the MICRO or control groups (7.7% vs 9.3%, odds ratio, 0.89; 95% confidence interval, 0.61-1.22, P = 0.46).
The finding of a LOS benefit for the MICRO group limited to patients transferred to the study unit prompted a comparison of patients transferred to the study unit and patients directly admitted to the study unit from the emergency department (Appendix 2). Compared to patients admitted directly to the study unit, patients transferred to the study unit were more likely to have Medicaid or no insurance, more likely to be discharged to a facility, had longer LOS, and were more likely to experience CD.
Patient Satisfaction
There were 175 and 140 responses to the HCAHPS survey for the MICRO and the control groups, respectively. Patients in the MICRO group were more likely to report that “doctors, nurses, or other hospital staff talk with you about whether you would have the help you needed when you left the hospital” (88% vs 78%, P = 0.01). Responses for all other HCAHPS items were similar for the 2 groups.
Figure 1
Clinician/Staff Survey
The response rate was 96% (30 nurses and 17 hospitalists) pre-intervention and 100% (30 nurses and 22 hospitalists) postintervention. Hospitalists and nurses gave significantly higher scores for the dimensions “teamwork within units,” “overall perception of patient safety,” and “patient safety grade” on the postintervention survey compared to the pre-intervention survey (Figure 1). Hospitalists and nurses rated the efficiency of IDR and the ability of IDR to identify safety issues higher on the postintervention survey compared to the pre-intervention survey (Figure 2).
Figure 2
DISCUSSION
We transformed daily IDR from a standard conference room model to a structured bedside model with scripted roles, and performed a rigorous comparison using patient-level data. Our finding that transforming daily IDR from a standard conference room model to a bedside model did not significantly reduce LOS suggests either that the model is ineffective or needs to be incorporated into more comprehensive efforts to improve clinical outcomes. Studies suggest that bedside rounding can improve outcomes when implemented in the context of comprehensive restructuring of patient care.4,9 Stein et al.4 have described the reorganization of a medical ward as an “accountable care unit.” The ACU model included daily IDRs at the bedside, as well as geographic-based teams, co-leadership by a hospitalist and nurse manager, and unit-level reporting. Although no definitive conclusions can be drawn based on their descriptive report, transformation of the unit was associated with reduced LOS and mortality. Similarly, Kara et al.9 found that the number of elements of an “accountable care team” model implemented by each unit was associated with greater reductions in LOS and cost. In contrast, our findings of a lack of an effect are consistent with a recent cluster-randomized trial by O’Leary et al,10 which found that implementation of patient-centered bedside rounds did not improve patient satisfaction or perceptions of shared decision-making compared to units using a model of structured IDRs in a conference room setting. It is notable that the control groups in both the O’Leary trial10 and this study did not represent usual care, because these groups featured localization of the clinical teams and high-quality IDR. In our trial, it is plausible that the control side of the unit was functioning at a high level, which would have decreased our ability to further improve outcomes. Whether restructuring unit processes, including implementation of bedside IDR, improves care compared to usual care without these processes is unknown.
We found that the MICRO intervention significantly decreased LOS compared to the control group for patients transferred to the study unit. This analysis was exploratory and the finding was unexpected. Patients were transferred to the study unit from units of higher acuity, and were more likely to have Medicaid or no insurance and be discharged to facilities rather than home, suggesting that these patients had substantial disposition challenges. It is plausible that this is the population for which bedside IDRs may have the greatest impact. This was a secondary analysis, however, and should be considered as hypothesis-generating for future investigations.
Although the impact on outcomes of bedside IDRs is uncertain, potential benefits and practical barriers have been examined. Gonzalo et al.11 surveyed inpatient physicians and nurses at a hospital employing bedside IDRs and found that the benefits ranked the highest were communication, coordination, and teamwork, and the lowest-ranked benefits were related to efficiency and outcomes. The 6 greatest barriers concerned the time required to complete bedside IDR. These results indicate that the time investiture by staff may be a barrier to widespread adoption. More modest changes, such as increasing the structure of standard conference room rounds, may improve care, although the data are mixed. O’Leary et al.12 assessed the value of a structured approach in a conference room setting, which primarily entailed implementing a checklist for newly admitted patients, and found no difference in LOS. Follow-up studies by these investigators found mixed results on the ability of structured IDR to decrease the incidence of adverse events.13,14
The results of our AHRQ survey of patient safety culture found that several important aspects of teamwork and safety were perceived as improved by the intervention, including the “overall grade on patient safety.” Other studies have similarly shown increases in teamwork and safety ratings through redesign of IDR. O’Leary et al.12 surveyed residents and nurses on a unit that implemented a structured, conference room-based IDR and found that providers on the intervention unit rated the teamwork climate higher than providers on the control unit. Our finding that hospitalists and nurses gave higher ratings for IDR being “efficient” and “a good use of my time” on the postintervention survey than the pre-intervention survey suggests that initial concerns about the additional time commitment may be offset by gains in overall efficiency and in development of an environment of enhanced communication, teamwork, and safety.
This study has several limitations. First, the trial may have been underpowered to find small differences between the groups. The trends for decreased LOS and clinical deterioration in the MICRO group may suggest that bedside IDR can provide a small but clinically significant benefit that would be statistically significant only in a larger trial. Second, patients were not randomized to the 2 groups. The impact is diminished, however, because the routine hospital process for assigning patients to the 2 areas in which the groups were located is random and based solely on bed availability. Third, nurses and hospitalists caring for patients in the control group likely experienced improved communication practices from the unit-wide TeamStepps education and from participating in the MICRO protocol when caring for patients on the intervention side of the unit. Fourth, we did not collect data on the effectiveness of communication and are unable to assess the fidelity with which the structured protocol was followed or whether interprofessional communication was fostered or hindered. Lastly, the study was implemented on a nonteaching unit at a single academic medical center. The protocol and the results may not be generalizable to other hospitals or units that include house staff.
In conclusion, transforming IDR from a conference room model to a bedside model did not reduce overall LOS or clinical deterioration on a unit using features of an ACU structure. Although several beneficial effects were noted, including a reduction in LOS for patients transferred to the study unit and higher ratings of the patient safety climate and efficiency of IDR, implementing bedside IDR in this setting has marginal benefit. Future studies should assess whether a comprehensive transformation of the inpatient model of care, including patient-centered bedside IDR, geographic cohorting of teams, and co-leadership, improves outcomes compared to models without these features.
Disclosures
This trial was funded by Medline’s Prevention Above All Discoveries Grant Program. The authors report no financial conflicts of interest.
The care of hospitalized patients requires practitioners from multiple disciplines to assess and communicate the patient’s status in a dynamic manner during hospitalization. Although optimal teamwork is needed for patient care to be delivered reliably and efficiently, care within hospitals is typically delivered in a fragmented manner.1 A bedside model for daily interdisciplinary rounds (IDR) has been proposed as a method to provide a structured process and engage all team members in a patient-centered, system-of-care delivery.2 Specific advantages of convening rounds in the presence of the patient include the ability to directly assess care (eg, presence of a potentially unnecessary urinary catheter), patient engagement in key aspects of their care and disposition, and an increased opportunity for team members to develop a shared understanding of the patient’s views and needs.
Implementing dramatic changes to the workflow of multiple disciplines will require rigorous evidence to support a concerted effort from leadership and buy-in from stakeholders at the front line of patient care. Despite the urgency for evidence, there has been little investigation of this strategy. A systematic review3 identified 30 studies published between 1998 and 2013 addressing interdisciplinary interventions on medical wards, none of which examined a bedside IDR model. In a study performed after the period assessed by the systematic review, Stein et al4 described the restructuring of a medical ward as an accountable care unit (ACU), which included a bedside model for rounds by the interdisciplinary team. The change was associated with decreased mortality and length of stay (LOS), although the study did not isolate the impact of rounds or use a concurrent control group and presented aggregate rather than patient-level outcomes. The lack of convincing data may be a reason bedside rounds are not widely employed by hospitals. To provide high-quality evidence, we performed a large, prospective controlled trial comparing a structured bedside model (mobile interdisciplinary care rounds [MICRO]) with standard rounds.
METHODS
This study took place at the Mount Sinai Hospital, which is a 1171-bed tertiary care academic medical center in New York City, New York. A nonteaching unit offered the ability to use a prospective controlled design. Patients were assigned to the north and the south wings of the unit in a quasi-randomized manner, rather than based on diagnosis or acuity. We transformed IDR to a bedside model on the north side of the unit (MICRO group), while the south side of the unit continued using standard conference room-based IDR (control group). The north and south sides of the unit contain 17 and 14 beds, respectively. During the study period, nurses and hospitalists cared for patients on both sides of the study unit, although on any given day were assigned only patients on 1 side of the unit. The unit uses a clinical microsystem model, which has been defined as “a group of clinicians and staff working together with a shared clinical purpose to provide care for a population of patients,” and has a defined set of characteristics associated with high performance.5,6 Our microsystem model has incorporated features as described by Stein’s ACU model,4 including co-leadership by a hospitalist and a nurse manager, geographic assignment of patients to teams, and unit-level data reports. One hospitalist is assigned geographically to each area of the unit in a 2- to 4-week rotation. Coverage of the unit does not include house staff; patients are primarily assigned to hospitalists working with nurse practitioners. Patients were enrolled prospectively during their initial IDR by a research coordinator. Patient-level data and outcomes were collected prospectively by a research coordinator who attended IDR on the intervention and the control sides of the study unit daily.
Inclusion Criteria
All patients admitted to the medicine service on the study unit were eligible. Patients were greater than 18 years and admitted for an acute medical condition. Patients admitted to another unit and later transferred to the study unit were enrolled at the time of transfer. Patients could be included more than once if hospitalized on the study unit on more than 1 occasion. Most patients were covered by hospitalists, although patients covered by private physicians were included. Patients from other departments, including family medicine, are uncommonly admitted to the unit and were excluded. Patients were also excluded if they were admitted and discharged over the same weekend, because the MICRO rounds occur during weekdays and there was no opportunity to offer the intervention on Saturdays and Sundays.
MICRO Intervention
Interdisciplinary rounds occurred daily at 10:00 am for the control group and at 10:30 am for the MICRO group, and were attended by the hospitalist caring for the majority of patients on the unit, staff nurses, and the unit medical director, nurse manager, social worker, and case manager. Rounds on the control unit focused on the plan of care and disposition but did not follow any set structure and were typically 25 to 30 minutes in duration.
The MICRO rounds occurred at the bedside and followed a structured script (Appendix 1) that was designed to limit discussion of each patient to 3 minutes or less, and included speaking roles for the hospitalist, nurse, and social worker. For private physicians, the nurse practitioner assigned to the patient performed the role of the hospitalist. Rounds were expected to be approximately 50 minutes in duration. Patients were further engaged by asking for their main goal for the day. A patient safety checklist was reviewed. Initially, this task was performed by the nurse manager, who did not verbalize the items unless a deficiency was noted. After 6 months’ experience, this responsibility was given to the staff nurse, who reviewed the checklist verbally as part of the bedside script. Patients were seen daily, including those being discharged later that same day.
Staff and Clinician Education
We developed and implemented a curriculum based on a modified version of the Agency for Healthcare Research and Quality’s TeamStepps® program to ensure that all team members were provided with the basic principles of communication within the healthcare setting. The curriculum consisted of interactive didactics on essential elements of teamwork, including team structure, communication, situation monitoring, and mutual support, as well as the purpose and structure of the MICRO model. The curriculum was delivered to nurses at 3 monthly staff meetings on the study unit and to hospitalists during 3 hospital medicine grand rounds over a 3-month period. Nurses and physicians providing care on both geographic areas of the study unit received the education program because no group of practitioners was designated to only 1 geographic area.
Outcomes
Primary and Secondary Outcomes
The primary outcomes were clinical deterioration (CD) and length of stay. Clinical deterioration was a composite outcome defined a priori as death; escalation of care (ie, transfer to an intensive care unit, intermediate care unit, or teaching unit); or a hospital-acquired complication (ie, venous thromboembolism, fall, stage III-IV pressure ulcer, catheter-associated urinary tract infection, central-line associated bloodstream infection, or Clostridium difficile-associated diarrhea). The LOS was calculated as the mean LOS with outliers excluded (outliers defined as having a LOS 100 days or longer or 2.5 or more standard deviations from the expected LOS).
Process metrics on IDR, such as the duration of rounds, attendance by members of the interdisciplinary team, the percentage of patients discussed, or the effectiveness of communication, were not collected. We assessed patient satisfaction based on the Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS) survey.
Patient Safety Culture Survey
To assess the impact on the perceptions of patient safety, we administered the Agency for Healthcare Research and Quality (AHRQ) Hospital Survey on Patient Safety Culture toall staff and clinicians working on both sides of the study unit immediately before and 12 months after implementation of the MICRO model. Results are reported for the AHRQ dimensions that were most relevant to the MICRO intervention: “teamwork within units,” “overall perceptions of safety,” “communication,” “openness,” “overall patient safety grade,” and “handoffs and transitions.” The survey represents pre- and post-comparison. All nurses and hospitalists on both the MICRO and control sides of the study unit had received the TeamStepps curriculum and participated in MICRO rounds by the time of the postintervention survey. We added 3 questions specifically assessing the perception of the efficiency and effectiveness of IDR. Postintervention respondents reflected on their overall impression of IDR, which included their experiences on both sides of the unit, because no group of nurses or hospitalists was exposed only to the MICRO side or the control side of the unit. Responses to survey questions were recorded on a 5-point Likert scale (from “strongly disagree” to “strongly agree” for opinion questions; and “never,” rarely,” sometimes,” “most of the time,” and “always” for frequency questions) and given a score from 1 to 5. The question asking for an overall grade for patient safety was scored from 1 to 5 points corresponding to letter grade choices F, D, C, B, A.
Statistical Analysis
The sample size was based on the estimate of the baseline rate of the primary outcome of CD and the projected decrease by the MICRO intervention. A study using the Global Trigger Tool developed by the Institute for Healthcare Improvement provided a best estimate of 16% as the baseline rate for CD.7 A total of 2000 hospitalizations were planned to be included to have a power of at least 80% to detect a 25% reduction in the annual incidence of CD with a 2-tailed type I error rate of 0.05. Comparisons of dichotomous event rates were made using chi square tests at a 2-tailed level for significance of 0.05. The LOS was analyzed using the nonparametric median test and multivariable regression analysis. We used a generalized linear model with gamma distribution and log link for all analyses of LOS, where LOS was the outcome variable, and intervention vs. control unit type was the predictor variable. Age, sex, race, payer, case mix, and comorbidities defined with the Elixhauser algorithm were used as covariates.8 We used multivariable logistic regression for analysis of CD, where the dependent variable was CD. Predictor variables included intervention, patient age, sex, race, payer, case mix and comorbidities. Patient satisfaction data were compared using the chi square test. The Student t test for dependent means was used to analyze the patient safety culture survey data.
The study protocol was submitted to the Icahn Mount Sinai School of Medicine’s institutional review board and determined to be exempt from full review.
RESULTS
A total of 2005 hospitalizations were included over the 12-month study period, consisting of 1089 hospitalizations in the MICRO group and 916 in the control group. Bedside and standard IDR were completed daily, Monday through Friday without exception. The demographic characteristics and comorbidities were similar for the 2 groups (Table). Hospitalizations of patients who were initially admitted to another unit and subsequently transferred to the study unit accounted for 11.1% of hospitalizations.
Table
Risk-adjusted LOS was similar for the groups (6.6 vs 7.0 days, P = 0.17, for the MICRO and control groups, respectively). On subgroup analysis, a reduction in LOS was noted for patients transferred to the study unit (10.4 vs 14.0 days, P = 0.02, for the MICRO and control groups, respectively). The LOS was unchanged for patients admitted directly to the study unit (6.0 vs 5.8 days, P = 0.93). There was no difference in the incidence of clinical deterioration for the MICRO or control groups (7.7% vs 9.3%, odds ratio, 0.89; 95% confidence interval, 0.61-1.22, P = 0.46).
The finding of a LOS benefit for the MICRO group limited to patients transferred to the study unit prompted a comparison of patients transferred to the study unit and patients directly admitted to the study unit from the emergency department (Appendix 2). Compared to patients admitted directly to the study unit, patients transferred to the study unit were more likely to have Medicaid or no insurance, more likely to be discharged to a facility, had longer LOS, and were more likely to experience CD.
Patient Satisfaction
There were 175 and 140 responses to the HCAHPS survey for the MICRO and the control groups, respectively. Patients in the MICRO group were more likely to report that “doctors, nurses, or other hospital staff talk with you about whether you would have the help you needed when you left the hospital” (88% vs 78%, P = 0.01). Responses for all other HCAHPS items were similar for the 2 groups.
Figure 1
Clinician/Staff Survey
The response rate was 96% (30 nurses and 17 hospitalists) pre-intervention and 100% (30 nurses and 22 hospitalists) postintervention. Hospitalists and nurses gave significantly higher scores for the dimensions “teamwork within units,” “overall perception of patient safety,” and “patient safety grade” on the postintervention survey compared to the pre-intervention survey (Figure 1). Hospitalists and nurses rated the efficiency of IDR and the ability of IDR to identify safety issues higher on the postintervention survey compared to the pre-intervention survey (Figure 2).
Figure 2
DISCUSSION
We transformed daily IDR from a standard conference room model to a structured bedside model with scripted roles, and performed a rigorous comparison using patient-level data. Our finding that transforming daily IDR from a standard conference room model to a bedside model did not significantly reduce LOS suggests either that the model is ineffective or needs to be incorporated into more comprehensive efforts to improve clinical outcomes. Studies suggest that bedside rounding can improve outcomes when implemented in the context of comprehensive restructuring of patient care.4,9 Stein et al.4 have described the reorganization of a medical ward as an “accountable care unit.” The ACU model included daily IDRs at the bedside, as well as geographic-based teams, co-leadership by a hospitalist and nurse manager, and unit-level reporting. Although no definitive conclusions can be drawn based on their descriptive report, transformation of the unit was associated with reduced LOS and mortality. Similarly, Kara et al.9 found that the number of elements of an “accountable care team” model implemented by each unit was associated with greater reductions in LOS and cost. In contrast, our findings of a lack of an effect are consistent with a recent cluster-randomized trial by O’Leary et al,10 which found that implementation of patient-centered bedside rounds did not improve patient satisfaction or perceptions of shared decision-making compared to units using a model of structured IDRs in a conference room setting. It is notable that the control groups in both the O’Leary trial10 and this study did not represent usual care, because these groups featured localization of the clinical teams and high-quality IDR. In our trial, it is plausible that the control side of the unit was functioning at a high level, which would have decreased our ability to further improve outcomes. Whether restructuring unit processes, including implementation of bedside IDR, improves care compared to usual care without these processes is unknown.
We found that the MICRO intervention significantly decreased LOS compared to the control group for patients transferred to the study unit. This analysis was exploratory and the finding was unexpected. Patients were transferred to the study unit from units of higher acuity, and were more likely to have Medicaid or no insurance and be discharged to facilities rather than home, suggesting that these patients had substantial disposition challenges. It is plausible that this is the population for which bedside IDRs may have the greatest impact. This was a secondary analysis, however, and should be considered as hypothesis-generating for future investigations.
Although the impact on outcomes of bedside IDRs is uncertain, potential benefits and practical barriers have been examined. Gonzalo et al.11 surveyed inpatient physicians and nurses at a hospital employing bedside IDRs and found that the benefits ranked the highest were communication, coordination, and teamwork, and the lowest-ranked benefits were related to efficiency and outcomes. The 6 greatest barriers concerned the time required to complete bedside IDR. These results indicate that the time investiture by staff may be a barrier to widespread adoption. More modest changes, such as increasing the structure of standard conference room rounds, may improve care, although the data are mixed. O’Leary et al.12 assessed the value of a structured approach in a conference room setting, which primarily entailed implementing a checklist for newly admitted patients, and found no difference in LOS. Follow-up studies by these investigators found mixed results on the ability of structured IDR to decrease the incidence of adverse events.13,14
The results of our AHRQ survey of patient safety culture found that several important aspects of teamwork and safety were perceived as improved by the intervention, including the “overall grade on patient safety.” Other studies have similarly shown increases in teamwork and safety ratings through redesign of IDR. O’Leary et al.12 surveyed residents and nurses on a unit that implemented a structured, conference room-based IDR and found that providers on the intervention unit rated the teamwork climate higher than providers on the control unit. Our finding that hospitalists and nurses gave higher ratings for IDR being “efficient” and “a good use of my time” on the postintervention survey than the pre-intervention survey suggests that initial concerns about the additional time commitment may be offset by gains in overall efficiency and in development of an environment of enhanced communication, teamwork, and safety.
This study has several limitations. First, the trial may have been underpowered to find small differences between the groups. The trends for decreased LOS and clinical deterioration in the MICRO group may suggest that bedside IDR can provide a small but clinically significant benefit that would be statistically significant only in a larger trial. Second, patients were not randomized to the 2 groups. The impact is diminished, however, because the routine hospital process for assigning patients to the 2 areas in which the groups were located is random and based solely on bed availability. Third, nurses and hospitalists caring for patients in the control group likely experienced improved communication practices from the unit-wide TeamStepps education and from participating in the MICRO protocol when caring for patients on the intervention side of the unit. Fourth, we did not collect data on the effectiveness of communication and are unable to assess the fidelity with which the structured protocol was followed or whether interprofessional communication was fostered or hindered. Lastly, the study was implemented on a nonteaching unit at a single academic medical center. The protocol and the results may not be generalizable to other hospitals or units that include house staff.
In conclusion, transforming IDR from a conference room model to a bedside model did not reduce overall LOS or clinical deterioration on a unit using features of an ACU structure. Although several beneficial effects were noted, including a reduction in LOS for patients transferred to the study unit and higher ratings of the patient safety climate and efficiency of IDR, implementing bedside IDR in this setting has marginal benefit. Future studies should assess whether a comprehensive transformation of the inpatient model of care, including patient-centered bedside IDR, geographic cohorting of teams, and co-leadership, improves outcomes compared to models without these features.
Disclosures
This trial was funded by Medline’s Prevention Above All Discoveries Grant Program. The authors report no financial conflicts of interest.
References
1. O’Leary KJ, Sehgal NL, Terrell G, Williams MV; High Performance Teams and the Hospital of the Future Project Team. Interdisciplinary teamwork in hospitals: A review and practical recommendations for improvement. J Hosp Med. 2011;7(1):48-54. PubMed 2. Gonzalo JD, Wolpaw DR, Lehman E, Chuang CH. Patient-centered interprofessional collaborative care: factors associated with bedside interprofessional rounds. J Gen Intern Med. 2014;29(7):1040-1047. PubMed 3. Pannick S, Davis R, Ashrafian H, et al. Effects of interdisciplinary team care interventions on general medical wards. A systematic review. JAMA Intern Med. 2015;175(8):1288-1298. PubMed 4. Stein J, Payne C, Methvin A, et al. Reorganizing a hospital ward as an accountable care unit. J Hosp Med. 2015;10(1):36-40. PubMed 5. Mohr J, Batalden P, Barach P. Integrating patient safety into the clinical microsystem. Qual Saf Health Care. 2004;13(suppl 2):ii34-ii38. PubMed 6. Nelson EC, Batalden PB, Huber TP, et al. Microsystems in health care: Part 1. Learning from high-performing front-line clinical units. Jt Comm J Qual Improv. 2002;28:472-493. PubMed 7. Rutberg H, Borgstedt Risberg MB, Sjödahl R, Nordqvist P, Valter L, Nilsson L. Characterisations of adverse events detected in a university hospital: a 4-year study using the Global Trigger Tool method. BMJ Open. 2014;4(5):e004879. PubMed 8. Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Med Care. 1998;36(1):8-27. PubMed 9. Kara A, Johnson CS, Nicley A, Niemeier MR, Hui SL. Redesigning accountable care: testing the effectiveness of an accountable care team model. J Hosp Med. 2015;10(12):773-779. PubMed 10. O’Leary KJ, Killarney A, Hansen LO, Jones S, Malladi M, Marks K, et al. Effect of patient-centred bedside rounds on hospitalised patients’ decision control, activation and satisfaction with care. BMJ Qual Saf. 2016;25(12):921-928. PubMed 11. Gonzalo JD, Kuperman E, Lehman E, Haidet P. Bedside interprofessional rounds: perceptions of benefits and barriers by internal medicine nursing staff, attending physicians, and housestaff physicians. J Hosp Med. 2014;9(10):646-651. PubMed 12. O’Leary KJ, Wayne DB, Haviley C, Slade ME, Lee J, Williams MV. Improving teamwork: impact of structured interdisciplinary rounds on a medical teaching unit. J Gen Intern Med. 2010;25:826-832. PubMed 13. O’Leary KJ, Buck R, Fligiel HM, et al. Structured interdisciplinary rounds in a medical teaching unit: improving patient safety. Arch Intern Med. 2011;171(7):678-684. PubMed 14. O’Leary KJ, Creden AJ, Slade ME, et al. Implementation of unit-based interventions to improve teamwork and patient safety on a medical service. Am J Med Qual. 2015;30(5):409-416. PubMed
References
1. O’Leary KJ, Sehgal NL, Terrell G, Williams MV; High Performance Teams and the Hospital of the Future Project Team. Interdisciplinary teamwork in hospitals: A review and practical recommendations for improvement. J Hosp Med. 2011;7(1):48-54. PubMed 2. Gonzalo JD, Wolpaw DR, Lehman E, Chuang CH. Patient-centered interprofessional collaborative care: factors associated with bedside interprofessional rounds. J Gen Intern Med. 2014;29(7):1040-1047. PubMed 3. Pannick S, Davis R, Ashrafian H, et al. Effects of interdisciplinary team care interventions on general medical wards. A systematic review. JAMA Intern Med. 2015;175(8):1288-1298. PubMed 4. Stein J, Payne C, Methvin A, et al. Reorganizing a hospital ward as an accountable care unit. J Hosp Med. 2015;10(1):36-40. PubMed 5. Mohr J, Batalden P, Barach P. Integrating patient safety into the clinical microsystem. Qual Saf Health Care. 2004;13(suppl 2):ii34-ii38. PubMed 6. Nelson EC, Batalden PB, Huber TP, et al. Microsystems in health care: Part 1. Learning from high-performing front-line clinical units. Jt Comm J Qual Improv. 2002;28:472-493. PubMed 7. Rutberg H, Borgstedt Risberg MB, Sjödahl R, Nordqvist P, Valter L, Nilsson L. Characterisations of adverse events detected in a university hospital: a 4-year study using the Global Trigger Tool method. BMJ Open. 2014;4(5):e004879. PubMed 8. Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Med Care. 1998;36(1):8-27. PubMed 9. Kara A, Johnson CS, Nicley A, Niemeier MR, Hui SL. Redesigning accountable care: testing the effectiveness of an accountable care team model. J Hosp Med. 2015;10(12):773-779. PubMed 10. O’Leary KJ, Killarney A, Hansen LO, Jones S, Malladi M, Marks K, et al. Effect of patient-centred bedside rounds on hospitalised patients’ decision control, activation and satisfaction with care. BMJ Qual Saf. 2016;25(12):921-928. PubMed 11. Gonzalo JD, Kuperman E, Lehman E, Haidet P. Bedside interprofessional rounds: perceptions of benefits and barriers by internal medicine nursing staff, attending physicians, and housestaff physicians. J Hosp Med. 2014;9(10):646-651. PubMed 12. O’Leary KJ, Wayne DB, Haviley C, Slade ME, Lee J, Williams MV. Improving teamwork: impact of structured interdisciplinary rounds on a medical teaching unit. J Gen Intern Med. 2010;25:826-832. PubMed 13. O’Leary KJ, Buck R, Fligiel HM, et al. Structured interdisciplinary rounds in a medical teaching unit: improving patient safety. Arch Intern Med. 2011;171(7):678-684. PubMed 14. O’Leary KJ, Creden AJ, Slade ME, et al. Implementation of unit-based interventions to improve teamwork and patient safety on a medical service. Am J Med Qual. 2015;30(5):409-416. PubMed
Address for correspondence and reprint requests: Andrew S. Dunn, MD, MPH, Chief, Division of Hospital Medicine, Mount Sinai Health System, 1468 Madison Ave, Box 1087, New York, NY 10029; Telephone: 212-241-2920; Fax: 212-289-6393; E-mail: andrew.dunn@mountsinai.org
Content Gating
Gated (full article locked unless allowed per User)
Attending rounds at academic medical centers are often disconnected from patients and non-physician care team members. Time spent bedside is consistently less than one third of total rounding time, with observational studies reporting a range of 9% to 33% over the past several decades.1-8 Rounds are often conducted outside patient rooms, denying patients, families, and nurses the opportunity to participate and offer valuable insights. Lack of bedside rounds thus limits patient and family engagement, patient input into the care plan, teaching of the physical examination, and communication and collaboration with nurses. In one study, physicians and nurses on rounds engaged in interprofessional communication in only 12% of patient cases.1 Studies have found interdisciplinary bedside rounds have several benefits, including subjectively improved communication and teamwork between physicians and nurses; increased patient satisfaction, including feeling more cared for by the medical team; and decreased length of stay and costs of care.2-10
However, there are many barriers to conducting interdisciplinary bedside rounds at large academic medical centers. Patients cared for by a single medical team are often geographically dispersed to several nursing units, and nurses are unable to predict when physicians will round on their patients. This situation limits nursing involvement on rounds and keeps doctors and nurses isolated from each other.2 Regionalization of care teams reduces this fragmentation by facilitating more interaction among doctors, patients, families, and nursing staff.
There are few data on how regionalized patients and interdisciplinary bedside rounds affect rounding time and the nature of rounds. This information is needed to understand how these structural changes mediate their effects, whether other steps are required to optimize outcomes, and how to maximize efficiency. We used time-motion analysis (TMA) to investigate how regionalization of medical teams, encouragement of bedside rounding, and systematic inclusion of nurses on ward rounds affect amount of time spent with patients, nursing presence on rounds, and total rounding time.
METHODS
Setting
This prospective interventional study, approved by the Institutional Review Board of Partners HealthCare, was conducted on the general medical wards at Brigham and Women’s Hospital, an academic 793-bed tertiary-care center in Boston, Massachusetts. Housestaff teams consist of 1 attending, 1 resident, and 2 interns with or without a medical student. Before June 20, 2013, daily rounds on medical inpatients were conducted largely on the patient unit but outside patient rooms. After completing most of a rounding discussion outside a patient’s room, the team might walk in to examine or speak with the patient. A typical medical team had patients dispersed over 7 medical units on average, and over as many as 13. As nurses were unit based, they did not consistently participate in rounds.
Intervention
In June 2013, as part of a general medical service care redesign initiative, the general medical teams were regionalized to specific inpatient units. The goal was to have teams admit patients predominantly to the team’s designated unit and to have all patients on a unit be cared for by the unit’s assigned team as often as possible, with an 85% goal for both. Toward those ends, the admitting structure was changed from a traditional 4-day call cycle to daily admitting for all teams, based on each unit’s bed availability.11
Teams were also expected to conduct rounds with nurses, and a system for facilitating these rounds was established. As physician and nurse care teams were now geographically co-located, it became possible for residents and nurses to check a rounding sheet for the planned patient rounding order, which had been set by the resident and nurse-in-charge before rounds. No more than about 5 minutes was needed to prepare each day’s order. The rounding sheet prioritized sick patients, newly admitted patients, and planned morning discharges, but patients were also always grouped by nurse. For example, the physician team rounded with the first nurse on all 3 of a nurse’s patients, and then proceeded to the next group of 3 patients with the next nurse, until all patients were seen.
Teams were encouraged to conduct patient- and family-centered rounds exclusively at bedside, except when bedside rounding was thought to be detrimental to a patient (eg, one with delirium). After an intern’s bedside presentation, which included a brief summary and details about overnight events and vital signs, the concerns of the patient, family, and nurse were shared, a focused physical examination performed, relevant data (eg, laboratory test results and imaging studies) reviewed, and the day’s plan formulated. The entire team, including the attending, was expected to have read new patients’ admission notes before rounds. Bedside rounds could thus be focused more on patient assessment and patient/family engagement and less on data transfer.
Several actions were taken to facilitate these changes. Residents, attendings, nurses, and other interdisciplinary team members participated in a series of focus groups and conferences to define workflows and share best practices for patient- and family-centered bedside rounds. Tips on bedside rounding were included in a general medicine rotation guidebook made available to residents and attendings. At the beginning of each post-intervention general medicine rotation, attendings and residents attended brief orientation sessions to review the new daily schedule, have interdisciplinary huddles, and share expectations for patient- and family-centered bedside rounds. On the general medicine units, new medical directors were hired to partner with existing nursing directors to support adoption of the workflows. Last, an interdisciplinary leadership team was formed to support the care redesign efforts. This team started meeting every 2 weeks.
Study Design
We used a pre–post analysis to study the effects of care redesign. Analysis was performed at the same time of year for 2 consecutive years to control for the stage of training and experience of the housestaff. TMA was performed by trained medical students using computer tablets linked to a customized Microsoft Access database form (Redmond, Washington). The form and the database were designed with specific buttons that, when pressed, recorded the time of particular events, such as the coming and going of each participant, the location of rounds, and the beginning and the end of rounding encounters with a patient. One research assistant using an Access entry form was able to dynamically track all events in real time, as they occurred. We collected data on 4 teams at baseline and 5 teams after the intervention. Each of the 4 baseline teams was followed for 4 consecutive weekdays—16 rounds total, April-June 2013—to capture the 4-day call cycle. Each of the 5 post-intervention teams was followed for 5 consecutive weekdays—25 rounds total, April–June 2014—to capture the 5-day cycle. (Because of technical difficulties, data from 1 rounding session were not captured.) For inclusion in the statistical analyses, TMA captured 166 on-service patients before the intervention and 304 afterward. Off-service patients, those with an attending other than the team attending, were excluded because their rounds were conducted separately.
We examined 2 primary outcomes, the proportion of time each clinical team member was present on rounds and the proportion of bedside rounding time. Secondary outcomes were round duration, rounding time per patient, and total non-patient time per rounding session (total rounding time minus total patient time).
Statistical Analysis
TMA data were organized in an Access database and analyzed with SAS Version 9.3 (SAS Institute, Cary, North Carolina). We analyzed the data by round session as well as by patient.
Data are presented as means with standard deviations, medians with interquartile ranges, and proportions, as appropriate. For analyses by round session, we used unadjusted linear regression; for patient-level analyses, we used general estimating equations to adjust for clustering of patients within each session; for nurse presence during any part of a round by patient, we used a χ2 test. Total non-patient time per round session was compared with use of patient-clustered general estimating equations using a γ distribution to account for the non-normality of the data.
Table 1
RESULTS
Patient and Care Team Characteristics
Over the first year of the initiative, 85% of a team’s patients were on their assigned unit, and 87% of a unit’s patients were with the assigned team. Census numbers were 10.4 patients per general medicine team in April-June 2013 and 12.7 patients per team in April-June 2014, a 22% increase after care redesign. There were no statistically significant differences in patient characteristics, including age, sex, race, language, admission source, and comorbidity measure (Elixhauser score), between the pre-intervention and post-intervention study periods, except for a slightly higher proportion of patients admitted from home and fewer patients admitted directly from clinic (Table 1).
Figure 1
Primary Outcomes
Mean proportion of time the nurse was present on rounds per round session increased significantly (P < 0.001), from 24.1% to 67.8% (Figure 1A, Table 2). For individual patient encounters, the increased overall nursing presence was attributable to having more nurses on rounds and having nurses present for a larger proportion of individual rounding encounters (Figure 1B, Table 2). Nurses were present for at least some part of rounds for 53% of patients before the intervention and 93% afterward (P < 0.001). Mean proportion of round time by each of the 2 interns on each team decreased from 59.6% to 49.6% (P = 0.007).
Total bedside rounding time increased significantly (P < 0.001), from 39.9% before the intervention to 55.8% afterward (Table 2). Meanwhile, percentage of rounding time spent on the unit but outside patient rooms decreased significantly (P = 0.004), from 55.2% to 42.2%, as did rounding time on a unit completely different from the patient’s (4.9% before intervention, 2.0% afterward; P = 0.03). Again, patient-level results were similar (Figure 2, Table 2), but the decreased time spent on the unit, outside the patient rooms, was not significant.
Table 2
Secondary Outcomes
Total rounding time decreased significantly, from a mean of 182 minutes (3.0 hours) at baseline to a mean of 146 minutes (2.4 hours) after the intervention, despite the higher post-intervention census. (When adjusted for patient census, the difference increased from 35.5 to 53.8 minutes; Table 2.) Mean rounding time per patient decreased significantly, from 14.7 minutes at baseline to 10.5 minutes after the intervention. For newly admitted patients, mean rounding time per patient decreased from 30.0 minutes before implementation to 16.3 minutes afterward. Mean rounding time also decreased, though much less, for subsequent-day patients (Table 2). For both new and existing patients, the decrease in rounding time largely was a reduction in time spent rounding outside patient rooms, with minimal impact on bedside time (Table 2). Mean time nurses were present during a patient’s rounds increased significantly, from 4.5 to 8.0 minutes (Table 2). Total nurse rounding time increased from 45.1 minutes per session to 98.8 minutes. Rounding time not related to patient discussion or evaluation decreased from 22.7 minutes per session to 13.3 minutes (P = 0.003).
Figure 2
DISCUSSION
TMA of our care redesign initiative showed that this multipronged intervention, which included team regionalization, encouragement of bedside rounding with nurses, call structure changes, and attendings’ reading of admission notes before rounds, resulted in an increased proportion of rounding time spent with patients and an increased proportion of time nurses were present on rounds. Secondarily, round duration decreased even as patient census increased.
Regionalized teams have been found to improve interdisciplinary communication.1 The present study elaborates on that finding by demonstrating a dramatic increase in nursing presence on rounds, likely resulting from the unit’s use of rounding schedules and nurses’ prioritization of rounding orders, both of which were made possible by geographic co-localization. Other research has noted that one of the most significant barriers to interdisciplinary rounds is difficulty coordinating the start times of physician/nurse bedside rounding encounters. The system we have studied directly addresses this difficulty.9 Of note, nursing presence on rounds is necessary but not sufficient for true physician–nurse collaboration and effective communication,1 as reflected in a separate study of the intervention showing no significant difference in the concordance of the patient care plan between nurses and physicians before and after regionalization.12 Additional interventions may be needed to ensure that communication during bedside rounds is effective.
Our regionalized teams spent a significantly higher proportion of rounding time bedside, likely because of a cultural shift in expectations and the increased convenience of seeing patients on the team’s unit. Nevertheless, bedside time was not 100%. Structural barriers (eg, patients off-unit for dialysis) and cultural barriers likely contributed to the less than full adoption of bedside rounding. As described previously, cultural barriers to bedside rounding include trainees’ anxiety about being questioned in front of patients, the desire to freely exchange academic ideas in a conference room, and attendings’ doubts about their bedside teaching ability.1,9,13 Bedside rounds provide an important opportunity to apply the principles of patient- and family-centered care, including promotion of dignity and respect, information sharing, and collaboration. Thus, overcoming the concerns of housestaff and attendings and helping them feel prepared for bedside rounds can benefit the patient experience. More attention should be given to these practices as these types of interventions are implemented at Brigham and Women’s Hospital and elsewhere.1,13-15
Another primary concern about interdisciplinary bedside rounding is the perception that it takes more time.9 Therefore, it was important for us to measure round duration as a balancing measure to be considered for our intervention. Fortunately, we found round duration decreased with regionalization and encouragement of bedside rounding. This decrease was driven largely by a significant decrease in mean rounding time per new patient, which may be attributable at least in part to setting expectations that attendings and residents will read admission notes before rounds and that interns will summarize rather than recount information from admission notes. However, we also found rounding time decreases for subsequent-day patients, suggesting an underlying time savings. Spending a larger proportion of time bedside may therefore result in more efficient rounds. Bedside presentations can reduce redundancies, such as discussing a patient’s case outside his or her room and subsequently walking in and going over much of the same information with the patient. Our model de-emphasizes data transfer in favor of discussion of care plans. There was also a decrease in non-patient time, likely reflecting reduced transit time for regionalized teams. This decrease aligns with a recent finding that bedside rounding was at least as efficient as rounding outside the room.16
Of note, though a larger percentage of time was spent bedside after implementation of the care redesign, the absolute amount of bedside time did not change significantly. Our data showed that, even with shorter rounds, the same amount of absolute time can be spent bedside, face to face with the patient, by increasing the proportion of bedside rounding time. In other words, teams on average did not spend more time with patients, though the content and the structure of those encounters may have changed. This finding may be attributable to eliminating redundancy, forgoing the outside-the-room discussion, and thus the largest time reductions were realized there. In addition, teams incompletely adopted beside rounds, as reflected in the data. We expect that, with more complete adoption, an even larger proportion of time will be spent bedside, and absolute time bedside might increase as a result.
An unexpected result of the care redesign was that interns’ proportion of rounding time decreased after the intervention. This decrease most likely is attributable to interns’ being less likely to participate in rounds for a co-intern’s patient, and to their staying outside that patient’s room to give themselves more time to advance the care of their own patients. Before the intervention, when more rounding time was spent outside patient rooms, interns were more likely to join rounds for their co-intern’s patients because they could easily break away, as needed, to continue care of their own patients. The resident is now encouraged to use the morning huddle to identify which patients likely have the most educational value, and both interns are expected to join the bedside rounds for these patients.
This study had a few limitations. First, the pre–post design made it difficult to exclude the possibility that other temporal changes may have affected outcomes, though we did account for time-of-year effects by aligning our data-collection phases. In addition, the authors, including the director of the general medical service, are unaware of any co-interventions during the study period. Second, the multipronged intervention included care team regionalization, encouragement of bedside rounding with nurses, call structure changes (from 4 days to daily admitting), and attendings’ reading of admission notes before rounds. Thus, parsing which component(s) contributed to the results was difficult, though all the changes instituted likely were necessary for system redesign. For example, regionalization of clinicians to unit-based teams was made possible by switching to a daily admitting system.
Time that team members spent preparing for rounds was not recorded before or after the intervention. Thus, the decrease in total rounding time could have been accompanied by an increase in time spent preparing for rounds. However, admission notes were available in our electronic medical record before and after the intervention, and most residents and attendings were already reading them pre-intervention. After the intervention, pre-round note reading was more clearly defined as an expectation, and we were able to set the expectation that interns should use their presentations to summarize rather than recount information. In addition, in the post-intervention period, we did not include time spent preparing rounding orders; as already noted, however, preparation took only 5 minutes per day. Also, we did not analyze the content or the quality of the discussion on rounds, but simply recorded who was present where and when. Regarding the effect of the intervention on patient care, results were mixed. As reported in 2016, we saw no difference in frequency of adverse events with this intervention.12 However, a more sensitive measure of adverse events—used in a study on handoffs—showed our regionalization efforts had an additive effect on reducing overnight adverse events.17Researchers should now focus on the effects of care redesign on clinical outcomes, interdisciplinary care team communication, patient engagement and satisfaction, provider opinions of communication, workflow, patient care, and housestaff education. Our methodology can be used as a model to link structure, process, and outcome related to rounds and thereby better understand how best to optimize patient care and efficiency. Additional studies are needed to analyze the content of rounds and their association with patient and educational outcomes. Last, it will be important to conduct a study to see if the effects we have identified can be sustained. Such a study is already under way.
In conclusion, creating regionalized care teams and encouraging focused bedside rounds increased the proportion of bedside time and the presence of nurses on rounds. Rounds were shorter despite higher patient census. TMA revealed that regionalized care teams and bedside rounding at a large academic hospital are feasible, and are useful in establishing the necessary structures for increasing physician–nurse and provider–patient interactions.
Acknowledgments
The authors acknowledge Dr. Stan Ashley, Dr. Jacqueline Somerville, and Sheila Harris for their support of the regionalization initiative.
Disclosures
Dr. Schnipper received funding from Sanofi-aventis to conduct an investigator-initiated study to implement and evaluate a multi-faceted intervention to improve transitions of care in patients discharged home on insulin. The study was also supported by funding from the Marshall A. Wolf Medical Education Fund, Brigham and Women’s Hospital, and Dr. Stan Ashley, Chief Medical Officer, Brigham and Women’s Hospital. Some of the content of this article was orally presented at the annual meeting of the Society of Hospital Medicine; March 29-April 1, 2015; National Harbor, MD.
References
1. Crumlish CM, Yialamas MA, McMahon GT. Quantification of bedside teaching by an academic hospitalist group. J Hosp Med. 2009;4(5):304-307. PubMed 2. Gonzalo JD, Masters PA, Simons RJ, Chuang CH. Attending rounds and bedside case presentations: medical student and medicine resident experiences and attitudes. Teach Learn Med. 2009;21(2):105-110. PubMed 3. Elliot DL, Hickam DH. Attending rounds on in-patient units: differences between medical and non-medical services. Med Educ. 1993;27(6):503-508. PubMed 4. Payson HE, Barchas JD. A time study of medical teaching rounds. N Engl J Med. 1965;273(27):1468-1471. PubMed 5. Tremonti LP, Biddle WB. Teaching behaviors of residents and faculty members. J Med Educ. 1982;57(11):854-859. PubMed 6. Miller M, Johnson B, Greene HL, Baier M, Nowlin S. An observational study of attending rounds. J Gen Intern Med. 1992;7(6):646-648. PubMed 7. Collins GF, Cassie JM, Daggett CJ. The role of the attending physician in clinical training. J Med Educ. 1978;53(5):429-431. PubMed 8. Ward DR, Ghali WA, Graham A, Lemaire JB. A real-time locating system observes physician time-motion patterns during walk-rounds: a pilot study. BMC Med Educ. 2014;14:37. PubMed 9. Gonzalo JD, Kuperman E, Lehman E, Haidet P. Bedside interprofessional rounds: perceptions of benefits and barriers by internal medicine nursing staff, attending physicians, and housestaff physicians. J Hosp Med. 2014;9(10):646-651. PubMed 10. Stickrath C, Noble M, Prochazka A, et al. Attending rounds in the current era: what is and is not happening. JAMA Intern Med. 2013;173(12):1084-1089. PubMed 11. Boxer R, Vitale M, Gershanik EF, et al. 5th time’s a charm: creation of unit-based care teams in a high occupancy hospital [abstract]. J Hosp Med. 2015;10(suppl 2). 12. Mueller SK, Schnipper JL, Giannelli K, Roy CL, Boxer R. Impact of regionalized care on concordance of plan and preventable adverse events on general medicine services. J Hosp Med. 2016;11(9):620-627. PubMed 13. Chauke HL, Pattinson RC. Ward rounds—bedside or conference room? S Afr Med J. 2006;96(5):398-400. PubMed 14. Wang-Cheng RM, Barnas GP, Sigmann P, Riendl PA, Young MJ. Bedside case presentations: why patients like them but learners don’t. J Gen Intern Med. 1989;4(4):284-287. PubMed 15. Lehmann LS, Brancati FL, Chen MC, Roter D, Dobs AS. The effect of bedside case presentations on patients’ perceptions of their medical care. N Engl J Med. 1997;336(16):1150-1155. PubMed 16. Gonzalo JD, Chuang CH, Huang G, Smith C. The return of bedside rounds: an educational intervention. J Gen Intern Med. 2010;25(8):792-798. PubMed 17. Mueller SK, Yoon C, Schnipper JL. Association of a web-based handoff tool with rates of medical errors. JAMA Intern Med. 2016;176(9):1400-1402. PubMed
Attending rounds at academic medical centers are often disconnected from patients and non-physician care team members. Time spent bedside is consistently less than one third of total rounding time, with observational studies reporting a range of 9% to 33% over the past several decades.1-8 Rounds are often conducted outside patient rooms, denying patients, families, and nurses the opportunity to participate and offer valuable insights. Lack of bedside rounds thus limits patient and family engagement, patient input into the care plan, teaching of the physical examination, and communication and collaboration with nurses. In one study, physicians and nurses on rounds engaged in interprofessional communication in only 12% of patient cases.1 Studies have found interdisciplinary bedside rounds have several benefits, including subjectively improved communication and teamwork between physicians and nurses; increased patient satisfaction, including feeling more cared for by the medical team; and decreased length of stay and costs of care.2-10
However, there are many barriers to conducting interdisciplinary bedside rounds at large academic medical centers. Patients cared for by a single medical team are often geographically dispersed to several nursing units, and nurses are unable to predict when physicians will round on their patients. This situation limits nursing involvement on rounds and keeps doctors and nurses isolated from each other.2 Regionalization of care teams reduces this fragmentation by facilitating more interaction among doctors, patients, families, and nursing staff.
There are few data on how regionalized patients and interdisciplinary bedside rounds affect rounding time and the nature of rounds. This information is needed to understand how these structural changes mediate their effects, whether other steps are required to optimize outcomes, and how to maximize efficiency. We used time-motion analysis (TMA) to investigate how regionalization of medical teams, encouragement of bedside rounding, and systematic inclusion of nurses on ward rounds affect amount of time spent with patients, nursing presence on rounds, and total rounding time.
METHODS
Setting
This prospective interventional study, approved by the Institutional Review Board of Partners HealthCare, was conducted on the general medical wards at Brigham and Women’s Hospital, an academic 793-bed tertiary-care center in Boston, Massachusetts. Housestaff teams consist of 1 attending, 1 resident, and 2 interns with or without a medical student. Before June 20, 2013, daily rounds on medical inpatients were conducted largely on the patient unit but outside patient rooms. After completing most of a rounding discussion outside a patient’s room, the team might walk in to examine or speak with the patient. A typical medical team had patients dispersed over 7 medical units on average, and over as many as 13. As nurses were unit based, they did not consistently participate in rounds.
Intervention
In June 2013, as part of a general medical service care redesign initiative, the general medical teams were regionalized to specific inpatient units. The goal was to have teams admit patients predominantly to the team’s designated unit and to have all patients on a unit be cared for by the unit’s assigned team as often as possible, with an 85% goal for both. Toward those ends, the admitting structure was changed from a traditional 4-day call cycle to daily admitting for all teams, based on each unit’s bed availability.11
Teams were also expected to conduct rounds with nurses, and a system for facilitating these rounds was established. As physician and nurse care teams were now geographically co-located, it became possible for residents and nurses to check a rounding sheet for the planned patient rounding order, which had been set by the resident and nurse-in-charge before rounds. No more than about 5 minutes was needed to prepare each day’s order. The rounding sheet prioritized sick patients, newly admitted patients, and planned morning discharges, but patients were also always grouped by nurse. For example, the physician team rounded with the first nurse on all 3 of a nurse’s patients, and then proceeded to the next group of 3 patients with the next nurse, until all patients were seen.
Teams were encouraged to conduct patient- and family-centered rounds exclusively at bedside, except when bedside rounding was thought to be detrimental to a patient (eg, one with delirium). After an intern’s bedside presentation, which included a brief summary and details about overnight events and vital signs, the concerns of the patient, family, and nurse were shared, a focused physical examination performed, relevant data (eg, laboratory test results and imaging studies) reviewed, and the day’s plan formulated. The entire team, including the attending, was expected to have read new patients’ admission notes before rounds. Bedside rounds could thus be focused more on patient assessment and patient/family engagement and less on data transfer.
Several actions were taken to facilitate these changes. Residents, attendings, nurses, and other interdisciplinary team members participated in a series of focus groups and conferences to define workflows and share best practices for patient- and family-centered bedside rounds. Tips on bedside rounding were included in a general medicine rotation guidebook made available to residents and attendings. At the beginning of each post-intervention general medicine rotation, attendings and residents attended brief orientation sessions to review the new daily schedule, have interdisciplinary huddles, and share expectations for patient- and family-centered bedside rounds. On the general medicine units, new medical directors were hired to partner with existing nursing directors to support adoption of the workflows. Last, an interdisciplinary leadership team was formed to support the care redesign efforts. This team started meeting every 2 weeks.
Study Design
We used a pre–post analysis to study the effects of care redesign. Analysis was performed at the same time of year for 2 consecutive years to control for the stage of training and experience of the housestaff. TMA was performed by trained medical students using computer tablets linked to a customized Microsoft Access database form (Redmond, Washington). The form and the database were designed with specific buttons that, when pressed, recorded the time of particular events, such as the coming and going of each participant, the location of rounds, and the beginning and the end of rounding encounters with a patient. One research assistant using an Access entry form was able to dynamically track all events in real time, as they occurred. We collected data on 4 teams at baseline and 5 teams after the intervention. Each of the 4 baseline teams was followed for 4 consecutive weekdays—16 rounds total, April-June 2013—to capture the 4-day call cycle. Each of the 5 post-intervention teams was followed for 5 consecutive weekdays—25 rounds total, April–June 2014—to capture the 5-day cycle. (Because of technical difficulties, data from 1 rounding session were not captured.) For inclusion in the statistical analyses, TMA captured 166 on-service patients before the intervention and 304 afterward. Off-service patients, those with an attending other than the team attending, were excluded because their rounds were conducted separately.
We examined 2 primary outcomes, the proportion of time each clinical team member was present on rounds and the proportion of bedside rounding time. Secondary outcomes were round duration, rounding time per patient, and total non-patient time per rounding session (total rounding time minus total patient time).
Statistical Analysis
TMA data were organized in an Access database and analyzed with SAS Version 9.3 (SAS Institute, Cary, North Carolina). We analyzed the data by round session as well as by patient.
Data are presented as means with standard deviations, medians with interquartile ranges, and proportions, as appropriate. For analyses by round session, we used unadjusted linear regression; for patient-level analyses, we used general estimating equations to adjust for clustering of patients within each session; for nurse presence during any part of a round by patient, we used a χ2 test. Total non-patient time per round session was compared with use of patient-clustered general estimating equations using a γ distribution to account for the non-normality of the data.
Table 1
RESULTS
Patient and Care Team Characteristics
Over the first year of the initiative, 85% of a team’s patients were on their assigned unit, and 87% of a unit’s patients were with the assigned team. Census numbers were 10.4 patients per general medicine team in April-June 2013 and 12.7 patients per team in April-June 2014, a 22% increase after care redesign. There were no statistically significant differences in patient characteristics, including age, sex, race, language, admission source, and comorbidity measure (Elixhauser score), between the pre-intervention and post-intervention study periods, except for a slightly higher proportion of patients admitted from home and fewer patients admitted directly from clinic (Table 1).
Figure 1
Primary Outcomes
Mean proportion of time the nurse was present on rounds per round session increased significantly (P < 0.001), from 24.1% to 67.8% (Figure 1A, Table 2). For individual patient encounters, the increased overall nursing presence was attributable to having more nurses on rounds and having nurses present for a larger proportion of individual rounding encounters (Figure 1B, Table 2). Nurses were present for at least some part of rounds for 53% of patients before the intervention and 93% afterward (P < 0.001). Mean proportion of round time by each of the 2 interns on each team decreased from 59.6% to 49.6% (P = 0.007).
Total bedside rounding time increased significantly (P < 0.001), from 39.9% before the intervention to 55.8% afterward (Table 2). Meanwhile, percentage of rounding time spent on the unit but outside patient rooms decreased significantly (P = 0.004), from 55.2% to 42.2%, as did rounding time on a unit completely different from the patient’s (4.9% before intervention, 2.0% afterward; P = 0.03). Again, patient-level results were similar (Figure 2, Table 2), but the decreased time spent on the unit, outside the patient rooms, was not significant.
Table 2
Secondary Outcomes
Total rounding time decreased significantly, from a mean of 182 minutes (3.0 hours) at baseline to a mean of 146 minutes (2.4 hours) after the intervention, despite the higher post-intervention census. (When adjusted for patient census, the difference increased from 35.5 to 53.8 minutes; Table 2.) Mean rounding time per patient decreased significantly, from 14.7 minutes at baseline to 10.5 minutes after the intervention. For newly admitted patients, mean rounding time per patient decreased from 30.0 minutes before implementation to 16.3 minutes afterward. Mean rounding time also decreased, though much less, for subsequent-day patients (Table 2). For both new and existing patients, the decrease in rounding time largely was a reduction in time spent rounding outside patient rooms, with minimal impact on bedside time (Table 2). Mean time nurses were present during a patient’s rounds increased significantly, from 4.5 to 8.0 minutes (Table 2). Total nurse rounding time increased from 45.1 minutes per session to 98.8 minutes. Rounding time not related to patient discussion or evaluation decreased from 22.7 minutes per session to 13.3 minutes (P = 0.003).
Figure 2
DISCUSSION
TMA of our care redesign initiative showed that this multipronged intervention, which included team regionalization, encouragement of bedside rounding with nurses, call structure changes, and attendings’ reading of admission notes before rounds, resulted in an increased proportion of rounding time spent with patients and an increased proportion of time nurses were present on rounds. Secondarily, round duration decreased even as patient census increased.
Regionalized teams have been found to improve interdisciplinary communication.1 The present study elaborates on that finding by demonstrating a dramatic increase in nursing presence on rounds, likely resulting from the unit’s use of rounding schedules and nurses’ prioritization of rounding orders, both of which were made possible by geographic co-localization. Other research has noted that one of the most significant barriers to interdisciplinary rounds is difficulty coordinating the start times of physician/nurse bedside rounding encounters. The system we have studied directly addresses this difficulty.9 Of note, nursing presence on rounds is necessary but not sufficient for true physician–nurse collaboration and effective communication,1 as reflected in a separate study of the intervention showing no significant difference in the concordance of the patient care plan between nurses and physicians before and after regionalization.12 Additional interventions may be needed to ensure that communication during bedside rounds is effective.
Our regionalized teams spent a significantly higher proportion of rounding time bedside, likely because of a cultural shift in expectations and the increased convenience of seeing patients on the team’s unit. Nevertheless, bedside time was not 100%. Structural barriers (eg, patients off-unit for dialysis) and cultural barriers likely contributed to the less than full adoption of bedside rounding. As described previously, cultural barriers to bedside rounding include trainees’ anxiety about being questioned in front of patients, the desire to freely exchange academic ideas in a conference room, and attendings’ doubts about their bedside teaching ability.1,9,13 Bedside rounds provide an important opportunity to apply the principles of patient- and family-centered care, including promotion of dignity and respect, information sharing, and collaboration. Thus, overcoming the concerns of housestaff and attendings and helping them feel prepared for bedside rounds can benefit the patient experience. More attention should be given to these practices as these types of interventions are implemented at Brigham and Women’s Hospital and elsewhere.1,13-15
Another primary concern about interdisciplinary bedside rounding is the perception that it takes more time.9 Therefore, it was important for us to measure round duration as a balancing measure to be considered for our intervention. Fortunately, we found round duration decreased with regionalization and encouragement of bedside rounding. This decrease was driven largely by a significant decrease in mean rounding time per new patient, which may be attributable at least in part to setting expectations that attendings and residents will read admission notes before rounds and that interns will summarize rather than recount information from admission notes. However, we also found rounding time decreases for subsequent-day patients, suggesting an underlying time savings. Spending a larger proportion of time bedside may therefore result in more efficient rounds. Bedside presentations can reduce redundancies, such as discussing a patient’s case outside his or her room and subsequently walking in and going over much of the same information with the patient. Our model de-emphasizes data transfer in favor of discussion of care plans. There was also a decrease in non-patient time, likely reflecting reduced transit time for regionalized teams. This decrease aligns with a recent finding that bedside rounding was at least as efficient as rounding outside the room.16
Of note, though a larger percentage of time was spent bedside after implementation of the care redesign, the absolute amount of bedside time did not change significantly. Our data showed that, even with shorter rounds, the same amount of absolute time can be spent bedside, face to face with the patient, by increasing the proportion of bedside rounding time. In other words, teams on average did not spend more time with patients, though the content and the structure of those encounters may have changed. This finding may be attributable to eliminating redundancy, forgoing the outside-the-room discussion, and thus the largest time reductions were realized there. In addition, teams incompletely adopted beside rounds, as reflected in the data. We expect that, with more complete adoption, an even larger proportion of time will be spent bedside, and absolute time bedside might increase as a result.
An unexpected result of the care redesign was that interns’ proportion of rounding time decreased after the intervention. This decrease most likely is attributable to interns’ being less likely to participate in rounds for a co-intern’s patient, and to their staying outside that patient’s room to give themselves more time to advance the care of their own patients. Before the intervention, when more rounding time was spent outside patient rooms, interns were more likely to join rounds for their co-intern’s patients because they could easily break away, as needed, to continue care of their own patients. The resident is now encouraged to use the morning huddle to identify which patients likely have the most educational value, and both interns are expected to join the bedside rounds for these patients.
This study had a few limitations. First, the pre–post design made it difficult to exclude the possibility that other temporal changes may have affected outcomes, though we did account for time-of-year effects by aligning our data-collection phases. In addition, the authors, including the director of the general medical service, are unaware of any co-interventions during the study period. Second, the multipronged intervention included care team regionalization, encouragement of bedside rounding with nurses, call structure changes (from 4 days to daily admitting), and attendings’ reading of admission notes before rounds. Thus, parsing which component(s) contributed to the results was difficult, though all the changes instituted likely were necessary for system redesign. For example, regionalization of clinicians to unit-based teams was made possible by switching to a daily admitting system.
Time that team members spent preparing for rounds was not recorded before or after the intervention. Thus, the decrease in total rounding time could have been accompanied by an increase in time spent preparing for rounds. However, admission notes were available in our electronic medical record before and after the intervention, and most residents and attendings were already reading them pre-intervention. After the intervention, pre-round note reading was more clearly defined as an expectation, and we were able to set the expectation that interns should use their presentations to summarize rather than recount information. In addition, in the post-intervention period, we did not include time spent preparing rounding orders; as already noted, however, preparation took only 5 minutes per day. Also, we did not analyze the content or the quality of the discussion on rounds, but simply recorded who was present where and when. Regarding the effect of the intervention on patient care, results were mixed. As reported in 2016, we saw no difference in frequency of adverse events with this intervention.12 However, a more sensitive measure of adverse events—used in a study on handoffs—showed our regionalization efforts had an additive effect on reducing overnight adverse events.17Researchers should now focus on the effects of care redesign on clinical outcomes, interdisciplinary care team communication, patient engagement and satisfaction, provider opinions of communication, workflow, patient care, and housestaff education. Our methodology can be used as a model to link structure, process, and outcome related to rounds and thereby better understand how best to optimize patient care and efficiency. Additional studies are needed to analyze the content of rounds and their association with patient and educational outcomes. Last, it will be important to conduct a study to see if the effects we have identified can be sustained. Such a study is already under way.
In conclusion, creating regionalized care teams and encouraging focused bedside rounds increased the proportion of bedside time and the presence of nurses on rounds. Rounds were shorter despite higher patient census. TMA revealed that regionalized care teams and bedside rounding at a large academic hospital are feasible, and are useful in establishing the necessary structures for increasing physician–nurse and provider–patient interactions.
Acknowledgments
The authors acknowledge Dr. Stan Ashley, Dr. Jacqueline Somerville, and Sheila Harris for their support of the regionalization initiative.
Disclosures
Dr. Schnipper received funding from Sanofi-aventis to conduct an investigator-initiated study to implement and evaluate a multi-faceted intervention to improve transitions of care in patients discharged home on insulin. The study was also supported by funding from the Marshall A. Wolf Medical Education Fund, Brigham and Women’s Hospital, and Dr. Stan Ashley, Chief Medical Officer, Brigham and Women’s Hospital. Some of the content of this article was orally presented at the annual meeting of the Society of Hospital Medicine; March 29-April 1, 2015; National Harbor, MD.
Attending rounds at academic medical centers are often disconnected from patients and non-physician care team members. Time spent bedside is consistently less than one third of total rounding time, with observational studies reporting a range of 9% to 33% over the past several decades.1-8 Rounds are often conducted outside patient rooms, denying patients, families, and nurses the opportunity to participate and offer valuable insights. Lack of bedside rounds thus limits patient and family engagement, patient input into the care plan, teaching of the physical examination, and communication and collaboration with nurses. In one study, physicians and nurses on rounds engaged in interprofessional communication in only 12% of patient cases.1 Studies have found interdisciplinary bedside rounds have several benefits, including subjectively improved communication and teamwork between physicians and nurses; increased patient satisfaction, including feeling more cared for by the medical team; and decreased length of stay and costs of care.2-10
However, there are many barriers to conducting interdisciplinary bedside rounds at large academic medical centers. Patients cared for by a single medical team are often geographically dispersed to several nursing units, and nurses are unable to predict when physicians will round on their patients. This situation limits nursing involvement on rounds and keeps doctors and nurses isolated from each other.2 Regionalization of care teams reduces this fragmentation by facilitating more interaction among doctors, patients, families, and nursing staff.
There are few data on how regionalized patients and interdisciplinary bedside rounds affect rounding time and the nature of rounds. This information is needed to understand how these structural changes mediate their effects, whether other steps are required to optimize outcomes, and how to maximize efficiency. We used time-motion analysis (TMA) to investigate how regionalization of medical teams, encouragement of bedside rounding, and systematic inclusion of nurses on ward rounds affect amount of time spent with patients, nursing presence on rounds, and total rounding time.
METHODS
Setting
This prospective interventional study, approved by the Institutional Review Board of Partners HealthCare, was conducted on the general medical wards at Brigham and Women’s Hospital, an academic 793-bed tertiary-care center in Boston, Massachusetts. Housestaff teams consist of 1 attending, 1 resident, and 2 interns with or without a medical student. Before June 20, 2013, daily rounds on medical inpatients were conducted largely on the patient unit but outside patient rooms. After completing most of a rounding discussion outside a patient’s room, the team might walk in to examine or speak with the patient. A typical medical team had patients dispersed over 7 medical units on average, and over as many as 13. As nurses were unit based, they did not consistently participate in rounds.
Intervention
In June 2013, as part of a general medical service care redesign initiative, the general medical teams were regionalized to specific inpatient units. The goal was to have teams admit patients predominantly to the team’s designated unit and to have all patients on a unit be cared for by the unit’s assigned team as often as possible, with an 85% goal for both. Toward those ends, the admitting structure was changed from a traditional 4-day call cycle to daily admitting for all teams, based on each unit’s bed availability.11
Teams were also expected to conduct rounds with nurses, and a system for facilitating these rounds was established. As physician and nurse care teams were now geographically co-located, it became possible for residents and nurses to check a rounding sheet for the planned patient rounding order, which had been set by the resident and nurse-in-charge before rounds. No more than about 5 minutes was needed to prepare each day’s order. The rounding sheet prioritized sick patients, newly admitted patients, and planned morning discharges, but patients were also always grouped by nurse. For example, the physician team rounded with the first nurse on all 3 of a nurse’s patients, and then proceeded to the next group of 3 patients with the next nurse, until all patients were seen.
Teams were encouraged to conduct patient- and family-centered rounds exclusively at bedside, except when bedside rounding was thought to be detrimental to a patient (eg, one with delirium). After an intern’s bedside presentation, which included a brief summary and details about overnight events and vital signs, the concerns of the patient, family, and nurse were shared, a focused physical examination performed, relevant data (eg, laboratory test results and imaging studies) reviewed, and the day’s plan formulated. The entire team, including the attending, was expected to have read new patients’ admission notes before rounds. Bedside rounds could thus be focused more on patient assessment and patient/family engagement and less on data transfer.
Several actions were taken to facilitate these changes. Residents, attendings, nurses, and other interdisciplinary team members participated in a series of focus groups and conferences to define workflows and share best practices for patient- and family-centered bedside rounds. Tips on bedside rounding were included in a general medicine rotation guidebook made available to residents and attendings. At the beginning of each post-intervention general medicine rotation, attendings and residents attended brief orientation sessions to review the new daily schedule, have interdisciplinary huddles, and share expectations for patient- and family-centered bedside rounds. On the general medicine units, new medical directors were hired to partner with existing nursing directors to support adoption of the workflows. Last, an interdisciplinary leadership team was formed to support the care redesign efforts. This team started meeting every 2 weeks.
Study Design
We used a pre–post analysis to study the effects of care redesign. Analysis was performed at the same time of year for 2 consecutive years to control for the stage of training and experience of the housestaff. TMA was performed by trained medical students using computer tablets linked to a customized Microsoft Access database form (Redmond, Washington). The form and the database were designed with specific buttons that, when pressed, recorded the time of particular events, such as the coming and going of each participant, the location of rounds, and the beginning and the end of rounding encounters with a patient. One research assistant using an Access entry form was able to dynamically track all events in real time, as they occurred. We collected data on 4 teams at baseline and 5 teams after the intervention. Each of the 4 baseline teams was followed for 4 consecutive weekdays—16 rounds total, April-June 2013—to capture the 4-day call cycle. Each of the 5 post-intervention teams was followed for 5 consecutive weekdays—25 rounds total, April–June 2014—to capture the 5-day cycle. (Because of technical difficulties, data from 1 rounding session were not captured.) For inclusion in the statistical analyses, TMA captured 166 on-service patients before the intervention and 304 afterward. Off-service patients, those with an attending other than the team attending, were excluded because their rounds were conducted separately.
We examined 2 primary outcomes, the proportion of time each clinical team member was present on rounds and the proportion of bedside rounding time. Secondary outcomes were round duration, rounding time per patient, and total non-patient time per rounding session (total rounding time minus total patient time).
Statistical Analysis
TMA data were organized in an Access database and analyzed with SAS Version 9.3 (SAS Institute, Cary, North Carolina). We analyzed the data by round session as well as by patient.
Data are presented as means with standard deviations, medians with interquartile ranges, and proportions, as appropriate. For analyses by round session, we used unadjusted linear regression; for patient-level analyses, we used general estimating equations to adjust for clustering of patients within each session; for nurse presence during any part of a round by patient, we used a χ2 test. Total non-patient time per round session was compared with use of patient-clustered general estimating equations using a γ distribution to account for the non-normality of the data.
Table 1
RESULTS
Patient and Care Team Characteristics
Over the first year of the initiative, 85% of a team’s patients were on their assigned unit, and 87% of a unit’s patients were with the assigned team. Census numbers were 10.4 patients per general medicine team in April-June 2013 and 12.7 patients per team in April-June 2014, a 22% increase after care redesign. There were no statistically significant differences in patient characteristics, including age, sex, race, language, admission source, and comorbidity measure (Elixhauser score), between the pre-intervention and post-intervention study periods, except for a slightly higher proportion of patients admitted from home and fewer patients admitted directly from clinic (Table 1).
Figure 1
Primary Outcomes
Mean proportion of time the nurse was present on rounds per round session increased significantly (P < 0.001), from 24.1% to 67.8% (Figure 1A, Table 2). For individual patient encounters, the increased overall nursing presence was attributable to having more nurses on rounds and having nurses present for a larger proportion of individual rounding encounters (Figure 1B, Table 2). Nurses were present for at least some part of rounds for 53% of patients before the intervention and 93% afterward (P < 0.001). Mean proportion of round time by each of the 2 interns on each team decreased from 59.6% to 49.6% (P = 0.007).
Total bedside rounding time increased significantly (P < 0.001), from 39.9% before the intervention to 55.8% afterward (Table 2). Meanwhile, percentage of rounding time spent on the unit but outside patient rooms decreased significantly (P = 0.004), from 55.2% to 42.2%, as did rounding time on a unit completely different from the patient’s (4.9% before intervention, 2.0% afterward; P = 0.03). Again, patient-level results were similar (Figure 2, Table 2), but the decreased time spent on the unit, outside the patient rooms, was not significant.
Table 2
Secondary Outcomes
Total rounding time decreased significantly, from a mean of 182 minutes (3.0 hours) at baseline to a mean of 146 minutes (2.4 hours) after the intervention, despite the higher post-intervention census. (When adjusted for patient census, the difference increased from 35.5 to 53.8 minutes; Table 2.) Mean rounding time per patient decreased significantly, from 14.7 minutes at baseline to 10.5 minutes after the intervention. For newly admitted patients, mean rounding time per patient decreased from 30.0 minutes before implementation to 16.3 minutes afterward. Mean rounding time also decreased, though much less, for subsequent-day patients (Table 2). For both new and existing patients, the decrease in rounding time largely was a reduction in time spent rounding outside patient rooms, with minimal impact on bedside time (Table 2). Mean time nurses were present during a patient’s rounds increased significantly, from 4.5 to 8.0 minutes (Table 2). Total nurse rounding time increased from 45.1 minutes per session to 98.8 minutes. Rounding time not related to patient discussion or evaluation decreased from 22.7 minutes per session to 13.3 minutes (P = 0.003).
Figure 2
DISCUSSION
TMA of our care redesign initiative showed that this multipronged intervention, which included team regionalization, encouragement of bedside rounding with nurses, call structure changes, and attendings’ reading of admission notes before rounds, resulted in an increased proportion of rounding time spent with patients and an increased proportion of time nurses were present on rounds. Secondarily, round duration decreased even as patient census increased.
Regionalized teams have been found to improve interdisciplinary communication.1 The present study elaborates on that finding by demonstrating a dramatic increase in nursing presence on rounds, likely resulting from the unit’s use of rounding schedules and nurses’ prioritization of rounding orders, both of which were made possible by geographic co-localization. Other research has noted that one of the most significant barriers to interdisciplinary rounds is difficulty coordinating the start times of physician/nurse bedside rounding encounters. The system we have studied directly addresses this difficulty.9 Of note, nursing presence on rounds is necessary but not sufficient for true physician–nurse collaboration and effective communication,1 as reflected in a separate study of the intervention showing no significant difference in the concordance of the patient care plan between nurses and physicians before and after regionalization.12 Additional interventions may be needed to ensure that communication during bedside rounds is effective.
Our regionalized teams spent a significantly higher proportion of rounding time bedside, likely because of a cultural shift in expectations and the increased convenience of seeing patients on the team’s unit. Nevertheless, bedside time was not 100%. Structural barriers (eg, patients off-unit for dialysis) and cultural barriers likely contributed to the less than full adoption of bedside rounding. As described previously, cultural barriers to bedside rounding include trainees’ anxiety about being questioned in front of patients, the desire to freely exchange academic ideas in a conference room, and attendings’ doubts about their bedside teaching ability.1,9,13 Bedside rounds provide an important opportunity to apply the principles of patient- and family-centered care, including promotion of dignity and respect, information sharing, and collaboration. Thus, overcoming the concerns of housestaff and attendings and helping them feel prepared for bedside rounds can benefit the patient experience. More attention should be given to these practices as these types of interventions are implemented at Brigham and Women’s Hospital and elsewhere.1,13-15
Another primary concern about interdisciplinary bedside rounding is the perception that it takes more time.9 Therefore, it was important for us to measure round duration as a balancing measure to be considered for our intervention. Fortunately, we found round duration decreased with regionalization and encouragement of bedside rounding. This decrease was driven largely by a significant decrease in mean rounding time per new patient, which may be attributable at least in part to setting expectations that attendings and residents will read admission notes before rounds and that interns will summarize rather than recount information from admission notes. However, we also found rounding time decreases for subsequent-day patients, suggesting an underlying time savings. Spending a larger proportion of time bedside may therefore result in more efficient rounds. Bedside presentations can reduce redundancies, such as discussing a patient’s case outside his or her room and subsequently walking in and going over much of the same information with the patient. Our model de-emphasizes data transfer in favor of discussion of care plans. There was also a decrease in non-patient time, likely reflecting reduced transit time for regionalized teams. This decrease aligns with a recent finding that bedside rounding was at least as efficient as rounding outside the room.16
Of note, though a larger percentage of time was spent bedside after implementation of the care redesign, the absolute amount of bedside time did not change significantly. Our data showed that, even with shorter rounds, the same amount of absolute time can be spent bedside, face to face with the patient, by increasing the proportion of bedside rounding time. In other words, teams on average did not spend more time with patients, though the content and the structure of those encounters may have changed. This finding may be attributable to eliminating redundancy, forgoing the outside-the-room discussion, and thus the largest time reductions were realized there. In addition, teams incompletely adopted beside rounds, as reflected in the data. We expect that, with more complete adoption, an even larger proportion of time will be spent bedside, and absolute time bedside might increase as a result.
An unexpected result of the care redesign was that interns’ proportion of rounding time decreased after the intervention. This decrease most likely is attributable to interns’ being less likely to participate in rounds for a co-intern’s patient, and to their staying outside that patient’s room to give themselves more time to advance the care of their own patients. Before the intervention, when more rounding time was spent outside patient rooms, interns were more likely to join rounds for their co-intern’s patients because they could easily break away, as needed, to continue care of their own patients. The resident is now encouraged to use the morning huddle to identify which patients likely have the most educational value, and both interns are expected to join the bedside rounds for these patients.
This study had a few limitations. First, the pre–post design made it difficult to exclude the possibility that other temporal changes may have affected outcomes, though we did account for time-of-year effects by aligning our data-collection phases. In addition, the authors, including the director of the general medical service, are unaware of any co-interventions during the study period. Second, the multipronged intervention included care team regionalization, encouragement of bedside rounding with nurses, call structure changes (from 4 days to daily admitting), and attendings’ reading of admission notes before rounds. Thus, parsing which component(s) contributed to the results was difficult, though all the changes instituted likely were necessary for system redesign. For example, regionalization of clinicians to unit-based teams was made possible by switching to a daily admitting system.
Time that team members spent preparing for rounds was not recorded before or after the intervention. Thus, the decrease in total rounding time could have been accompanied by an increase in time spent preparing for rounds. However, admission notes were available in our electronic medical record before and after the intervention, and most residents and attendings were already reading them pre-intervention. After the intervention, pre-round note reading was more clearly defined as an expectation, and we were able to set the expectation that interns should use their presentations to summarize rather than recount information. In addition, in the post-intervention period, we did not include time spent preparing rounding orders; as already noted, however, preparation took only 5 minutes per day. Also, we did not analyze the content or the quality of the discussion on rounds, but simply recorded who was present where and when. Regarding the effect of the intervention on patient care, results were mixed. As reported in 2016, we saw no difference in frequency of adverse events with this intervention.12 However, a more sensitive measure of adverse events—used in a study on handoffs—showed our regionalization efforts had an additive effect on reducing overnight adverse events.17Researchers should now focus on the effects of care redesign on clinical outcomes, interdisciplinary care team communication, patient engagement and satisfaction, provider opinions of communication, workflow, patient care, and housestaff education. Our methodology can be used as a model to link structure, process, and outcome related to rounds and thereby better understand how best to optimize patient care and efficiency. Additional studies are needed to analyze the content of rounds and their association with patient and educational outcomes. Last, it will be important to conduct a study to see if the effects we have identified can be sustained. Such a study is already under way.
In conclusion, creating regionalized care teams and encouraging focused bedside rounds increased the proportion of bedside time and the presence of nurses on rounds. Rounds were shorter despite higher patient census. TMA revealed that regionalized care teams and bedside rounding at a large academic hospital are feasible, and are useful in establishing the necessary structures for increasing physician–nurse and provider–patient interactions.
Acknowledgments
The authors acknowledge Dr. Stan Ashley, Dr. Jacqueline Somerville, and Sheila Harris for their support of the regionalization initiative.
Disclosures
Dr. Schnipper received funding from Sanofi-aventis to conduct an investigator-initiated study to implement and evaluate a multi-faceted intervention to improve transitions of care in patients discharged home on insulin. The study was also supported by funding from the Marshall A. Wolf Medical Education Fund, Brigham and Women’s Hospital, and Dr. Stan Ashley, Chief Medical Officer, Brigham and Women’s Hospital. Some of the content of this article was orally presented at the annual meeting of the Society of Hospital Medicine; March 29-April 1, 2015; National Harbor, MD.
References
1. Crumlish CM, Yialamas MA, McMahon GT. Quantification of bedside teaching by an academic hospitalist group. J Hosp Med. 2009;4(5):304-307. PubMed 2. Gonzalo JD, Masters PA, Simons RJ, Chuang CH. Attending rounds and bedside case presentations: medical student and medicine resident experiences and attitudes. Teach Learn Med. 2009;21(2):105-110. PubMed 3. Elliot DL, Hickam DH. Attending rounds on in-patient units: differences between medical and non-medical services. Med Educ. 1993;27(6):503-508. PubMed 4. Payson HE, Barchas JD. A time study of medical teaching rounds. N Engl J Med. 1965;273(27):1468-1471. PubMed 5. Tremonti LP, Biddle WB. Teaching behaviors of residents and faculty members. J Med Educ. 1982;57(11):854-859. PubMed 6. Miller M, Johnson B, Greene HL, Baier M, Nowlin S. An observational study of attending rounds. J Gen Intern Med. 1992;7(6):646-648. PubMed 7. Collins GF, Cassie JM, Daggett CJ. The role of the attending physician in clinical training. J Med Educ. 1978;53(5):429-431. PubMed 8. Ward DR, Ghali WA, Graham A, Lemaire JB. A real-time locating system observes physician time-motion patterns during walk-rounds: a pilot study. BMC Med Educ. 2014;14:37. PubMed 9. Gonzalo JD, Kuperman E, Lehman E, Haidet P. Bedside interprofessional rounds: perceptions of benefits and barriers by internal medicine nursing staff, attending physicians, and housestaff physicians. J Hosp Med. 2014;9(10):646-651. PubMed 10. Stickrath C, Noble M, Prochazka A, et al. Attending rounds in the current era: what is and is not happening. JAMA Intern Med. 2013;173(12):1084-1089. PubMed 11. Boxer R, Vitale M, Gershanik EF, et al. 5th time’s a charm: creation of unit-based care teams in a high occupancy hospital [abstract]. J Hosp Med. 2015;10(suppl 2). 12. Mueller SK, Schnipper JL, Giannelli K, Roy CL, Boxer R. Impact of regionalized care on concordance of plan and preventable adverse events on general medicine services. J Hosp Med. 2016;11(9):620-627. PubMed 13. Chauke HL, Pattinson RC. Ward rounds—bedside or conference room? S Afr Med J. 2006;96(5):398-400. PubMed 14. Wang-Cheng RM, Barnas GP, Sigmann P, Riendl PA, Young MJ. Bedside case presentations: why patients like them but learners don’t. J Gen Intern Med. 1989;4(4):284-287. PubMed 15. Lehmann LS, Brancati FL, Chen MC, Roter D, Dobs AS. The effect of bedside case presentations on patients’ perceptions of their medical care. N Engl J Med. 1997;336(16):1150-1155. PubMed 16. Gonzalo JD, Chuang CH, Huang G, Smith C. The return of bedside rounds: an educational intervention. J Gen Intern Med. 2010;25(8):792-798. PubMed 17. Mueller SK, Yoon C, Schnipper JL. Association of a web-based handoff tool with rates of medical errors. JAMA Intern Med. 2016;176(9):1400-1402. PubMed
References
1. Crumlish CM, Yialamas MA, McMahon GT. Quantification of bedside teaching by an academic hospitalist group. J Hosp Med. 2009;4(5):304-307. PubMed 2. Gonzalo JD, Masters PA, Simons RJ, Chuang CH. Attending rounds and bedside case presentations: medical student and medicine resident experiences and attitudes. Teach Learn Med. 2009;21(2):105-110. PubMed 3. Elliot DL, Hickam DH. Attending rounds on in-patient units: differences between medical and non-medical services. Med Educ. 1993;27(6):503-508. PubMed 4. Payson HE, Barchas JD. A time study of medical teaching rounds. N Engl J Med. 1965;273(27):1468-1471. PubMed 5. Tremonti LP, Biddle WB. Teaching behaviors of residents and faculty members. J Med Educ. 1982;57(11):854-859. PubMed 6. Miller M, Johnson B, Greene HL, Baier M, Nowlin S. An observational study of attending rounds. J Gen Intern Med. 1992;7(6):646-648. PubMed 7. Collins GF, Cassie JM, Daggett CJ. The role of the attending physician in clinical training. J Med Educ. 1978;53(5):429-431. PubMed 8. Ward DR, Ghali WA, Graham A, Lemaire JB. A real-time locating system observes physician time-motion patterns during walk-rounds: a pilot study. BMC Med Educ. 2014;14:37. PubMed 9. Gonzalo JD, Kuperman E, Lehman E, Haidet P. Bedside interprofessional rounds: perceptions of benefits and barriers by internal medicine nursing staff, attending physicians, and housestaff physicians. J Hosp Med. 2014;9(10):646-651. PubMed 10. Stickrath C, Noble M, Prochazka A, et al. Attending rounds in the current era: what is and is not happening. JAMA Intern Med. 2013;173(12):1084-1089. PubMed 11. Boxer R, Vitale M, Gershanik EF, et al. 5th time’s a charm: creation of unit-based care teams in a high occupancy hospital [abstract]. J Hosp Med. 2015;10(suppl 2). 12. Mueller SK, Schnipper JL, Giannelli K, Roy CL, Boxer R. Impact of regionalized care on concordance of plan and preventable adverse events on general medicine services. J Hosp Med. 2016;11(9):620-627. PubMed 13. Chauke HL, Pattinson RC. Ward rounds—bedside or conference room? S Afr Med J. 2006;96(5):398-400. PubMed 14. Wang-Cheng RM, Barnas GP, Sigmann P, Riendl PA, Young MJ. Bedside case presentations: why patients like them but learners don’t. J Gen Intern Med. 1989;4(4):284-287. PubMed 15. Lehmann LS, Brancati FL, Chen MC, Roter D, Dobs AS. The effect of bedside case presentations on patients’ perceptions of their medical care. N Engl J Med. 1997;336(16):1150-1155. PubMed 16. Gonzalo JD, Chuang CH, Huang G, Smith C. The return of bedside rounds: an educational intervention. J Gen Intern Med. 2010;25(8):792-798. PubMed 17. Mueller SK, Yoon C, Schnipper JL. Association of a web-based handoff tool with rates of medical errors. JAMA Intern Med. 2016;176(9):1400-1402. PubMed
Address for correspondence and reprint requests: Robert Boxer, MD, PhD, Brigham and Women’s Hospital, 75 Francis St, PBB-B 412, Boston, MA 02115; Telephone: 617-278-0055; Fax: 617-278-6906; E-mail: rboxer@partners.org
Content Gating
Gated (full article locked unless allowed per User)
In recent years, rapid response teams (RRTs) have been widely implemented to improve patient safety and quality of care. RRTs traditionally are activated by providers to address a clinically deteriorating patient; trained nurses, respiratory care specialists, and physicians are brought bedside to assist in triage and management. After the Joint Commission1 endorsed patient engagement as a strategy for enhancing patient safety, new initiatives were developed to meet the challenge. Programs designed to enhance patient engagement have taken a variety of forms, including educational campaigns encouraging patients to report adverse events, requests for handwashing by providers, and the institution of patient- and family-activated RRTs.2 Patient involvement is viewed favorably and has been shown to increase patients’ perception of health care quality.3 Although these initiatives are presumed helpful in encouraging communication, there is limited evidence that more communication leads to safety improvements. Despite the increasing prevalence of patient-activated RRTs in the United States, they have gone largely unevaluated in the adult population, and their efficacy remains unclear.
CONDITION HELP
Condition Help (CH) is a patient- and family-initiated RRT designed to prevent medical errors and communication problems and improve patient safety. Patients and families are encouraged to call the CH hotline if they believe that there has been a breakdown in care or that their health is in imminent danger. This RRT was inspired by the case of Josie King, an 18-month-old girl who died of preventable causes at a large children’s hospital.4 After her daughter’s death, Sorrel King started the Josie King Foundation, an organization committed to preventing medical errors and creating a culture of patient safety. With the support of this foundation, CH was launched in 2005 at the Children’s Hospital of Pittsburgh at the University of Pittsburgh Medical Center (UPMC). Later it was implemented at the UPMC adult tertiary-care center, and now it is available in all UPMC facilities.
On admission, patients receive a brochure that details the purpose of CH and provides examples of when and how to call the CH hotline. In this brochure, patients are instructed to call CH in 3 situations: “1) There is an emergency and you cannot get the attention of hospital staff, 2) You see a change in the patient’s condition and the health care team is not recognizing the concern, or 3) There is breakdown in how care is given or uncertainty over what needs to be done.” These instructions are printed on bulletins placed in elevators and hallways throughout the hospital. Patients and families may activate the system at any time and can even do so from home.
When a patient or family member calls the hotline, an operator notifies the CH team. This team, which consists of a patient care liaison (or an on-duty administrator) and the unit charge nurse, convenes bedside to address the patient’s concern. The team was designed without a physician to ensure that the primary team remains in charge of the care plan. CH is kept separate from our traditional RRT and does not compete for resources (personnel, equipment, time) with the RRT, which is designed to address a clinically deteriorating patient.
In this article, we describe the characteristics of patients for whom CH was activated at our adult hospital. We also describe reasons for calls, whether changes in care were implemented, and outcomes, including traditional RRT activation, transfer to intensive care unit (ICU), and inpatient mortality. As CH was designed with patient safety as a goal, we tracked 2 types of calls, those involving safety issues and those involving nonsafety issues.
METHODS
This study was approved by the quality improvement committee at the University of Pittsburgh and was considered exempt from review by the university’s Institutional Review Board.
Our integrated health system consists of more than 20 hospitals serving a tristate region. UPMC Presbyterian and UPMC Montefiore are adult tertiary-care referral hospitals with more than 750 medical/surgical beds and 150 critical care beds and more than 30,000 annual inpatient admissions. These hospitals are physically connected and function as a single large medical center. We reviewed all CH events that occurred at this combined hospital during the period January 2012 through June 2015. The dates coincided with CH data acquisition.
CH was available 24 hours a day 7 days a week. A patient care liaison (or an on-duty administrator) and the unit charge nurse responded to CH calls. Data from all calls included date and time of call, day of week, primary service, patient location, unique patient identifiers, call initiator (patient or family), whether a call led to changes in care, and primary reason for call. Each call reason was sorted into 1 of 10 categories: pain control, staff problem, lack of communication between patient/family and care team, questions about patient management, care delays, delays in a particular service, questions about discharge, administrative issues, acute psychiatric needs, and unknown/other. In addition, after a call, we reviewed all charts to determine if a safety issue was involved; Dr. Eden and Dr. Bump independently reviewed calls for safety issues and discussed any differences until they reached consensus. We also recorded outcomes, including activation of a traditional RRT or transfer to ICU within 24 hours of CH call, inpatient mortality, and against medical advice (AMA) discharges. Given that many calls were made by patients who called more than once (during a single admission or over multiple admissions), we also sorted patients into one-time callers and repeat callers for comparison. Patient satisfaction data were unavailable for review.
Patient demographic data are presented as means, standard deviations, and percentages, and call characteristics as percentages. Chi-square tests and t tests were used for analyses except for comparisons having few observations. For those, Fisher exact test was used. All analyses were performed with SAS Version 9.4 (SAS Institute, Cary, North Carolina).
RESULTS
From January 2012 through June 2015, 367 CH calls were made, about 105 annually. During this period, there were about 33,000 admissions, 800 combined grievances and complaints, 170 AMA discharges, 155 cardiac arrests, 2300 traditional RRT activations, and 1200 inpatient deaths per year. The 367 CH calls were made by 240 patients (Table 1). Of these 240 patients, 43 (18%) activated the CH team with multiple calls; their calls accounted for (46.3%) of all calls (170/367). The majority of calls were made by patients (76.8%) rather than family members (21.8%). Mean (SD) patient age was 45.8 (17.4) years. Mean (SD) number of admissions per patient per year was 2.7 (3.5). More events were activated for patients admitted to medical services (66%) than surgical services (34%). Calls were evenly distributed between time of day and day of week.
Table 1
The most common reason for CH calls was inadequate pain control (48.2%), followed by dissatisfaction with staff (12.5%); the remaining calls were evenly distributed among the other categories. The majority of calls involved nonsafety issues (83.4%) rather than safety issues (11.4%); in 5.2% of calls, the distinction could not be made because of lack of information (Table 2). In 152 (41.4%) of the 367 total calls, a change in care or alteration in management was made. Of these 152 calls, 99 (65.1%) involved distinct changes in the care plan, such as medication changes, imaging or additional testing, or consultation with other physicians; the other 53 calls (34.9%) involved additional patient counseling or nonmedical changes. Our traditional RRT was activated within 24 hours of CH in 19 cases (5.2%); of the 19 patients, 6 were transferred to ICU. Seven patients (2.9%) died during admission. Twelve (3.3%) were discharged AMA. We compared outcomes of patients who made safety-issue calls with those of patients who made nonsafety-issue calls. The composite outcome of RRT activation, ICU transfer, and mortality was found for 6 (14.3%) of the 42 safety-issue calls and 15 (4.9%) of the 306 nonsafety-issue calls (P = 0.0291).
Table 2 The unexpected high rate of repeat calling prompted us to compare the characteristics of one-time and repeat callers. Repeat callers were younger: Mean age was 39.3 (12.8) years for repeat callers and 47.2 (17.9) years for one-time callers (P = 0.0012). Repeat callers had more admissions per year: Mean (SD) number of admissions was 5.67 (5.4) for repeat callers and 2.09 (2.5) for one-time callers (P = 0.0001). One-time and repeat callers did not differ with respect to race or sex. Compared with one-time callers, repeat callers were more often (P = 0.002) admitted to medical services (74.7%) than surgical services (58.9%). For repeat callers, a larger percentage of calls (P < 0.0001) were made by patients (93.5%) rather than families (62.4%). Calls about pain were more often (P < 0.0001) made by repeat callers (62.3%) than one-time callers (36%), calls involving safety issues were less often (P < 0.0001) made by repeat callers (5.9%) than one-time callers (16.2%), and changes in care were made less often (P < 0.0001) for repeat callers (32.9%) than one-time callers (48.7%). Between-group differences in rates of RRT activation, transfer to ICU, inpatient mortality, and AMA discharges were not significant.
DISCUSSION
Patient- and family-activated RRTs provide unique opportunities for patient and family engagement during inpatient hospital stays. Our study described the results obtained with use of a well-established patient-activated RRT over several years, one of the longer observation periods reported in the literature. We found that, with use of patient-activated RRTs, patient safety issues were identified, though these were far outnumbered by nonsafety issues.
Almost half of all CH events were related to pain. Pain as the primary driver for RRT activation may be attributable to several factors, including degree of illness, poor communication about pain management expectations, positive reinforcement of narcotic-seeking behavior as a result of CH activation, and high rate of opiate use in the catchment area. A striking finding of our analysis was repeat calling; only 43 (18%) of the 240 callers were repeat callers, but they made almost half of all the calls. In some cases, during a single admission, multiple calls were made because the first had no effect on care or management; more typically, though, multiple calls were made over several admissions. Repeat callers were admitted more often per year, and they used hospital services more. They should be further studied with a goal of designing programs that better meet their needs and that prospectively address expectations of pain control.
Our study was unique in describing several outcomes related to CH events. We found that traditional RRTs were seldom activated, level of care was seldom escalated, and mortality was rare, though these outcomes occurred more often for safety-issue calls than nonsafety-issue calls. We also found that activation of CH teams often led to changes in medical management, though we could not determine whether these changes in care led to different patient outcomes.
Patient-initiated RRTs are described in a limited number of pediatric and adult studies, all with findings differing from ours. In the pediatric models, most calls were initiated by family members, were less frequent, and tended to signal higher patient acuity.5,6 For example, in a pediatric RRT model,5 family members activated the RRT only twice within the study year, but both calls resulted in ICU transfer. Most descriptions of patient-activated RRTs in adult hospitals are from pilot studies, which similarly identified infrequent RRT calls but often did not identify call reasons or specific outcomes.7 A single-center study concluded that, after implementation of a mixed-model RRT8—a traditional practitioner-activated RRT later enhanced with a patient/family activation mechanism—non-ICU codes decreased, and there was a statistically significant drop in hospital-wide mortality rates. However, this RRT was patient-activated only 25 times over 2 years, and the specific outcomes of those events were not described.
Other initiatives have been designed to enhance patient care and communication. Purposeful rounding systems9 involve hourly rounding by bedside nurses and daily rounding by nurse leaders to improve timely patient care and provide proactive service. Such systems ideally preempt calls involving dissatisfaction and nonsafety issues. Although they would reduce the number of patient-dissatisfaction calls made in the CH system, they may not be any better than the CH system is in its main purpose, identifying safety issues. In addition, whether patient-activated RRTs or purposeful rounding systems are better at addressing patient dissatisfaction is unclear.
This study had its limitations. First, like other studies, it was a single-center observational study without a concurrent control group. Second, because CH was first implemented 10 years ago, we could not compare patient outcomes or traditional RRT use before and after program initiation. Third, our study cohort consisted of patients hospitalized at one academic tertiary-care center in one region, and the hospital is a training site for multiple residencies and fellowships. These factors likely affect the generalizability of our data to smaller or community-based centers. Fourth, some determinations were subjective (eg, whether calls involved safety or nonsafety issues). We tried to minimize bias by having 2 authors independently review cases, but the process did not reflect patient experience or perspective. Fifth, our hospital adopted its traditional RRT years before its CH system. The criteria used by hospital personnel for traditional RRT activation are designed to encourage staff to call for help at early signs of patient deterioration. Consequently, traditional RRT activations substantially outnumber CH calls. Whether this resulted in fewer CH safety calls is unclear. Sixth, we did not capture the financial implications of using CH teams.
Although patient-activated RRTs identified patient safety issues, questions about the utility or necessity of these RRTs remain. In our era of limited hospital resources, the case has not been definitively made that these teams are practical, based on patient outcomes, though other studies have found improved patient satisfaction.7 Most of the RRT calls in our study involved patient dissatisfaction and communication issues. CH may not be the ideal approach for managing these issues, but it represents the last line of patient advocacy once other systems have failed.
We think patient-activated RRTs have the potential to effect patient engagement in safe care. Given the importance of establishing a culture of patient safety and engagement, and increased detection of safety-related events, CH remains active throughout our hospital system. Newer iterations of CH may benefit from stricter language in defining appropriate occasions for calling RRTs, and from descriptions of other resources for patient advocacy within the hospital. These modifications could end up restricting RRT activations to patient complaints and preserving CH resources for patients with safety concerns. Our study lays the groundwork for other institutions that are considering similar interventions. Studies should now start evaluating how well patient- and family-activated RRTs improve patient satisfaction, staff satisfaction, and patient outcomes.
CONCLUSION
Patient- and family-activated RRTs were designed to engage patients and families in safe care. Although CH detects patient safety issues, these are far outnumbered by nonsafety issues. CH demonstrates a commitment to patient engagement and a culture that emphasizes patient safety.
Acknowledgements
This work was presented as a poster at the annual meeting of the Society of Hospital Medicine; March 6-9, 2016; San Diego, CA.
Disclosure
Nothing to report.
References
1. Joint Commission. Improving America’s Hospitals: The Joint Commission’s Annual Report on Quality and Safety 2008. http://www.jointcommission.org/assets/1/6/2008_Annual_Report.pdf. Published November 2008. Accessed May 4, 2016. PubMed 2. Berger Z, Flickinger TE, Pfoh E, Martinez KA, Dy SM. Promoting engagement by patients and families to reduce adverse events in acute care settings: a systematic review. BMJ Qual Saf. 2014;23(7):548-555. PubMed 3. Weingart SN, Zhu J, Chiappetta L, et al. Hospitalized patients’ participation and its impact on quality of care and patient safety. Int J Qual Health Care. 2011;23(3):269-277. PubMed 4. Kennedy P, Pronovost P. Shepherding change: how the market, healthcare providers, and public policy can deliver quality care for the 21st century. Crit Care Med. 2006;34(3 suppl):S1-S6. PubMed 5. Ray EM, Smith R, Massie S, et al. Family alert: implementing direct family activation of a pediatric response team. Jt Comm J Qual Patient Saf. 2009;35(11):575-580. PubMed 6. Dean BS, Decker MJ, Hupp D, Urbach AH, Lewis E, Benes-Stickle J. Condition Help: a pediatric rapid response team triggered by patients and parents. J Healthc Qual. 2008;30(3):28-31. PubMed 7. Vorwerk J, King L. Consumer participation in early detection of the deteriorating patient and call activation to rapid response systems: a literature review. J Clin Nurs. 2015;25(1-2):38-52. PubMed 8. Gerdik C, Vallish RO, Miles K, Godwin SA, Wludyka PS, Panni MK. Successful implementation of a family and patient activated rapid response team in an adult level 1 trauma center. Resuscitation. 2010;81(12):1676-1681. PubMed 9. Hancock KK. From the bedside: purposeful rounding essential to patient experience. Association for Patient Experience website. http://www.patient-experience.org/Resources/Newsletter/Newsletters/Articles/2014/From-the-Bedside-Purposeful-Rounding-Essential-to.aspx. Published February 27, 2014. Accessed July 25, 2016.
In recent years, rapid response teams (RRTs) have been widely implemented to improve patient safety and quality of care. RRTs traditionally are activated by providers to address a clinically deteriorating patient; trained nurses, respiratory care specialists, and physicians are brought bedside to assist in triage and management. After the Joint Commission1 endorsed patient engagement as a strategy for enhancing patient safety, new initiatives were developed to meet the challenge. Programs designed to enhance patient engagement have taken a variety of forms, including educational campaigns encouraging patients to report adverse events, requests for handwashing by providers, and the institution of patient- and family-activated RRTs.2 Patient involvement is viewed favorably and has been shown to increase patients’ perception of health care quality.3 Although these initiatives are presumed helpful in encouraging communication, there is limited evidence that more communication leads to safety improvements. Despite the increasing prevalence of patient-activated RRTs in the United States, they have gone largely unevaluated in the adult population, and their efficacy remains unclear.
CONDITION HELP
Condition Help (CH) is a patient- and family-initiated RRT designed to prevent medical errors and communication problems and improve patient safety. Patients and families are encouraged to call the CH hotline if they believe that there has been a breakdown in care or that their health is in imminent danger. This RRT was inspired by the case of Josie King, an 18-month-old girl who died of preventable causes at a large children’s hospital.4 After her daughter’s death, Sorrel King started the Josie King Foundation, an organization committed to preventing medical errors and creating a culture of patient safety. With the support of this foundation, CH was launched in 2005 at the Children’s Hospital of Pittsburgh at the University of Pittsburgh Medical Center (UPMC). Later it was implemented at the UPMC adult tertiary-care center, and now it is available in all UPMC facilities.
On admission, patients receive a brochure that details the purpose of CH and provides examples of when and how to call the CH hotline. In this brochure, patients are instructed to call CH in 3 situations: “1) There is an emergency and you cannot get the attention of hospital staff, 2) You see a change in the patient’s condition and the health care team is not recognizing the concern, or 3) There is breakdown in how care is given or uncertainty over what needs to be done.” These instructions are printed on bulletins placed in elevators and hallways throughout the hospital. Patients and families may activate the system at any time and can even do so from home.
When a patient or family member calls the hotline, an operator notifies the CH team. This team, which consists of a patient care liaison (or an on-duty administrator) and the unit charge nurse, convenes bedside to address the patient’s concern. The team was designed without a physician to ensure that the primary team remains in charge of the care plan. CH is kept separate from our traditional RRT and does not compete for resources (personnel, equipment, time) with the RRT, which is designed to address a clinically deteriorating patient.
In this article, we describe the characteristics of patients for whom CH was activated at our adult hospital. We also describe reasons for calls, whether changes in care were implemented, and outcomes, including traditional RRT activation, transfer to intensive care unit (ICU), and inpatient mortality. As CH was designed with patient safety as a goal, we tracked 2 types of calls, those involving safety issues and those involving nonsafety issues.
METHODS
This study was approved by the quality improvement committee at the University of Pittsburgh and was considered exempt from review by the university’s Institutional Review Board.
Our integrated health system consists of more than 20 hospitals serving a tristate region. UPMC Presbyterian and UPMC Montefiore are adult tertiary-care referral hospitals with more than 750 medical/surgical beds and 150 critical care beds and more than 30,000 annual inpatient admissions. These hospitals are physically connected and function as a single large medical center. We reviewed all CH events that occurred at this combined hospital during the period January 2012 through June 2015. The dates coincided with CH data acquisition.
CH was available 24 hours a day 7 days a week. A patient care liaison (or an on-duty administrator) and the unit charge nurse responded to CH calls. Data from all calls included date and time of call, day of week, primary service, patient location, unique patient identifiers, call initiator (patient or family), whether a call led to changes in care, and primary reason for call. Each call reason was sorted into 1 of 10 categories: pain control, staff problem, lack of communication between patient/family and care team, questions about patient management, care delays, delays in a particular service, questions about discharge, administrative issues, acute psychiatric needs, and unknown/other. In addition, after a call, we reviewed all charts to determine if a safety issue was involved; Dr. Eden and Dr. Bump independently reviewed calls for safety issues and discussed any differences until they reached consensus. We also recorded outcomes, including activation of a traditional RRT or transfer to ICU within 24 hours of CH call, inpatient mortality, and against medical advice (AMA) discharges. Given that many calls were made by patients who called more than once (during a single admission or over multiple admissions), we also sorted patients into one-time callers and repeat callers for comparison. Patient satisfaction data were unavailable for review.
Patient demographic data are presented as means, standard deviations, and percentages, and call characteristics as percentages. Chi-square tests and t tests were used for analyses except for comparisons having few observations. For those, Fisher exact test was used. All analyses were performed with SAS Version 9.4 (SAS Institute, Cary, North Carolina).
RESULTS
From January 2012 through June 2015, 367 CH calls were made, about 105 annually. During this period, there were about 33,000 admissions, 800 combined grievances and complaints, 170 AMA discharges, 155 cardiac arrests, 2300 traditional RRT activations, and 1200 inpatient deaths per year. The 367 CH calls were made by 240 patients (Table 1). Of these 240 patients, 43 (18%) activated the CH team with multiple calls; their calls accounted for (46.3%) of all calls (170/367). The majority of calls were made by patients (76.8%) rather than family members (21.8%). Mean (SD) patient age was 45.8 (17.4) years. Mean (SD) number of admissions per patient per year was 2.7 (3.5). More events were activated for patients admitted to medical services (66%) than surgical services (34%). Calls were evenly distributed between time of day and day of week.
Table 1
The most common reason for CH calls was inadequate pain control (48.2%), followed by dissatisfaction with staff (12.5%); the remaining calls were evenly distributed among the other categories. The majority of calls involved nonsafety issues (83.4%) rather than safety issues (11.4%); in 5.2% of calls, the distinction could not be made because of lack of information (Table 2). In 152 (41.4%) of the 367 total calls, a change in care or alteration in management was made. Of these 152 calls, 99 (65.1%) involved distinct changes in the care plan, such as medication changes, imaging or additional testing, or consultation with other physicians; the other 53 calls (34.9%) involved additional patient counseling or nonmedical changes. Our traditional RRT was activated within 24 hours of CH in 19 cases (5.2%); of the 19 patients, 6 were transferred to ICU. Seven patients (2.9%) died during admission. Twelve (3.3%) were discharged AMA. We compared outcomes of patients who made safety-issue calls with those of patients who made nonsafety-issue calls. The composite outcome of RRT activation, ICU transfer, and mortality was found for 6 (14.3%) of the 42 safety-issue calls and 15 (4.9%) of the 306 nonsafety-issue calls (P = 0.0291).
Table 2 The unexpected high rate of repeat calling prompted us to compare the characteristics of one-time and repeat callers. Repeat callers were younger: Mean age was 39.3 (12.8) years for repeat callers and 47.2 (17.9) years for one-time callers (P = 0.0012). Repeat callers had more admissions per year: Mean (SD) number of admissions was 5.67 (5.4) for repeat callers and 2.09 (2.5) for one-time callers (P = 0.0001). One-time and repeat callers did not differ with respect to race or sex. Compared with one-time callers, repeat callers were more often (P = 0.002) admitted to medical services (74.7%) than surgical services (58.9%). For repeat callers, a larger percentage of calls (P < 0.0001) were made by patients (93.5%) rather than families (62.4%). Calls about pain were more often (P < 0.0001) made by repeat callers (62.3%) than one-time callers (36%), calls involving safety issues were less often (P < 0.0001) made by repeat callers (5.9%) than one-time callers (16.2%), and changes in care were made less often (P < 0.0001) for repeat callers (32.9%) than one-time callers (48.7%). Between-group differences in rates of RRT activation, transfer to ICU, inpatient mortality, and AMA discharges were not significant.
DISCUSSION
Patient- and family-activated RRTs provide unique opportunities for patient and family engagement during inpatient hospital stays. Our study described the results obtained with use of a well-established patient-activated RRT over several years, one of the longer observation periods reported in the literature. We found that, with use of patient-activated RRTs, patient safety issues were identified, though these were far outnumbered by nonsafety issues.
Almost half of all CH events were related to pain. Pain as the primary driver for RRT activation may be attributable to several factors, including degree of illness, poor communication about pain management expectations, positive reinforcement of narcotic-seeking behavior as a result of CH activation, and high rate of opiate use in the catchment area. A striking finding of our analysis was repeat calling; only 43 (18%) of the 240 callers were repeat callers, but they made almost half of all the calls. In some cases, during a single admission, multiple calls were made because the first had no effect on care or management; more typically, though, multiple calls were made over several admissions. Repeat callers were admitted more often per year, and they used hospital services more. They should be further studied with a goal of designing programs that better meet their needs and that prospectively address expectations of pain control.
Our study was unique in describing several outcomes related to CH events. We found that traditional RRTs were seldom activated, level of care was seldom escalated, and mortality was rare, though these outcomes occurred more often for safety-issue calls than nonsafety-issue calls. We also found that activation of CH teams often led to changes in medical management, though we could not determine whether these changes in care led to different patient outcomes.
Patient-initiated RRTs are described in a limited number of pediatric and adult studies, all with findings differing from ours. In the pediatric models, most calls were initiated by family members, were less frequent, and tended to signal higher patient acuity.5,6 For example, in a pediatric RRT model,5 family members activated the RRT only twice within the study year, but both calls resulted in ICU transfer. Most descriptions of patient-activated RRTs in adult hospitals are from pilot studies, which similarly identified infrequent RRT calls but often did not identify call reasons or specific outcomes.7 A single-center study concluded that, after implementation of a mixed-model RRT8—a traditional practitioner-activated RRT later enhanced with a patient/family activation mechanism—non-ICU codes decreased, and there was a statistically significant drop in hospital-wide mortality rates. However, this RRT was patient-activated only 25 times over 2 years, and the specific outcomes of those events were not described.
Other initiatives have been designed to enhance patient care and communication. Purposeful rounding systems9 involve hourly rounding by bedside nurses and daily rounding by nurse leaders to improve timely patient care and provide proactive service. Such systems ideally preempt calls involving dissatisfaction and nonsafety issues. Although they would reduce the number of patient-dissatisfaction calls made in the CH system, they may not be any better than the CH system is in its main purpose, identifying safety issues. In addition, whether patient-activated RRTs or purposeful rounding systems are better at addressing patient dissatisfaction is unclear.
This study had its limitations. First, like other studies, it was a single-center observational study without a concurrent control group. Second, because CH was first implemented 10 years ago, we could not compare patient outcomes or traditional RRT use before and after program initiation. Third, our study cohort consisted of patients hospitalized at one academic tertiary-care center in one region, and the hospital is a training site for multiple residencies and fellowships. These factors likely affect the generalizability of our data to smaller or community-based centers. Fourth, some determinations were subjective (eg, whether calls involved safety or nonsafety issues). We tried to minimize bias by having 2 authors independently review cases, but the process did not reflect patient experience or perspective. Fifth, our hospital adopted its traditional RRT years before its CH system. The criteria used by hospital personnel for traditional RRT activation are designed to encourage staff to call for help at early signs of patient deterioration. Consequently, traditional RRT activations substantially outnumber CH calls. Whether this resulted in fewer CH safety calls is unclear. Sixth, we did not capture the financial implications of using CH teams.
Although patient-activated RRTs identified patient safety issues, questions about the utility or necessity of these RRTs remain. In our era of limited hospital resources, the case has not been definitively made that these teams are practical, based on patient outcomes, though other studies have found improved patient satisfaction.7 Most of the RRT calls in our study involved patient dissatisfaction and communication issues. CH may not be the ideal approach for managing these issues, but it represents the last line of patient advocacy once other systems have failed.
We think patient-activated RRTs have the potential to effect patient engagement in safe care. Given the importance of establishing a culture of patient safety and engagement, and increased detection of safety-related events, CH remains active throughout our hospital system. Newer iterations of CH may benefit from stricter language in defining appropriate occasions for calling RRTs, and from descriptions of other resources for patient advocacy within the hospital. These modifications could end up restricting RRT activations to patient complaints and preserving CH resources for patients with safety concerns. Our study lays the groundwork for other institutions that are considering similar interventions. Studies should now start evaluating how well patient- and family-activated RRTs improve patient satisfaction, staff satisfaction, and patient outcomes.
CONCLUSION
Patient- and family-activated RRTs were designed to engage patients and families in safe care. Although CH detects patient safety issues, these are far outnumbered by nonsafety issues. CH demonstrates a commitment to patient engagement and a culture that emphasizes patient safety.
Acknowledgements
This work was presented as a poster at the annual meeting of the Society of Hospital Medicine; March 6-9, 2016; San Diego, CA.
Disclosure
Nothing to report.
In recent years, rapid response teams (RRTs) have been widely implemented to improve patient safety and quality of care. RRTs traditionally are activated by providers to address a clinically deteriorating patient; trained nurses, respiratory care specialists, and physicians are brought bedside to assist in triage and management. After the Joint Commission1 endorsed patient engagement as a strategy for enhancing patient safety, new initiatives were developed to meet the challenge. Programs designed to enhance patient engagement have taken a variety of forms, including educational campaigns encouraging patients to report adverse events, requests for handwashing by providers, and the institution of patient- and family-activated RRTs.2 Patient involvement is viewed favorably and has been shown to increase patients’ perception of health care quality.3 Although these initiatives are presumed helpful in encouraging communication, there is limited evidence that more communication leads to safety improvements. Despite the increasing prevalence of patient-activated RRTs in the United States, they have gone largely unevaluated in the adult population, and their efficacy remains unclear.
CONDITION HELP
Condition Help (CH) is a patient- and family-initiated RRT designed to prevent medical errors and communication problems and improve patient safety. Patients and families are encouraged to call the CH hotline if they believe that there has been a breakdown in care or that their health is in imminent danger. This RRT was inspired by the case of Josie King, an 18-month-old girl who died of preventable causes at a large children’s hospital.4 After her daughter’s death, Sorrel King started the Josie King Foundation, an organization committed to preventing medical errors and creating a culture of patient safety. With the support of this foundation, CH was launched in 2005 at the Children’s Hospital of Pittsburgh at the University of Pittsburgh Medical Center (UPMC). Later it was implemented at the UPMC adult tertiary-care center, and now it is available in all UPMC facilities.
On admission, patients receive a brochure that details the purpose of CH and provides examples of when and how to call the CH hotline. In this brochure, patients are instructed to call CH in 3 situations: “1) There is an emergency and you cannot get the attention of hospital staff, 2) You see a change in the patient’s condition and the health care team is not recognizing the concern, or 3) There is breakdown in how care is given or uncertainty over what needs to be done.” These instructions are printed on bulletins placed in elevators and hallways throughout the hospital. Patients and families may activate the system at any time and can even do so from home.
When a patient or family member calls the hotline, an operator notifies the CH team. This team, which consists of a patient care liaison (or an on-duty administrator) and the unit charge nurse, convenes bedside to address the patient’s concern. The team was designed without a physician to ensure that the primary team remains in charge of the care plan. CH is kept separate from our traditional RRT and does not compete for resources (personnel, equipment, time) with the RRT, which is designed to address a clinically deteriorating patient.
In this article, we describe the characteristics of patients for whom CH was activated at our adult hospital. We also describe reasons for calls, whether changes in care were implemented, and outcomes, including traditional RRT activation, transfer to intensive care unit (ICU), and inpatient mortality. As CH was designed with patient safety as a goal, we tracked 2 types of calls, those involving safety issues and those involving nonsafety issues.
METHODS
This study was approved by the quality improvement committee at the University of Pittsburgh and was considered exempt from review by the university’s Institutional Review Board.
Our integrated health system consists of more than 20 hospitals serving a tristate region. UPMC Presbyterian and UPMC Montefiore are adult tertiary-care referral hospitals with more than 750 medical/surgical beds and 150 critical care beds and more than 30,000 annual inpatient admissions. These hospitals are physically connected and function as a single large medical center. We reviewed all CH events that occurred at this combined hospital during the period January 2012 through June 2015. The dates coincided with CH data acquisition.
CH was available 24 hours a day 7 days a week. A patient care liaison (or an on-duty administrator) and the unit charge nurse responded to CH calls. Data from all calls included date and time of call, day of week, primary service, patient location, unique patient identifiers, call initiator (patient or family), whether a call led to changes in care, and primary reason for call. Each call reason was sorted into 1 of 10 categories: pain control, staff problem, lack of communication between patient/family and care team, questions about patient management, care delays, delays in a particular service, questions about discharge, administrative issues, acute psychiatric needs, and unknown/other. In addition, after a call, we reviewed all charts to determine if a safety issue was involved; Dr. Eden and Dr. Bump independently reviewed calls for safety issues and discussed any differences until they reached consensus. We also recorded outcomes, including activation of a traditional RRT or transfer to ICU within 24 hours of CH call, inpatient mortality, and against medical advice (AMA) discharges. Given that many calls were made by patients who called more than once (during a single admission or over multiple admissions), we also sorted patients into one-time callers and repeat callers for comparison. Patient satisfaction data were unavailable for review.
Patient demographic data are presented as means, standard deviations, and percentages, and call characteristics as percentages. Chi-square tests and t tests were used for analyses except for comparisons having few observations. For those, Fisher exact test was used. All analyses were performed with SAS Version 9.4 (SAS Institute, Cary, North Carolina).
RESULTS
From January 2012 through June 2015, 367 CH calls were made, about 105 annually. During this period, there were about 33,000 admissions, 800 combined grievances and complaints, 170 AMA discharges, 155 cardiac arrests, 2300 traditional RRT activations, and 1200 inpatient deaths per year. The 367 CH calls were made by 240 patients (Table 1). Of these 240 patients, 43 (18%) activated the CH team with multiple calls; their calls accounted for (46.3%) of all calls (170/367). The majority of calls were made by patients (76.8%) rather than family members (21.8%). Mean (SD) patient age was 45.8 (17.4) years. Mean (SD) number of admissions per patient per year was 2.7 (3.5). More events were activated for patients admitted to medical services (66%) than surgical services (34%). Calls were evenly distributed between time of day and day of week.
Table 1
The most common reason for CH calls was inadequate pain control (48.2%), followed by dissatisfaction with staff (12.5%); the remaining calls were evenly distributed among the other categories. The majority of calls involved nonsafety issues (83.4%) rather than safety issues (11.4%); in 5.2% of calls, the distinction could not be made because of lack of information (Table 2). In 152 (41.4%) of the 367 total calls, a change in care or alteration in management was made. Of these 152 calls, 99 (65.1%) involved distinct changes in the care plan, such as medication changes, imaging or additional testing, or consultation with other physicians; the other 53 calls (34.9%) involved additional patient counseling or nonmedical changes. Our traditional RRT was activated within 24 hours of CH in 19 cases (5.2%); of the 19 patients, 6 were transferred to ICU. Seven patients (2.9%) died during admission. Twelve (3.3%) were discharged AMA. We compared outcomes of patients who made safety-issue calls with those of patients who made nonsafety-issue calls. The composite outcome of RRT activation, ICU transfer, and mortality was found for 6 (14.3%) of the 42 safety-issue calls and 15 (4.9%) of the 306 nonsafety-issue calls (P = 0.0291).
Table 2 The unexpected high rate of repeat calling prompted us to compare the characteristics of one-time and repeat callers. Repeat callers were younger: Mean age was 39.3 (12.8) years for repeat callers and 47.2 (17.9) years for one-time callers (P = 0.0012). Repeat callers had more admissions per year: Mean (SD) number of admissions was 5.67 (5.4) for repeat callers and 2.09 (2.5) for one-time callers (P = 0.0001). One-time and repeat callers did not differ with respect to race or sex. Compared with one-time callers, repeat callers were more often (P = 0.002) admitted to medical services (74.7%) than surgical services (58.9%). For repeat callers, a larger percentage of calls (P < 0.0001) were made by patients (93.5%) rather than families (62.4%). Calls about pain were more often (P < 0.0001) made by repeat callers (62.3%) than one-time callers (36%), calls involving safety issues were less often (P < 0.0001) made by repeat callers (5.9%) than one-time callers (16.2%), and changes in care were made less often (P < 0.0001) for repeat callers (32.9%) than one-time callers (48.7%). Between-group differences in rates of RRT activation, transfer to ICU, inpatient mortality, and AMA discharges were not significant.
DISCUSSION
Patient- and family-activated RRTs provide unique opportunities for patient and family engagement during inpatient hospital stays. Our study described the results obtained with use of a well-established patient-activated RRT over several years, one of the longer observation periods reported in the literature. We found that, with use of patient-activated RRTs, patient safety issues were identified, though these were far outnumbered by nonsafety issues.
Almost half of all CH events were related to pain. Pain as the primary driver for RRT activation may be attributable to several factors, including degree of illness, poor communication about pain management expectations, positive reinforcement of narcotic-seeking behavior as a result of CH activation, and high rate of opiate use in the catchment area. A striking finding of our analysis was repeat calling; only 43 (18%) of the 240 callers were repeat callers, but they made almost half of all the calls. In some cases, during a single admission, multiple calls were made because the first had no effect on care or management; more typically, though, multiple calls were made over several admissions. Repeat callers were admitted more often per year, and they used hospital services more. They should be further studied with a goal of designing programs that better meet their needs and that prospectively address expectations of pain control.
Our study was unique in describing several outcomes related to CH events. We found that traditional RRTs were seldom activated, level of care was seldom escalated, and mortality was rare, though these outcomes occurred more often for safety-issue calls than nonsafety-issue calls. We also found that activation of CH teams often led to changes in medical management, though we could not determine whether these changes in care led to different patient outcomes.
Patient-initiated RRTs are described in a limited number of pediatric and adult studies, all with findings differing from ours. In the pediatric models, most calls were initiated by family members, were less frequent, and tended to signal higher patient acuity.5,6 For example, in a pediatric RRT model,5 family members activated the RRT only twice within the study year, but both calls resulted in ICU transfer. Most descriptions of patient-activated RRTs in adult hospitals are from pilot studies, which similarly identified infrequent RRT calls but often did not identify call reasons or specific outcomes.7 A single-center study concluded that, after implementation of a mixed-model RRT8—a traditional practitioner-activated RRT later enhanced with a patient/family activation mechanism—non-ICU codes decreased, and there was a statistically significant drop in hospital-wide mortality rates. However, this RRT was patient-activated only 25 times over 2 years, and the specific outcomes of those events were not described.
Other initiatives have been designed to enhance patient care and communication. Purposeful rounding systems9 involve hourly rounding by bedside nurses and daily rounding by nurse leaders to improve timely patient care and provide proactive service. Such systems ideally preempt calls involving dissatisfaction and nonsafety issues. Although they would reduce the number of patient-dissatisfaction calls made in the CH system, they may not be any better than the CH system is in its main purpose, identifying safety issues. In addition, whether patient-activated RRTs or purposeful rounding systems are better at addressing patient dissatisfaction is unclear.
This study had its limitations. First, like other studies, it was a single-center observational study without a concurrent control group. Second, because CH was first implemented 10 years ago, we could not compare patient outcomes or traditional RRT use before and after program initiation. Third, our study cohort consisted of patients hospitalized at one academic tertiary-care center in one region, and the hospital is a training site for multiple residencies and fellowships. These factors likely affect the generalizability of our data to smaller or community-based centers. Fourth, some determinations were subjective (eg, whether calls involved safety or nonsafety issues). We tried to minimize bias by having 2 authors independently review cases, but the process did not reflect patient experience or perspective. Fifth, our hospital adopted its traditional RRT years before its CH system. The criteria used by hospital personnel for traditional RRT activation are designed to encourage staff to call for help at early signs of patient deterioration. Consequently, traditional RRT activations substantially outnumber CH calls. Whether this resulted in fewer CH safety calls is unclear. Sixth, we did not capture the financial implications of using CH teams.
Although patient-activated RRTs identified patient safety issues, questions about the utility or necessity of these RRTs remain. In our era of limited hospital resources, the case has not been definitively made that these teams are practical, based on patient outcomes, though other studies have found improved patient satisfaction.7 Most of the RRT calls in our study involved patient dissatisfaction and communication issues. CH may not be the ideal approach for managing these issues, but it represents the last line of patient advocacy once other systems have failed.
We think patient-activated RRTs have the potential to effect patient engagement in safe care. Given the importance of establishing a culture of patient safety and engagement, and increased detection of safety-related events, CH remains active throughout our hospital system. Newer iterations of CH may benefit from stricter language in defining appropriate occasions for calling RRTs, and from descriptions of other resources for patient advocacy within the hospital. These modifications could end up restricting RRT activations to patient complaints and preserving CH resources for patients with safety concerns. Our study lays the groundwork for other institutions that are considering similar interventions. Studies should now start evaluating how well patient- and family-activated RRTs improve patient satisfaction, staff satisfaction, and patient outcomes.
CONCLUSION
Patient- and family-activated RRTs were designed to engage patients and families in safe care. Although CH detects patient safety issues, these are far outnumbered by nonsafety issues. CH demonstrates a commitment to patient engagement and a culture that emphasizes patient safety.
Acknowledgements
This work was presented as a poster at the annual meeting of the Society of Hospital Medicine; March 6-9, 2016; San Diego, CA.
Disclosure
Nothing to report.
References
1. Joint Commission. Improving America’s Hospitals: The Joint Commission’s Annual Report on Quality and Safety 2008. http://www.jointcommission.org/assets/1/6/2008_Annual_Report.pdf. Published November 2008. Accessed May 4, 2016. PubMed 2. Berger Z, Flickinger TE, Pfoh E, Martinez KA, Dy SM. Promoting engagement by patients and families to reduce adverse events in acute care settings: a systematic review. BMJ Qual Saf. 2014;23(7):548-555. PubMed 3. Weingart SN, Zhu J, Chiappetta L, et al. Hospitalized patients’ participation and its impact on quality of care and patient safety. Int J Qual Health Care. 2011;23(3):269-277. PubMed 4. Kennedy P, Pronovost P. Shepherding change: how the market, healthcare providers, and public policy can deliver quality care for the 21st century. Crit Care Med. 2006;34(3 suppl):S1-S6. PubMed 5. Ray EM, Smith R, Massie S, et al. Family alert: implementing direct family activation of a pediatric response team. Jt Comm J Qual Patient Saf. 2009;35(11):575-580. PubMed 6. Dean BS, Decker MJ, Hupp D, Urbach AH, Lewis E, Benes-Stickle J. Condition Help: a pediatric rapid response team triggered by patients and parents. J Healthc Qual. 2008;30(3):28-31. PubMed 7. Vorwerk J, King L. Consumer participation in early detection of the deteriorating patient and call activation to rapid response systems: a literature review. J Clin Nurs. 2015;25(1-2):38-52. PubMed 8. Gerdik C, Vallish RO, Miles K, Godwin SA, Wludyka PS, Panni MK. Successful implementation of a family and patient activated rapid response team in an adult level 1 trauma center. Resuscitation. 2010;81(12):1676-1681. PubMed 9. Hancock KK. From the bedside: purposeful rounding essential to patient experience. Association for Patient Experience website. http://www.patient-experience.org/Resources/Newsletter/Newsletters/Articles/2014/From-the-Bedside-Purposeful-Rounding-Essential-to.aspx. Published February 27, 2014. Accessed July 25, 2016.
References
1. Joint Commission. Improving America’s Hospitals: The Joint Commission’s Annual Report on Quality and Safety 2008. http://www.jointcommission.org/assets/1/6/2008_Annual_Report.pdf. Published November 2008. Accessed May 4, 2016. PubMed 2. Berger Z, Flickinger TE, Pfoh E, Martinez KA, Dy SM. Promoting engagement by patients and families to reduce adverse events in acute care settings: a systematic review. BMJ Qual Saf. 2014;23(7):548-555. PubMed 3. Weingart SN, Zhu J, Chiappetta L, et al. Hospitalized patients’ participation and its impact on quality of care and patient safety. Int J Qual Health Care. 2011;23(3):269-277. PubMed 4. Kennedy P, Pronovost P. Shepherding change: how the market, healthcare providers, and public policy can deliver quality care for the 21st century. Crit Care Med. 2006;34(3 suppl):S1-S6. PubMed 5. Ray EM, Smith R, Massie S, et al. Family alert: implementing direct family activation of a pediatric response team. Jt Comm J Qual Patient Saf. 2009;35(11):575-580. PubMed 6. Dean BS, Decker MJ, Hupp D, Urbach AH, Lewis E, Benes-Stickle J. Condition Help: a pediatric rapid response team triggered by patients and parents. J Healthc Qual. 2008;30(3):28-31. PubMed 7. Vorwerk J, King L. Consumer participation in early detection of the deteriorating patient and call activation to rapid response systems: a literature review. J Clin Nurs. 2015;25(1-2):38-52. PubMed 8. Gerdik C, Vallish RO, Miles K, Godwin SA, Wludyka PS, Panni MK. Successful implementation of a family and patient activated rapid response team in an adult level 1 trauma center. Resuscitation. 2010;81(12):1676-1681. PubMed 9. Hancock KK. From the bedside: purposeful rounding essential to patient experience. Association for Patient Experience website. http://www.patient-experience.org/Resources/Newsletter/Newsletters/Articles/2014/From-the-Bedside-Purposeful-Rounding-Essential-to.aspx. Published February 27, 2014. Accessed July 25, 2016.
Address for correspondence and reprint requests: Elizabeth L. Eden, MD, Internal Medicine Residency Program, 7th Floor, Montefiore Hospital, 200 Lothrop St, Pittsburgh PA 15213; Telephone: 412-802-6648; Fax: 412-692-4499; E-mail: edenel@upmc.edu
Content Gating
Gated (full article locked unless allowed per User)
Millions of individuals with chronic diseases are hospitalized annually in the United States. More than 90% of hospitalized adults have at least 1 chronic disease,1 and almost half of Medicare beneficiaries in the hospital have 4 or more chronic conditions.2 While many patients are admitted for worsening of a single chronic disease, patients are hospitalized more commonly for other causes. For instance, although acute heart failure is among the most frequent causes of hospitalizations among older adults, three-fourths of hospitalizations of patients with heart failure are for reasons other than acute heart failure.3
When a patient with a chronic disease is hospitalized, the inpatient provider must consider whether to actively or passively manage the chronic disease. Studies have suggested that intervening in chronic diseases during hospitalizations can lead to long-term improvement in treatment;4-6 for instance, stroke patients who were started on antihypertensive therapy at discharge were more likely to have their blood pressure controlled in the next year.5 However, some authors have argued that aggressive hypertension management by inpatient providers may result in patient harm.7 One case-based survey suggested that hospitalists were mixed in their interest in participating in chronic disease management in the hospital.8 This study found that providers were less likely to participate in chronic disease management if it was unrelated to the reason for hospitalization.8 However, to our knowledge, no studies have broadly evaluated inpatient provider attitudes, motivating factors, or barriers to participation in chronic disease management.
The purpose of this study was to understand provider attitudes towards chronic disease management for patients who are hospitalized for other causes. We were particularly interested in perceptions of barriers and facilitators to delivery of inpatient chronic disease management. Ultimately, such findings can inform future interventions to improve inpatient care of chronic disease.
METHODS
In this qualitative study, we conducted in-depth interviews with providers to understand attitudes, barriers, and facilitators towards inpatient management of chronic disease; this study was part of a larger studyto implement an electronic health record-based clinical decision-support system intervention to improve quality of care for hospitalized patients with heart failure.
We included providers who care for and can write medication orders for hospitalized adult patients at New York University (NYU) Langone Medical Center, an urban academic medical center. As patients with chronic conditions are commonly hospitalized for many reasons, we sought to interview providers from a range of clinical services without consideration of factors such as frequency of caring for patients with heart failure. We used a purposive sampling framework: we invited participants to ensure a range of services, including medicine, surgery, and neurology, and provider types, including attending physicians, resident physicians, nurse practitioners, and physician assistants. Potential participants, therefore, included all providers for adult hospitalized patients.
We identified potential participants through study team members, referrals from department heads and prior interviewees, and e-mails to department list serves. We did not formally track declinations to being interviewed, although we estimate them as fewer than 20% of providers directly approached. While we focused on inpatient providers at New York University Langone Medical Center, many of the attending physicians and residents spend a portion of their time at the Manhattan Veterans Affairs Hospital and Bellevue Hospital, a safety-net city hospital; providers could have outpatient responsibilities as well.
All participants provided verbal consent to participate. The study was approved by the New York University Institutional Review Board, which granted a waiver of documentation of consent. Participants received a $25 gift card following the interview.
We used a semi-structured interview guide (Appendix) to elicit in-depth accounts of provider attitudes, experiences with, and barriers and facilitators towards chronic disease management in the hospital. The interview began by asking about chronic disease in general and then asked more specific questions about heart failure; we included responses to both groups of questions in the current study. The interview also included questions related to the clinical decision-support system being developed as part of the larger implementation study, although we do not report on these results in the current study. The semi-structured interview guide was informed by the consolidated framework for advancing implementation science (CFIR), which offers an overarching typology for delineating factors that influence guideline implementation;9 we also used CFIR constructs in theme development. We conducted in-depth interviews with providers.
A priori, we estimated 25 interviews would be sufficient to include the purposive sample and achieve data saturation,10 which was reached after 31 interviews. Interviews were held in person or by telephone, at the convenience of the subject. All interviews were transcribed by a professional service. Transcriptions were reviewed against recordings with any mistakes corrected. Prior to each interview, we conducted a brief demographic survey.
Qualitative data were analyzed using a constant comparative analytic technique.11 The investigative team met after reviewing the first 10 interviews and discussed emergent themes from these early transcripts, which led to the initial code list. Two investigators coded the transcripts. Reliability was evaluated by independent coding of a 20% subset of interviews. Differences were reviewed and discussed until consensus was reached. Final intercoder reliability was determined to be greater than 95%.12 All investigators reviewed and refined the code list during the analysis phase. Codes were clustered into themes based on CFIR constructs.9 Analyses were performed using Atlas.ti v. 7 (ATLAS.ti Scientific Software Development GmbH, Berlin, Germany).
Table 1
RESULTS
We conducted interviews with 31 providers. Of these, 12 were on the medicine service, 12 were on the surgery or a surgical subspecialty service, and 7 were on other services; 11 were attending physicians, 12 were resident physicians, 5 were NPs, and 3 were PAs. Only 2 providers—an attending in medicine and a resident in surgery—had a specialty focus that was cardiac-related. Median time in current position was 4 years (Table 1). Seventeen of the interviews were in person, and 14 were conducted by telephone. The mean interview time was 20 minutes and ranged from 11 to 41 minutes.
Table 2
We identified 5 main themes with 29 supporting codes (Table 2) describing provider attitudes towards the management of chronic disease for hospitalized patients. These themes, with related CFIR constructs, were: 1) perceived impact on patient outcomes (CFIR construct: intervention characteristics, relative advantage); 2) hospital structural characteristics (inner setting, structural characteristics); 3) provider knowledge and self-efficacy (characteristic of individual, knowledge and beliefs about the intervention and self-efficacy); 4) hospital priorities (inner setting, implementation climate, relative priority); and 5) continuity and communication (inner setting, networks and communications). For most themes, subjects described both positive and negative aspects of chronic disease management, as well as related facilitators and barriers to delivery of chronic disease care for hospitalized patients. Illustrative quotes for each theme are shown in Table 3.
Perceived Impact on Patient Outcomes
Perceived impact on patient outcomes was mixed. Most providers believed the management of chronic diseases could lead to improvement in important patient outcomes, including decreased length of stay (LOS), prevention of hospital complication, and decreased readmissions. Surgical providers focused particularly on the benefits of preventing surgical complications and noted that they were more likely to manage chronic conditions—primarily through use of specialist consultation—when they perceived a benefit to prevention of surgical outcomes or a fear that surgery may worsen a stable chronic condition:
“Most of the surgery I do is pretty stressful on the body and is very likely to induce acute on chronic exacerbations of heart failure. For someone with Class II or higher heart failure, I’m definitely gonna have cardiology on board or at least internal medicine on board right from the beginning.”
Table 3 However, some providers acknowledged that there were potential risks to such management, including “prolonging hospital stays for nonemergent indications” and treatment with therapies that had previously led to an “adverse reaction that wasn’t clearly documented.” Providers were also concerned that treating chronic conditions may take focus away from acute conditions, which could lead to worse patient-centered outcomes. One attending in medicine described it:
“If you do potentially focus on those chronic issues, and there’s already a lot of other stuff going on with the patient, you might not be prioritizing the patient’s active issues appropriately. The patient’s saying, ‘I’m in pain. I’m in pain. I’m in pain,’ and you’re saying, ‘Thank you very much. Look, your heart failure, you didn’t get your beta-blocker.’ There could be a disconnect between patient’s goals, expectations, and your goals and expectations.”
Hospital Structural Characteristics
For many providers, the hospital setting provides a unique opportunity for care of patients with chronic disease. First, a hospitalization is a time for a patient’s management to be reviewed by a new care team. The hospital team reviews the management plan for patients at admission, which is a time to reevaluate whether patients are on evidence-based therapies: “It’s helpful to have a new set of eyes on somebody, like fresh information.” According to providers, this reevaluation can overcome instances of therapeutic inertia by the outpatient physician. Second, the hospital has many resources, including readily available specialist services and diagnostic tests, which can allow a patient-centered approach that coordinates care in 1 place, as a surgery NP described: “I think the advantage for the patient is that they wind up stopping in for 1 thing but we wind up taking care of a few without requiring the need for him or her to go to all these different specialists on the outside. They’re mostly elderly and not able to get around.” Third, the high availability of services and frequent monitoring allows rapid titration of evidence-based medicines, as discussed by a medicine resident: “It’s easier and faster to titrate medication—they’re in a monitored setting; you can ensure compliance.”
Patients may also differ from their usual state while hospitalized, creating both risks and benefits. The hospital setting can provide an opportunity to educate patients on their chronic disease(s) because they are motivated: “They’re in an office visit and their sugars are out of whack or something, they may take it a little bit more seriously if they were just in the hospital even though it was on an unrelated issue. I think it probably just changes their perspective on their disease.” However, in the hospital, patients are in an unusual environment with a restricted diet and forced medication compliance. Furthermore, the acute condition can lead to changes in their chronic disease, as described by 1 medicine attending: “their sugar is high because they’re acutely ill.” Providers expressed concern that changing medications in this setting may lead to adverse events (AEs) when patients return to their usual environment.
Provider Knowledge and Self-Efficacy
Insufficient knowledge of treatments for chronic conditions was cited as a barrier to some providers’ ability to actively manage chronic disease for hospitalized patients. Some providers described management of conditions outside their area as less satisfying than their primary focus. For example, an orthopedic surgeon explained: “…it’s very simple. You see your bone is broken, you fix it, that’s it…it’s intellectually satisfying…managing chronic diseases is less like that.” Reliance on consultants was 1 approach to deal with knowledge gaps in areas outside a provider’s expertise.
For a number of providers, management of stable chronic disease is the responsibility of the outpatient provider. Providers expressed concern that inpatient management was a reach into the domain of the primary care provider (PCP) and might take “away from the primary focus” of the hospitalization. Nonetheless, some providers noted an “ethical responsibility to manage [a] patient correctly,” and some providers believed that engaging in chronic disease management in the hospital would present an opportunity to expand their own expertise.
A few providers were worried about legal risk related to chronic disease management: “we don’t typically deal too much with managing some of these other medical issues for medical and legal reasons.” Providers again suggested that consults can help overcome this concern for risk, as discussed by 1 surgical attending: “We’re all not wanting to be sued, and we want to do the right thing. It costs me nothing to have a cardiologist on board, so like—why not.”
Hospital Priorities
Providers explained that the hospital has strong interests in early discharge and minimizing LOS. These priorities are based on goals of improving patient outcomes, increasing bed availability and hospital volume, and reducing costs. Providers perceive these hospital priorities as potential barriers to chronic disease management, which can increase LOS and costs through additional testing and treatment. As a medicine resident described: “The DBN philosophy, ‘discharge before noon’ philosophy, which is part of the hospital efficiency to get people in and out of the hospital as quickly as [is] safe, or maybe faster. And I think that there’s a culture where you’re encouraged to only focus on the acute issue and tend to defer everything else.”
Continuity and Communication
According to many providers, care continuity between the outpatient setting and the hospital played a major role in management of chronic disease. One barrier to starting a new evidence-based medication was lack of knowledge of patient history. As noted, providers expressed concern that a patient may not be on a given therapy because of an adverse reaction that was not documented in the hospital chart. This is particularly true because, as discussed by a surgery resident, patients with “PCPs outside the system [in which providers] don’t have access to the electronic medical record.” To overcome this barrier, providers attempt to communicate with the outpatient provider to confirm a lack of contraindications to therapies prior to any changes; notably, communication is easier if the inpatient provider has a relationship with the outpatient PCP.
Some providers were more likely to start chronic disease therapies if the patient had no prior outpatient care, because the provider was reassured that there was no rationale for missing therapies. One neurology attending noted that if a patient had newly documented “hypertension even if they were in for something else, I might start them on an antihypertensive, but then arrange for a close follow-up with a new PCP.”
Following hospitalization, providers wanted assurance that any changes to chronic disease management would be followed up by an outpatient physician. Any changes are relayed to the outpatient provider and the “level of communication…with the outpatient provider who’s gonna inherit” these changes can influence how aggressively the inpatient provider manages chronic diseases. Providers may be reluctant to start therapy for patients if they are concerned about outpatient follow up: “they have diabetes and they should really technically be on an ACE [angiotensin converting enzyme]inhibitor and aspirin, but they’re not. I might send them out on the aspirin but I might either start ACE inhibitor and have them follow up with their PCP in 2 weeks if I’m confident that they’ll do it or if I’m really confident that they’ll not follow up, I will help them get the appointment and then the discharge instruction is to the PCP is ‘Please start this patient on ACE inhibitor if they show up.’”
DISCUSSION
Providers frequently perceive benefit to chronic disease management in the hospital, including improvements in clinical outcomes. Notably, providers see opportunities to improve compliance with evidence-based care to overcome potential barriers to managing chronic disease in the outpatient setting, which can be limited by pressure for brief encounters,13 clinical inertia,14 difficulty with close monitoring of patients,15 and care fragmentation.16 Concurrently, inpatient providers are concerned about potential for patient harm related to chronic disease management, primarily related to AEs from medications. Similar to a case study about a patient with outpatient hypotension following aggressive inpatient hypertension management,7 providers fear that changing a patient’s chronic disease management in a hospital setting may cause harm when the patient returns home.
Although some clinicians have argued against aggressive in-hospital chronic disease management because of concerns for risk of AEs,7 our study and others8 have suggested that many clinicians perceive benefit. In some cases, such as smoking cessation counseling for all current smokers and prescribing an angiotensin converting enzyme inhibitor for patients with systolic heart failure, the perceived importance is so great that chronic disease management has been used as a national quality metric for hospitals. While these hospital metrics may be justified for short-term benefits after hospitalization, studies have demonstrated only weak improvement in short-term postdischarge outcomes related to chronic disease management.17 The true benefit is likely from improved processes of care in the short term that lead to long-term improvement in outcomes.4,5,18 Thus, the advantage of starting a patient hospitalized for a stroke on blood pressure medication is the increased likelihood that the patient will continue the medication as an outpatient, which may reduce long-term mortality.
For hospital delivery systems that are concerned with such care process improvement through in-hospital chronic disease management, we identified a number of barriers and facilitators to delivering this care. One significant barrier was poor transitions between the inpatient and the outpatient settings. When a patient transitions into the hospital, providers need to understand prior management choices. Facilitators to help inpatient providers understand prior management included either knowing the outpatient provider, or understanding that there was a lack of regular outpatient care; in both these cases, inpatient providers felt more comfortable managing chronic diseases because they had insight into the outpatient plan, or lack thereof. However, these facilitators may not be practical to incorporate in interventions to improve chronic disease care, which should consider overcoming these communication barriers. Use of shared electronic health records or standardized telephone calls with well-documented care plans obtained through health information exchanges may facilitate an inpatient provider to manage appropriately chronic disease. Similarly, discontinuity between the inpatient provider and the outpatient provider is a barrier that must be overcome to ease concerns that any chronic disease management changes do not result in harm in the postdischarge period. These findings again point to the need for improved documentation and communication between inpatient and outpatient providers. Of course, the transitional care period is one of high risk, and improving communication between providers has been an area of ongoing work.19
Lack of comfort among inpatient providers with managing chronic diseases is another important barrier, which appears to be largely overcome through the use of consultation services. Ready availability of specialists, common in academic medical centers, can facilitate delivery of chronic disease management. Inpatient interventions designed to improve evidence-based care for a chronic disease may benefit from involvement or at least availability of specialists in the effort. Another major barrier relates to hospital priorities, which in our study were closely aligned with external factors such as payment models. As hospitalizations are typically paid based on the discharge diagnosis, hospitals have incentives to discharge quickly and not order extra diagnostic tests. As a result, there are disincentives for chronic disease management that may require additional testing or monitoring in the hospital. Conversely, as hospitals accept postdischarge financial risks through readmission penalties or postdischarge cost savings, hospitals may perceive that long-term benefits of chronic disease management may outweigh short-term costs.
The study findings should be interpreted in the context of its limitations. Findings of our study of providers from a single academic medical center may not be generalizable. Nearly half of our interviews were conducted by telephone, which limits our ability to capture nonverbal cues in communication. Providers may have had social desirability bias towards positive aspects of chronic disease management. We did not have the power to determine differences in response by provider characteristic because this was an exploratory qualitative study. Future studies with representative sampling, a larger sample size, and measures for constructs such as provider self-efficacy are needed to examine differences by specialty, provider type, and experience level.
In conclusion, inpatient providers believe that hospital chronic disease management has the potential to be beneficial for both process of care and clinical outcomes; providers also express concern about potential adverse consequences of managing chronic disease during acute hospitalizations. To maximize both quality of care and patient safety, overcoming communication barriers between inpatient and outpatient providers is needed. Both a supportive hospital environment and availability of specialty support can facilitate in-hospital chronic disease management. Interventions that incorporate these factors may be well-suited to improve chronic disease care and long-term outcomes.
Disclosures
This work was supported by the Agency for Healthcare Research and Quality (AHRQ) grant K08HS23683. The authors report no financial conflicts of interest.
1. Friedman B, Jiang HJ, Elixhauser A, Segal A. Hospital inpatient costs for adults with multiple chronic conditions. Med Care Res Rev. 2006;63(3):327-346. PubMed 2. Steiner CA, Friedman B. Hospital utilization, costs, and mortality for adults with multiple chronic conditions, Nationwide Inpatient Sample, 2009. Prev Chronic Dis. 2013;10:E62. PubMed 3. Blecker S, Paul M, Taksler G, Ogedegbe G, Katz S. Heart failure-associated hospitalizations in the United States. J Am Coll Cardiol. 2013;61(12):1259-1267. PubMed 4. Fonarow GC. Role of in-hospital initiation of carvedilol to improve treatment rates and clinical outcomes. Am J Cardiol. 2004;93(9A):77B-81B. PubMed 5. Touze E, Coste J, Voicu M, et al. Importance of in-hospital initiation of therapies and therapeutic inertia in secondary stroke prevention: IMplementation of Prevention After a Cerebrovascular evenT (IMPACT) Study. Stroke. 2008;39(6):1834-1843. PubMed 6. Ovbiagele B, Saver JL, Fredieu A, et al. In-hospital initiation of secondary stroke prevention therapies yields high rates of adherence at follow-up. Stroke. 2004;35(12):2879-2883. PubMed 7. Steinman MA, Auerbach AD. Managing chronic disease in hospitalized patients. JAMA Intern Med. 2013;173(20):1857-1858. PubMed 8. Breu AC, Allen-Dicker J, Mueller S, Palamara K, Hinami K, Herzig SJ. Hospitalist and primary care physician perspectives on medication management of chronic conditions for hospitalized patients. J Hosp Med. 2014;9(5):303-309. PubMed 9. Damschroder LJ, Aron DC, Keith RE, Kirsh SR, Alexander JA, Lowery JC. Fostering implementation of health services research findings into practice: a consolidated framework for advancing implementation science. Implement Sci. 2009;4:50. PubMed 10. Morse JM. The significance of saturation. Qualitative Health Research. 1995;5(2):147-149. 11. Bradley EH, Curry LA, Devers KJ. Qualitative data analysis for health services research: developing taxonomy, themes, and theory. Health Services Research. 2007;42(4):1758-1772. PubMed 12. Riegel B, Dickson VV, Topaz M. Qualitative analysis of naturalistic decision making in adults with chronic heart failure. Nurs Res. 2013;62(2):91-98. PubMed 13. Linzer M, Konrad TR, Douglas J, et al. Managed care, time pressure, and physician job satisfaction: results from the physician worklife study. J Gen Intern Med. 2000;15(7):441-450. PubMed 14. Phillips LS, Branch WT, Cook CB, et al. Clinical inertia. Ann Intern Med. 2001;135(9):825-834. 15. Dev S, Hoffman TK, Kavalieratos D, et al. Barriers to adoption of mineralocorticoid receptor antagonists in patients with heart failure: A mixed-methods study. J Am Heart Assoc. 2016;4(3):e002493. PubMed 16. Stange KC. The problem of fragmentation and the need for integrative solutions. Ann Fam Med. 2009;7(2):100-103. PubMed 17. Fonarow GC, Abraham WT, Albert NM, et al. Association between performance measures and clinical outcomes for patients hospitalized with heart failure. JAMA. 2007;297(1):61-70. PubMed 18. Shah M, Norwood CA, Farias S, Ibrahim S, Chong PH, Fogelfeld L. Diabetes transitional care from inpatient to outpatient setting: pharmacist discharge counseling. J Pharm Pract. 2013;26(2):120-124. PubMed 19. Kripalani S, LeFevre F, Phillips CO, Williams MV, Basaviah P, Baker DW. Deficits in communication and information transfer between hospital-based and primary care physicians: implications for patient safety and continuity of care. JAMA. 2007;297(8):831-841. PubMed
Millions of individuals with chronic diseases are hospitalized annually in the United States. More than 90% of hospitalized adults have at least 1 chronic disease,1 and almost half of Medicare beneficiaries in the hospital have 4 or more chronic conditions.2 While many patients are admitted for worsening of a single chronic disease, patients are hospitalized more commonly for other causes. For instance, although acute heart failure is among the most frequent causes of hospitalizations among older adults, three-fourths of hospitalizations of patients with heart failure are for reasons other than acute heart failure.3
When a patient with a chronic disease is hospitalized, the inpatient provider must consider whether to actively or passively manage the chronic disease. Studies have suggested that intervening in chronic diseases during hospitalizations can lead to long-term improvement in treatment;4-6 for instance, stroke patients who were started on antihypertensive therapy at discharge were more likely to have their blood pressure controlled in the next year.5 However, some authors have argued that aggressive hypertension management by inpatient providers may result in patient harm.7 One case-based survey suggested that hospitalists were mixed in their interest in participating in chronic disease management in the hospital.8 This study found that providers were less likely to participate in chronic disease management if it was unrelated to the reason for hospitalization.8 However, to our knowledge, no studies have broadly evaluated inpatient provider attitudes, motivating factors, or barriers to participation in chronic disease management.
The purpose of this study was to understand provider attitudes towards chronic disease management for patients who are hospitalized for other causes. We were particularly interested in perceptions of barriers and facilitators to delivery of inpatient chronic disease management. Ultimately, such findings can inform future interventions to improve inpatient care of chronic disease.
METHODS
In this qualitative study, we conducted in-depth interviews with providers to understand attitudes, barriers, and facilitators towards inpatient management of chronic disease; this study was part of a larger studyto implement an electronic health record-based clinical decision-support system intervention to improve quality of care for hospitalized patients with heart failure.
We included providers who care for and can write medication orders for hospitalized adult patients at New York University (NYU) Langone Medical Center, an urban academic medical center. As patients with chronic conditions are commonly hospitalized for many reasons, we sought to interview providers from a range of clinical services without consideration of factors such as frequency of caring for patients with heart failure. We used a purposive sampling framework: we invited participants to ensure a range of services, including medicine, surgery, and neurology, and provider types, including attending physicians, resident physicians, nurse practitioners, and physician assistants. Potential participants, therefore, included all providers for adult hospitalized patients.
We identified potential participants through study team members, referrals from department heads and prior interviewees, and e-mails to department list serves. We did not formally track declinations to being interviewed, although we estimate them as fewer than 20% of providers directly approached. While we focused on inpatient providers at New York University Langone Medical Center, many of the attending physicians and residents spend a portion of their time at the Manhattan Veterans Affairs Hospital and Bellevue Hospital, a safety-net city hospital; providers could have outpatient responsibilities as well.
All participants provided verbal consent to participate. The study was approved by the New York University Institutional Review Board, which granted a waiver of documentation of consent. Participants received a $25 gift card following the interview.
We used a semi-structured interview guide (Appendix) to elicit in-depth accounts of provider attitudes, experiences with, and barriers and facilitators towards chronic disease management in the hospital. The interview began by asking about chronic disease in general and then asked more specific questions about heart failure; we included responses to both groups of questions in the current study. The interview also included questions related to the clinical decision-support system being developed as part of the larger implementation study, although we do not report on these results in the current study. The semi-structured interview guide was informed by the consolidated framework for advancing implementation science (CFIR), which offers an overarching typology for delineating factors that influence guideline implementation;9 we also used CFIR constructs in theme development. We conducted in-depth interviews with providers.
A priori, we estimated 25 interviews would be sufficient to include the purposive sample and achieve data saturation,10 which was reached after 31 interviews. Interviews were held in person or by telephone, at the convenience of the subject. All interviews were transcribed by a professional service. Transcriptions were reviewed against recordings with any mistakes corrected. Prior to each interview, we conducted a brief demographic survey.
Qualitative data were analyzed using a constant comparative analytic technique.11 The investigative team met after reviewing the first 10 interviews and discussed emergent themes from these early transcripts, which led to the initial code list. Two investigators coded the transcripts. Reliability was evaluated by independent coding of a 20% subset of interviews. Differences were reviewed and discussed until consensus was reached. Final intercoder reliability was determined to be greater than 95%.12 All investigators reviewed and refined the code list during the analysis phase. Codes were clustered into themes based on CFIR constructs.9 Analyses were performed using Atlas.ti v. 7 (ATLAS.ti Scientific Software Development GmbH, Berlin, Germany).
Table 1
RESULTS
We conducted interviews with 31 providers. Of these, 12 were on the medicine service, 12 were on the surgery or a surgical subspecialty service, and 7 were on other services; 11 were attending physicians, 12 were resident physicians, 5 were NPs, and 3 were PAs. Only 2 providers—an attending in medicine and a resident in surgery—had a specialty focus that was cardiac-related. Median time in current position was 4 years (Table 1). Seventeen of the interviews were in person, and 14 were conducted by telephone. The mean interview time was 20 minutes and ranged from 11 to 41 minutes.
Table 2
We identified 5 main themes with 29 supporting codes (Table 2) describing provider attitudes towards the management of chronic disease for hospitalized patients. These themes, with related CFIR constructs, were: 1) perceived impact on patient outcomes (CFIR construct: intervention characteristics, relative advantage); 2) hospital structural characteristics (inner setting, structural characteristics); 3) provider knowledge and self-efficacy (characteristic of individual, knowledge and beliefs about the intervention and self-efficacy); 4) hospital priorities (inner setting, implementation climate, relative priority); and 5) continuity and communication (inner setting, networks and communications). For most themes, subjects described both positive and negative aspects of chronic disease management, as well as related facilitators and barriers to delivery of chronic disease care for hospitalized patients. Illustrative quotes for each theme are shown in Table 3.
Perceived Impact on Patient Outcomes
Perceived impact on patient outcomes was mixed. Most providers believed the management of chronic diseases could lead to improvement in important patient outcomes, including decreased length of stay (LOS), prevention of hospital complication, and decreased readmissions. Surgical providers focused particularly on the benefits of preventing surgical complications and noted that they were more likely to manage chronic conditions—primarily through use of specialist consultation—when they perceived a benefit to prevention of surgical outcomes or a fear that surgery may worsen a stable chronic condition:
“Most of the surgery I do is pretty stressful on the body and is very likely to induce acute on chronic exacerbations of heart failure. For someone with Class II or higher heart failure, I’m definitely gonna have cardiology on board or at least internal medicine on board right from the beginning.”
Table 3 However, some providers acknowledged that there were potential risks to such management, including “prolonging hospital stays for nonemergent indications” and treatment with therapies that had previously led to an “adverse reaction that wasn’t clearly documented.” Providers were also concerned that treating chronic conditions may take focus away from acute conditions, which could lead to worse patient-centered outcomes. One attending in medicine described it:
“If you do potentially focus on those chronic issues, and there’s already a lot of other stuff going on with the patient, you might not be prioritizing the patient’s active issues appropriately. The patient’s saying, ‘I’m in pain. I’m in pain. I’m in pain,’ and you’re saying, ‘Thank you very much. Look, your heart failure, you didn’t get your beta-blocker.’ There could be a disconnect between patient’s goals, expectations, and your goals and expectations.”
Hospital Structural Characteristics
For many providers, the hospital setting provides a unique opportunity for care of patients with chronic disease. First, a hospitalization is a time for a patient’s management to be reviewed by a new care team. The hospital team reviews the management plan for patients at admission, which is a time to reevaluate whether patients are on evidence-based therapies: “It’s helpful to have a new set of eyes on somebody, like fresh information.” According to providers, this reevaluation can overcome instances of therapeutic inertia by the outpatient physician. Second, the hospital has many resources, including readily available specialist services and diagnostic tests, which can allow a patient-centered approach that coordinates care in 1 place, as a surgery NP described: “I think the advantage for the patient is that they wind up stopping in for 1 thing but we wind up taking care of a few without requiring the need for him or her to go to all these different specialists on the outside. They’re mostly elderly and not able to get around.” Third, the high availability of services and frequent monitoring allows rapid titration of evidence-based medicines, as discussed by a medicine resident: “It’s easier and faster to titrate medication—they’re in a monitored setting; you can ensure compliance.”
Patients may also differ from their usual state while hospitalized, creating both risks and benefits. The hospital setting can provide an opportunity to educate patients on their chronic disease(s) because they are motivated: “They’re in an office visit and their sugars are out of whack or something, they may take it a little bit more seriously if they were just in the hospital even though it was on an unrelated issue. I think it probably just changes their perspective on their disease.” However, in the hospital, patients are in an unusual environment with a restricted diet and forced medication compliance. Furthermore, the acute condition can lead to changes in their chronic disease, as described by 1 medicine attending: “their sugar is high because they’re acutely ill.” Providers expressed concern that changing medications in this setting may lead to adverse events (AEs) when patients return to their usual environment.
Provider Knowledge and Self-Efficacy
Insufficient knowledge of treatments for chronic conditions was cited as a barrier to some providers’ ability to actively manage chronic disease for hospitalized patients. Some providers described management of conditions outside their area as less satisfying than their primary focus. For example, an orthopedic surgeon explained: “…it’s very simple. You see your bone is broken, you fix it, that’s it…it’s intellectually satisfying…managing chronic diseases is less like that.” Reliance on consultants was 1 approach to deal with knowledge gaps in areas outside a provider’s expertise.
For a number of providers, management of stable chronic disease is the responsibility of the outpatient provider. Providers expressed concern that inpatient management was a reach into the domain of the primary care provider (PCP) and might take “away from the primary focus” of the hospitalization. Nonetheless, some providers noted an “ethical responsibility to manage [a] patient correctly,” and some providers believed that engaging in chronic disease management in the hospital would present an opportunity to expand their own expertise.
A few providers were worried about legal risk related to chronic disease management: “we don’t typically deal too much with managing some of these other medical issues for medical and legal reasons.” Providers again suggested that consults can help overcome this concern for risk, as discussed by 1 surgical attending: “We’re all not wanting to be sued, and we want to do the right thing. It costs me nothing to have a cardiologist on board, so like—why not.”
Hospital Priorities
Providers explained that the hospital has strong interests in early discharge and minimizing LOS. These priorities are based on goals of improving patient outcomes, increasing bed availability and hospital volume, and reducing costs. Providers perceive these hospital priorities as potential barriers to chronic disease management, which can increase LOS and costs through additional testing and treatment. As a medicine resident described: “The DBN philosophy, ‘discharge before noon’ philosophy, which is part of the hospital efficiency to get people in and out of the hospital as quickly as [is] safe, or maybe faster. And I think that there’s a culture where you’re encouraged to only focus on the acute issue and tend to defer everything else.”
Continuity and Communication
According to many providers, care continuity between the outpatient setting and the hospital played a major role in management of chronic disease. One barrier to starting a new evidence-based medication was lack of knowledge of patient history. As noted, providers expressed concern that a patient may not be on a given therapy because of an adverse reaction that was not documented in the hospital chart. This is particularly true because, as discussed by a surgery resident, patients with “PCPs outside the system [in which providers] don’t have access to the electronic medical record.” To overcome this barrier, providers attempt to communicate with the outpatient provider to confirm a lack of contraindications to therapies prior to any changes; notably, communication is easier if the inpatient provider has a relationship with the outpatient PCP.
Some providers were more likely to start chronic disease therapies if the patient had no prior outpatient care, because the provider was reassured that there was no rationale for missing therapies. One neurology attending noted that if a patient had newly documented “hypertension even if they were in for something else, I might start them on an antihypertensive, but then arrange for a close follow-up with a new PCP.”
Following hospitalization, providers wanted assurance that any changes to chronic disease management would be followed up by an outpatient physician. Any changes are relayed to the outpatient provider and the “level of communication…with the outpatient provider who’s gonna inherit” these changes can influence how aggressively the inpatient provider manages chronic diseases. Providers may be reluctant to start therapy for patients if they are concerned about outpatient follow up: “they have diabetes and they should really technically be on an ACE [angiotensin converting enzyme]inhibitor and aspirin, but they’re not. I might send them out on the aspirin but I might either start ACE inhibitor and have them follow up with their PCP in 2 weeks if I’m confident that they’ll do it or if I’m really confident that they’ll not follow up, I will help them get the appointment and then the discharge instruction is to the PCP is ‘Please start this patient on ACE inhibitor if they show up.’”
DISCUSSION
Providers frequently perceive benefit to chronic disease management in the hospital, including improvements in clinical outcomes. Notably, providers see opportunities to improve compliance with evidence-based care to overcome potential barriers to managing chronic disease in the outpatient setting, which can be limited by pressure for brief encounters,13 clinical inertia,14 difficulty with close monitoring of patients,15 and care fragmentation.16 Concurrently, inpatient providers are concerned about potential for patient harm related to chronic disease management, primarily related to AEs from medications. Similar to a case study about a patient with outpatient hypotension following aggressive inpatient hypertension management,7 providers fear that changing a patient’s chronic disease management in a hospital setting may cause harm when the patient returns home.
Although some clinicians have argued against aggressive in-hospital chronic disease management because of concerns for risk of AEs,7 our study and others8 have suggested that many clinicians perceive benefit. In some cases, such as smoking cessation counseling for all current smokers and prescribing an angiotensin converting enzyme inhibitor for patients with systolic heart failure, the perceived importance is so great that chronic disease management has been used as a national quality metric for hospitals. While these hospital metrics may be justified for short-term benefits after hospitalization, studies have demonstrated only weak improvement in short-term postdischarge outcomes related to chronic disease management.17 The true benefit is likely from improved processes of care in the short term that lead to long-term improvement in outcomes.4,5,18 Thus, the advantage of starting a patient hospitalized for a stroke on blood pressure medication is the increased likelihood that the patient will continue the medication as an outpatient, which may reduce long-term mortality.
For hospital delivery systems that are concerned with such care process improvement through in-hospital chronic disease management, we identified a number of barriers and facilitators to delivering this care. One significant barrier was poor transitions between the inpatient and the outpatient settings. When a patient transitions into the hospital, providers need to understand prior management choices. Facilitators to help inpatient providers understand prior management included either knowing the outpatient provider, or understanding that there was a lack of regular outpatient care; in both these cases, inpatient providers felt more comfortable managing chronic diseases because they had insight into the outpatient plan, or lack thereof. However, these facilitators may not be practical to incorporate in interventions to improve chronic disease care, which should consider overcoming these communication barriers. Use of shared electronic health records or standardized telephone calls with well-documented care plans obtained through health information exchanges may facilitate an inpatient provider to manage appropriately chronic disease. Similarly, discontinuity between the inpatient provider and the outpatient provider is a barrier that must be overcome to ease concerns that any chronic disease management changes do not result in harm in the postdischarge period. These findings again point to the need for improved documentation and communication between inpatient and outpatient providers. Of course, the transitional care period is one of high risk, and improving communication between providers has been an area of ongoing work.19
Lack of comfort among inpatient providers with managing chronic diseases is another important barrier, which appears to be largely overcome through the use of consultation services. Ready availability of specialists, common in academic medical centers, can facilitate delivery of chronic disease management. Inpatient interventions designed to improve evidence-based care for a chronic disease may benefit from involvement or at least availability of specialists in the effort. Another major barrier relates to hospital priorities, which in our study were closely aligned with external factors such as payment models. As hospitalizations are typically paid based on the discharge diagnosis, hospitals have incentives to discharge quickly and not order extra diagnostic tests. As a result, there are disincentives for chronic disease management that may require additional testing or monitoring in the hospital. Conversely, as hospitals accept postdischarge financial risks through readmission penalties or postdischarge cost savings, hospitals may perceive that long-term benefits of chronic disease management may outweigh short-term costs.
The study findings should be interpreted in the context of its limitations. Findings of our study of providers from a single academic medical center may not be generalizable. Nearly half of our interviews were conducted by telephone, which limits our ability to capture nonverbal cues in communication. Providers may have had social desirability bias towards positive aspects of chronic disease management. We did not have the power to determine differences in response by provider characteristic because this was an exploratory qualitative study. Future studies with representative sampling, a larger sample size, and measures for constructs such as provider self-efficacy are needed to examine differences by specialty, provider type, and experience level.
In conclusion, inpatient providers believe that hospital chronic disease management has the potential to be beneficial for both process of care and clinical outcomes; providers also express concern about potential adverse consequences of managing chronic disease during acute hospitalizations. To maximize both quality of care and patient safety, overcoming communication barriers between inpatient and outpatient providers is needed. Both a supportive hospital environment and availability of specialty support can facilitate in-hospital chronic disease management. Interventions that incorporate these factors may be well-suited to improve chronic disease care and long-term outcomes.
Disclosures
This work was supported by the Agency for Healthcare Research and Quality (AHRQ) grant K08HS23683. The authors report no financial conflicts of interest.
Millions of individuals with chronic diseases are hospitalized annually in the United States. More than 90% of hospitalized adults have at least 1 chronic disease,1 and almost half of Medicare beneficiaries in the hospital have 4 or more chronic conditions.2 While many patients are admitted for worsening of a single chronic disease, patients are hospitalized more commonly for other causes. For instance, although acute heart failure is among the most frequent causes of hospitalizations among older adults, three-fourths of hospitalizations of patients with heart failure are for reasons other than acute heart failure.3
When a patient with a chronic disease is hospitalized, the inpatient provider must consider whether to actively or passively manage the chronic disease. Studies have suggested that intervening in chronic diseases during hospitalizations can lead to long-term improvement in treatment;4-6 for instance, stroke patients who were started on antihypertensive therapy at discharge were more likely to have their blood pressure controlled in the next year.5 However, some authors have argued that aggressive hypertension management by inpatient providers may result in patient harm.7 One case-based survey suggested that hospitalists were mixed in their interest in participating in chronic disease management in the hospital.8 This study found that providers were less likely to participate in chronic disease management if it was unrelated to the reason for hospitalization.8 However, to our knowledge, no studies have broadly evaluated inpatient provider attitudes, motivating factors, or barriers to participation in chronic disease management.
The purpose of this study was to understand provider attitudes towards chronic disease management for patients who are hospitalized for other causes. We were particularly interested in perceptions of barriers and facilitators to delivery of inpatient chronic disease management. Ultimately, such findings can inform future interventions to improve inpatient care of chronic disease.
METHODS
In this qualitative study, we conducted in-depth interviews with providers to understand attitudes, barriers, and facilitators towards inpatient management of chronic disease; this study was part of a larger studyto implement an electronic health record-based clinical decision-support system intervention to improve quality of care for hospitalized patients with heart failure.
We included providers who care for and can write medication orders for hospitalized adult patients at New York University (NYU) Langone Medical Center, an urban academic medical center. As patients with chronic conditions are commonly hospitalized for many reasons, we sought to interview providers from a range of clinical services without consideration of factors such as frequency of caring for patients with heart failure. We used a purposive sampling framework: we invited participants to ensure a range of services, including medicine, surgery, and neurology, and provider types, including attending physicians, resident physicians, nurse practitioners, and physician assistants. Potential participants, therefore, included all providers for adult hospitalized patients.
We identified potential participants through study team members, referrals from department heads and prior interviewees, and e-mails to department list serves. We did not formally track declinations to being interviewed, although we estimate them as fewer than 20% of providers directly approached. While we focused on inpatient providers at New York University Langone Medical Center, many of the attending physicians and residents spend a portion of their time at the Manhattan Veterans Affairs Hospital and Bellevue Hospital, a safety-net city hospital; providers could have outpatient responsibilities as well.
All participants provided verbal consent to participate. The study was approved by the New York University Institutional Review Board, which granted a waiver of documentation of consent. Participants received a $25 gift card following the interview.
We used a semi-structured interview guide (Appendix) to elicit in-depth accounts of provider attitudes, experiences with, and barriers and facilitators towards chronic disease management in the hospital. The interview began by asking about chronic disease in general and then asked more specific questions about heart failure; we included responses to both groups of questions in the current study. The interview also included questions related to the clinical decision-support system being developed as part of the larger implementation study, although we do not report on these results in the current study. The semi-structured interview guide was informed by the consolidated framework for advancing implementation science (CFIR), which offers an overarching typology for delineating factors that influence guideline implementation;9 we also used CFIR constructs in theme development. We conducted in-depth interviews with providers.
A priori, we estimated 25 interviews would be sufficient to include the purposive sample and achieve data saturation,10 which was reached after 31 interviews. Interviews were held in person or by telephone, at the convenience of the subject. All interviews were transcribed by a professional service. Transcriptions were reviewed against recordings with any mistakes corrected. Prior to each interview, we conducted a brief demographic survey.
Qualitative data were analyzed using a constant comparative analytic technique.11 The investigative team met after reviewing the first 10 interviews and discussed emergent themes from these early transcripts, which led to the initial code list. Two investigators coded the transcripts. Reliability was evaluated by independent coding of a 20% subset of interviews. Differences were reviewed and discussed until consensus was reached. Final intercoder reliability was determined to be greater than 95%.12 All investigators reviewed and refined the code list during the analysis phase. Codes were clustered into themes based on CFIR constructs.9 Analyses were performed using Atlas.ti v. 7 (ATLAS.ti Scientific Software Development GmbH, Berlin, Germany).
Table 1
RESULTS
We conducted interviews with 31 providers. Of these, 12 were on the medicine service, 12 were on the surgery or a surgical subspecialty service, and 7 were on other services; 11 were attending physicians, 12 were resident physicians, 5 were NPs, and 3 were PAs. Only 2 providers—an attending in medicine and a resident in surgery—had a specialty focus that was cardiac-related. Median time in current position was 4 years (Table 1). Seventeen of the interviews were in person, and 14 were conducted by telephone. The mean interview time was 20 minutes and ranged from 11 to 41 minutes.
Table 2
We identified 5 main themes with 29 supporting codes (Table 2) describing provider attitudes towards the management of chronic disease for hospitalized patients. These themes, with related CFIR constructs, were: 1) perceived impact on patient outcomes (CFIR construct: intervention characteristics, relative advantage); 2) hospital structural characteristics (inner setting, structural characteristics); 3) provider knowledge and self-efficacy (characteristic of individual, knowledge and beliefs about the intervention and self-efficacy); 4) hospital priorities (inner setting, implementation climate, relative priority); and 5) continuity and communication (inner setting, networks and communications). For most themes, subjects described both positive and negative aspects of chronic disease management, as well as related facilitators and barriers to delivery of chronic disease care for hospitalized patients. Illustrative quotes for each theme are shown in Table 3.
Perceived Impact on Patient Outcomes
Perceived impact on patient outcomes was mixed. Most providers believed the management of chronic diseases could lead to improvement in important patient outcomes, including decreased length of stay (LOS), prevention of hospital complication, and decreased readmissions. Surgical providers focused particularly on the benefits of preventing surgical complications and noted that they were more likely to manage chronic conditions—primarily through use of specialist consultation—when they perceived a benefit to prevention of surgical outcomes or a fear that surgery may worsen a stable chronic condition:
“Most of the surgery I do is pretty stressful on the body and is very likely to induce acute on chronic exacerbations of heart failure. For someone with Class II or higher heart failure, I’m definitely gonna have cardiology on board or at least internal medicine on board right from the beginning.”
Table 3 However, some providers acknowledged that there were potential risks to such management, including “prolonging hospital stays for nonemergent indications” and treatment with therapies that had previously led to an “adverse reaction that wasn’t clearly documented.” Providers were also concerned that treating chronic conditions may take focus away from acute conditions, which could lead to worse patient-centered outcomes. One attending in medicine described it:
“If you do potentially focus on those chronic issues, and there’s already a lot of other stuff going on with the patient, you might not be prioritizing the patient’s active issues appropriately. The patient’s saying, ‘I’m in pain. I’m in pain. I’m in pain,’ and you’re saying, ‘Thank you very much. Look, your heart failure, you didn’t get your beta-blocker.’ There could be a disconnect between patient’s goals, expectations, and your goals and expectations.”
Hospital Structural Characteristics
For many providers, the hospital setting provides a unique opportunity for care of patients with chronic disease. First, a hospitalization is a time for a patient’s management to be reviewed by a new care team. The hospital team reviews the management plan for patients at admission, which is a time to reevaluate whether patients are on evidence-based therapies: “It’s helpful to have a new set of eyes on somebody, like fresh information.” According to providers, this reevaluation can overcome instances of therapeutic inertia by the outpatient physician. Second, the hospital has many resources, including readily available specialist services and diagnostic tests, which can allow a patient-centered approach that coordinates care in 1 place, as a surgery NP described: “I think the advantage for the patient is that they wind up stopping in for 1 thing but we wind up taking care of a few without requiring the need for him or her to go to all these different specialists on the outside. They’re mostly elderly and not able to get around.” Third, the high availability of services and frequent monitoring allows rapid titration of evidence-based medicines, as discussed by a medicine resident: “It’s easier and faster to titrate medication—they’re in a monitored setting; you can ensure compliance.”
Patients may also differ from their usual state while hospitalized, creating both risks and benefits. The hospital setting can provide an opportunity to educate patients on their chronic disease(s) because they are motivated: “They’re in an office visit and their sugars are out of whack or something, they may take it a little bit more seriously if they were just in the hospital even though it was on an unrelated issue. I think it probably just changes their perspective on their disease.” However, in the hospital, patients are in an unusual environment with a restricted diet and forced medication compliance. Furthermore, the acute condition can lead to changes in their chronic disease, as described by 1 medicine attending: “their sugar is high because they’re acutely ill.” Providers expressed concern that changing medications in this setting may lead to adverse events (AEs) when patients return to their usual environment.
Provider Knowledge and Self-Efficacy
Insufficient knowledge of treatments for chronic conditions was cited as a barrier to some providers’ ability to actively manage chronic disease for hospitalized patients. Some providers described management of conditions outside their area as less satisfying than their primary focus. For example, an orthopedic surgeon explained: “…it’s very simple. You see your bone is broken, you fix it, that’s it…it’s intellectually satisfying…managing chronic diseases is less like that.” Reliance on consultants was 1 approach to deal with knowledge gaps in areas outside a provider’s expertise.
For a number of providers, management of stable chronic disease is the responsibility of the outpatient provider. Providers expressed concern that inpatient management was a reach into the domain of the primary care provider (PCP) and might take “away from the primary focus” of the hospitalization. Nonetheless, some providers noted an “ethical responsibility to manage [a] patient correctly,” and some providers believed that engaging in chronic disease management in the hospital would present an opportunity to expand their own expertise.
A few providers were worried about legal risk related to chronic disease management: “we don’t typically deal too much with managing some of these other medical issues for medical and legal reasons.” Providers again suggested that consults can help overcome this concern for risk, as discussed by 1 surgical attending: “We’re all not wanting to be sued, and we want to do the right thing. It costs me nothing to have a cardiologist on board, so like—why not.”
Hospital Priorities
Providers explained that the hospital has strong interests in early discharge and minimizing LOS. These priorities are based on goals of improving patient outcomes, increasing bed availability and hospital volume, and reducing costs. Providers perceive these hospital priorities as potential barriers to chronic disease management, which can increase LOS and costs through additional testing and treatment. As a medicine resident described: “The DBN philosophy, ‘discharge before noon’ philosophy, which is part of the hospital efficiency to get people in and out of the hospital as quickly as [is] safe, or maybe faster. And I think that there’s a culture where you’re encouraged to only focus on the acute issue and tend to defer everything else.”
Continuity and Communication
According to many providers, care continuity between the outpatient setting and the hospital played a major role in management of chronic disease. One barrier to starting a new evidence-based medication was lack of knowledge of patient history. As noted, providers expressed concern that a patient may not be on a given therapy because of an adverse reaction that was not documented in the hospital chart. This is particularly true because, as discussed by a surgery resident, patients with “PCPs outside the system [in which providers] don’t have access to the electronic medical record.” To overcome this barrier, providers attempt to communicate with the outpatient provider to confirm a lack of contraindications to therapies prior to any changes; notably, communication is easier if the inpatient provider has a relationship with the outpatient PCP.
Some providers were more likely to start chronic disease therapies if the patient had no prior outpatient care, because the provider was reassured that there was no rationale for missing therapies. One neurology attending noted that if a patient had newly documented “hypertension even if they were in for something else, I might start them on an antihypertensive, but then arrange for a close follow-up with a new PCP.”
Following hospitalization, providers wanted assurance that any changes to chronic disease management would be followed up by an outpatient physician. Any changes are relayed to the outpatient provider and the “level of communication…with the outpatient provider who’s gonna inherit” these changes can influence how aggressively the inpatient provider manages chronic diseases. Providers may be reluctant to start therapy for patients if they are concerned about outpatient follow up: “they have diabetes and they should really technically be on an ACE [angiotensin converting enzyme]inhibitor and aspirin, but they’re not. I might send them out on the aspirin but I might either start ACE inhibitor and have them follow up with their PCP in 2 weeks if I’m confident that they’ll do it or if I’m really confident that they’ll not follow up, I will help them get the appointment and then the discharge instruction is to the PCP is ‘Please start this patient on ACE inhibitor if they show up.’”
DISCUSSION
Providers frequently perceive benefit to chronic disease management in the hospital, including improvements in clinical outcomes. Notably, providers see opportunities to improve compliance with evidence-based care to overcome potential barriers to managing chronic disease in the outpatient setting, which can be limited by pressure for brief encounters,13 clinical inertia,14 difficulty with close monitoring of patients,15 and care fragmentation.16 Concurrently, inpatient providers are concerned about potential for patient harm related to chronic disease management, primarily related to AEs from medications. Similar to a case study about a patient with outpatient hypotension following aggressive inpatient hypertension management,7 providers fear that changing a patient’s chronic disease management in a hospital setting may cause harm when the patient returns home.
Although some clinicians have argued against aggressive in-hospital chronic disease management because of concerns for risk of AEs,7 our study and others8 have suggested that many clinicians perceive benefit. In some cases, such as smoking cessation counseling for all current smokers and prescribing an angiotensin converting enzyme inhibitor for patients with systolic heart failure, the perceived importance is so great that chronic disease management has been used as a national quality metric for hospitals. While these hospital metrics may be justified for short-term benefits after hospitalization, studies have demonstrated only weak improvement in short-term postdischarge outcomes related to chronic disease management.17 The true benefit is likely from improved processes of care in the short term that lead to long-term improvement in outcomes.4,5,18 Thus, the advantage of starting a patient hospitalized for a stroke on blood pressure medication is the increased likelihood that the patient will continue the medication as an outpatient, which may reduce long-term mortality.
For hospital delivery systems that are concerned with such care process improvement through in-hospital chronic disease management, we identified a number of barriers and facilitators to delivering this care. One significant barrier was poor transitions between the inpatient and the outpatient settings. When a patient transitions into the hospital, providers need to understand prior management choices. Facilitators to help inpatient providers understand prior management included either knowing the outpatient provider, or understanding that there was a lack of regular outpatient care; in both these cases, inpatient providers felt more comfortable managing chronic diseases because they had insight into the outpatient plan, or lack thereof. However, these facilitators may not be practical to incorporate in interventions to improve chronic disease care, which should consider overcoming these communication barriers. Use of shared electronic health records or standardized telephone calls with well-documented care plans obtained through health information exchanges may facilitate an inpatient provider to manage appropriately chronic disease. Similarly, discontinuity between the inpatient provider and the outpatient provider is a barrier that must be overcome to ease concerns that any chronic disease management changes do not result in harm in the postdischarge period. These findings again point to the need for improved documentation and communication between inpatient and outpatient providers. Of course, the transitional care period is one of high risk, and improving communication between providers has been an area of ongoing work.19
Lack of comfort among inpatient providers with managing chronic diseases is another important barrier, which appears to be largely overcome through the use of consultation services. Ready availability of specialists, common in academic medical centers, can facilitate delivery of chronic disease management. Inpatient interventions designed to improve evidence-based care for a chronic disease may benefit from involvement or at least availability of specialists in the effort. Another major barrier relates to hospital priorities, which in our study were closely aligned with external factors such as payment models. As hospitalizations are typically paid based on the discharge diagnosis, hospitals have incentives to discharge quickly and not order extra diagnostic tests. As a result, there are disincentives for chronic disease management that may require additional testing or monitoring in the hospital. Conversely, as hospitals accept postdischarge financial risks through readmission penalties or postdischarge cost savings, hospitals may perceive that long-term benefits of chronic disease management may outweigh short-term costs.
The study findings should be interpreted in the context of its limitations. Findings of our study of providers from a single academic medical center may not be generalizable. Nearly half of our interviews were conducted by telephone, which limits our ability to capture nonverbal cues in communication. Providers may have had social desirability bias towards positive aspects of chronic disease management. We did not have the power to determine differences in response by provider characteristic because this was an exploratory qualitative study. Future studies with representative sampling, a larger sample size, and measures for constructs such as provider self-efficacy are needed to examine differences by specialty, provider type, and experience level.
In conclusion, inpatient providers believe that hospital chronic disease management has the potential to be beneficial for both process of care and clinical outcomes; providers also express concern about potential adverse consequences of managing chronic disease during acute hospitalizations. To maximize both quality of care and patient safety, overcoming communication barriers between inpatient and outpatient providers is needed. Both a supportive hospital environment and availability of specialty support can facilitate in-hospital chronic disease management. Interventions that incorporate these factors may be well-suited to improve chronic disease care and long-term outcomes.
Disclosures
This work was supported by the Agency for Healthcare Research and Quality (AHRQ) grant K08HS23683. The authors report no financial conflicts of interest.
References
1. Friedman B, Jiang HJ, Elixhauser A, Segal A. Hospital inpatient costs for adults with multiple chronic conditions. Med Care Res Rev. 2006;63(3):327-346. PubMed 2. Steiner CA, Friedman B. Hospital utilization, costs, and mortality for adults with multiple chronic conditions, Nationwide Inpatient Sample, 2009. Prev Chronic Dis. 2013;10:E62. PubMed 3. Blecker S, Paul M, Taksler G, Ogedegbe G, Katz S. Heart failure-associated hospitalizations in the United States. J Am Coll Cardiol. 2013;61(12):1259-1267. PubMed 4. Fonarow GC. Role of in-hospital initiation of carvedilol to improve treatment rates and clinical outcomes. Am J Cardiol. 2004;93(9A):77B-81B. PubMed 5. Touze E, Coste J, Voicu M, et al. Importance of in-hospital initiation of therapies and therapeutic inertia in secondary stroke prevention: IMplementation of Prevention After a Cerebrovascular evenT (IMPACT) Study. Stroke. 2008;39(6):1834-1843. PubMed 6. Ovbiagele B, Saver JL, Fredieu A, et al. In-hospital initiation of secondary stroke prevention therapies yields high rates of adherence at follow-up. Stroke. 2004;35(12):2879-2883. PubMed 7. Steinman MA, Auerbach AD. Managing chronic disease in hospitalized patients. JAMA Intern Med. 2013;173(20):1857-1858. PubMed 8. Breu AC, Allen-Dicker J, Mueller S, Palamara K, Hinami K, Herzig SJ. Hospitalist and primary care physician perspectives on medication management of chronic conditions for hospitalized patients. J Hosp Med. 2014;9(5):303-309. PubMed 9. Damschroder LJ, Aron DC, Keith RE, Kirsh SR, Alexander JA, Lowery JC. Fostering implementation of health services research findings into practice: a consolidated framework for advancing implementation science. Implement Sci. 2009;4:50. PubMed 10. Morse JM. The significance of saturation. Qualitative Health Research. 1995;5(2):147-149. 11. Bradley EH, Curry LA, Devers KJ. Qualitative data analysis for health services research: developing taxonomy, themes, and theory. Health Services Research. 2007;42(4):1758-1772. PubMed 12. Riegel B, Dickson VV, Topaz M. Qualitative analysis of naturalistic decision making in adults with chronic heart failure. Nurs Res. 2013;62(2):91-98. PubMed 13. Linzer M, Konrad TR, Douglas J, et al. Managed care, time pressure, and physician job satisfaction: results from the physician worklife study. J Gen Intern Med. 2000;15(7):441-450. PubMed 14. Phillips LS, Branch WT, Cook CB, et al. Clinical inertia. Ann Intern Med. 2001;135(9):825-834. 15. Dev S, Hoffman TK, Kavalieratos D, et al. Barriers to adoption of mineralocorticoid receptor antagonists in patients with heart failure: A mixed-methods study. J Am Heart Assoc. 2016;4(3):e002493. PubMed 16. Stange KC. The problem of fragmentation and the need for integrative solutions. Ann Fam Med. 2009;7(2):100-103. PubMed 17. Fonarow GC, Abraham WT, Albert NM, et al. Association between performance measures and clinical outcomes for patients hospitalized with heart failure. JAMA. 2007;297(1):61-70. PubMed 18. Shah M, Norwood CA, Farias S, Ibrahim S, Chong PH, Fogelfeld L. Diabetes transitional care from inpatient to outpatient setting: pharmacist discharge counseling. J Pharm Pract. 2013;26(2):120-124. PubMed 19. Kripalani S, LeFevre F, Phillips CO, Williams MV, Basaviah P, Baker DW. Deficits in communication and information transfer between hospital-based and primary care physicians: implications for patient safety and continuity of care. JAMA. 2007;297(8):831-841. PubMed
References
1. Friedman B, Jiang HJ, Elixhauser A, Segal A. Hospital inpatient costs for adults with multiple chronic conditions. Med Care Res Rev. 2006;63(3):327-346. PubMed 2. Steiner CA, Friedman B. Hospital utilization, costs, and mortality for adults with multiple chronic conditions, Nationwide Inpatient Sample, 2009. Prev Chronic Dis. 2013;10:E62. PubMed 3. Blecker S, Paul M, Taksler G, Ogedegbe G, Katz S. Heart failure-associated hospitalizations in the United States. J Am Coll Cardiol. 2013;61(12):1259-1267. PubMed 4. Fonarow GC. Role of in-hospital initiation of carvedilol to improve treatment rates and clinical outcomes. Am J Cardiol. 2004;93(9A):77B-81B. PubMed 5. Touze E, Coste J, Voicu M, et al. Importance of in-hospital initiation of therapies and therapeutic inertia in secondary stroke prevention: IMplementation of Prevention After a Cerebrovascular evenT (IMPACT) Study. Stroke. 2008;39(6):1834-1843. PubMed 6. Ovbiagele B, Saver JL, Fredieu A, et al. In-hospital initiation of secondary stroke prevention therapies yields high rates of adherence at follow-up. Stroke. 2004;35(12):2879-2883. PubMed 7. Steinman MA, Auerbach AD. Managing chronic disease in hospitalized patients. JAMA Intern Med. 2013;173(20):1857-1858. PubMed 8. Breu AC, Allen-Dicker J, Mueller S, Palamara K, Hinami K, Herzig SJ. Hospitalist and primary care physician perspectives on medication management of chronic conditions for hospitalized patients. J Hosp Med. 2014;9(5):303-309. PubMed 9. Damschroder LJ, Aron DC, Keith RE, Kirsh SR, Alexander JA, Lowery JC. Fostering implementation of health services research findings into practice: a consolidated framework for advancing implementation science. Implement Sci. 2009;4:50. PubMed 10. Morse JM. The significance of saturation. Qualitative Health Research. 1995;5(2):147-149. 11. Bradley EH, Curry LA, Devers KJ. Qualitative data analysis for health services research: developing taxonomy, themes, and theory. Health Services Research. 2007;42(4):1758-1772. PubMed 12. Riegel B, Dickson VV, Topaz M. Qualitative analysis of naturalistic decision making in adults with chronic heart failure. Nurs Res. 2013;62(2):91-98. PubMed 13. Linzer M, Konrad TR, Douglas J, et al. Managed care, time pressure, and physician job satisfaction: results from the physician worklife study. J Gen Intern Med. 2000;15(7):441-450. PubMed 14. Phillips LS, Branch WT, Cook CB, et al. Clinical inertia. Ann Intern Med. 2001;135(9):825-834. 15. Dev S, Hoffman TK, Kavalieratos D, et al. Barriers to adoption of mineralocorticoid receptor antagonists in patients with heart failure: A mixed-methods study. J Am Heart Assoc. 2016;4(3):e002493. PubMed 16. Stange KC. The problem of fragmentation and the need for integrative solutions. Ann Fam Med. 2009;7(2):100-103. PubMed 17. Fonarow GC, Abraham WT, Albert NM, et al. Association between performance measures and clinical outcomes for patients hospitalized with heart failure. JAMA. 2007;297(1):61-70. PubMed 18. Shah M, Norwood CA, Farias S, Ibrahim S, Chong PH, Fogelfeld L. Diabetes transitional care from inpatient to outpatient setting: pharmacist discharge counseling. J Pharm Pract. 2013;26(2):120-124. PubMed 19. Kripalani S, LeFevre F, Phillips CO, Williams MV, Basaviah P, Baker DW. Deficits in communication and information transfer between hospital-based and primary care physicians: implications for patient safety and continuity of care. JAMA. 2007;297(8):831-841. PubMed
Address for correspondence and reprint requests: Saul Blecker, MD, MHS, New York University School of Medicine, 227 E. 30th St., Room 648, New York, NY 10016; Telephone: 646-501-2513; Fax: 646-501-2706; E-mail: saul.blecker@nyumc.org
Content Gating
Open Access (article Unlocked/Open Access)
Alternative CME
Consolidated Pubs: Do Not Show Source Publication Logo
As the US population ages and becomes increasingly frail, the need for rehabilitation rises. By 2030, an estimated 20% of the population will be 65 years old or older, and almost 10% will be over 75.1 About 20% of hospitalized Medicare patients receive subsequent care in postacute inpatient rehabilitation (PAIR) facilities, accounting for $31 billion in Medicare expenditures in 2014.2 Although the need for rehabilitation will continue to rise, Medicare policy restricts access to it.
Under Medicare policy, PAIR services are covered for certain hospitalized patients but not others. Hospitalized patients are either inpatients, who are billed under Medicare Part A, or outpatients, billed under Part B. When hospital length of stay (LOS) is anticipated to be less than 2 midnights, patients are admitted as outpatients under the term observation status; when longer stays are expected, patients are admitted as inpatients.3 This recently implemented time-based distinction has been criticized as arbitrary, and as potentially shifting many patients from inpatient to outpatient (observation) status.4
The distinction between inpatient and observation status has significant consequences for posthospital care. Medicare Part A covers care in skilled nursing facilities (SNFs) and acute inpatient rehabilitation facilities (IRFs); after hospitalization, inpatients have access to either, without copay. As observation patients are covered under Medicare Part B, they are technically not covered for either service after their hospital stay. IRFs sometimes accept patients from ambulatory and nonacute settings; observation patients may be accepted in rare circumstances, but they pay the Part A deductible ($1288 in 2016) to have the services covered by Medicare. SNF services are never covered for observation patients, and access to this care requires an average out-of-pocket payment of more than $10,503 per beneficiary for a typical SNF stay.5 Given that about 70% of Medicare patients fall below 300% of the federal poverty line,6 the out-of-pocket costs for PAIR services for observation patients can be prohibitive.
Although only 0.75% of community-dwelling Medicare observation patients are discharged to PAIR facilities,7 it is unclear if the need for this care is higher but remains unmet secondary to cost concerns of Medicare beneficiaries. Also unclear is whether observation patients who would benefit from this care but do not receive it end up with poorer health outcomes and therefore use more healthcare services.
The purpose of this study was to estimate the proportion of Medicare observation patients who are admitted from home and receive a recommendation for placement in a PAIR facility, and to determine the ultimate disposition of such patients. We also sought to evaluate the association between recommendation for PAIR placement, LOS, and 30-day hospital revisit rate.
METHODS
The Institutional Review Board of Christiana Care Health System (CCHS) approved this study.
Sample and Design
This was an observational study of community-dwelling Medicare patients admitted under observation status to Delaware’s CCHS, which consists of a 907-bed regional tertiary-care facility in Newark and a 241-bed community hospital in Wilmington. The study period was January 1 to December 31, 2013. We limited our sample to patients treated by hospitalists on hospital wards, as this care constitutes 80% of the care provided to observation patients at CCHS and the majority of care nationally.8 As neither SNF care nor IRF care is covered under Medicare Part B, and both would result in high out-of-pocket costs for Medicare observation patients, we combined them into a single variable, PAIR.
All data were obtained from institutional electronic medical record and administrative data systems. Study inclusion criteria were Medicare as primary insurance, admission to hospital from home, and care received at either CCHS facility. Exclusion criteria were admission from PAIR facility, long-term care facility, assisted-living facility, or inpatient psychiatric facility; death; discharge against medical advice (AMA) or to hospice, non-SNF, or inpatient psychiatric facility; and discovery (during review of case management [CM] notes) of erroneous listing of Medicare as primary insurance, or of inpatient admission (within 30 days before index observation stay) that qualified for PAIR coverage under Medicare Part A.
We reviewed the medical charts of a representative (~30%) sample of the cohort and examined physical therapy (PT) and CM notes to determine the proportions of patients with recommendations for home with no services, home-based PT, possible PAIR, and PAIR. Charts were sorted by medical record number and were reviewed in consecutive order. We coded a patient as having a recommendation for possible PAIR if the PT notes indicated the patient may benefit from PAIR but could have home PT if PAIR placement was not possible. CM notes were also reviewed for evidence of patient or family preference regarding PAIR placement. All questions about PT and CM recommendations were resolved by consensus.
Measures
For the total study sample, we calculated descriptive statistics and frequencies for demographic and administrative variables, including age, sex, race (Caucasian, African American, other), ethnicity (Hispanic/non-Hispanic), ICD-9 (International Classification of Diseases, Ninth Revision) primary diagnosis code, LOS (in hours) for index observation admission, discharge disposition (home with no services, home PT, possible PAIR, PAIR), and 30-day hospital revisit (emergency department, observation, inpatient admission). We used χ2 test, Student t test, and analysis of variance (ANOVA) to test for statistically significant differences in characteristics between the chart review subgroup and the rest of the sample and between the groups with different disposition recommendations from PT notes.
For the chart review subgroup, we used ANOVA to calculate the unadjusted association between PT recommendation and LOS. We then adjusted for potential confounders, using multivariable linear regression with PT recommendation as a predictor and LOS as the outcome, controlling for variables previously associated with increased LOS among observation patients (primary diagnosis category, age, sex).6 We also adjusted for hospitalist group to account for potential variability in care delivery. As LOS was not normally distributed, we calculated the fourth root of LOS, which resulted in a more normal distribution, and used the transformed values in the regression model. We then calculated predicted values from the regression and back-transformed these to obtain adjusted mean values for LOS.
Figure
RESULTS
Of the 1417 unique patients who had Medicare as primary insurance and were admitted under observation status to a hospitalist service during the study period (2013), 94 were excluded (Figure). Of the remaining 1323 patients, the majority were 65 years old or older, female, white, and non-Hispanic. The most common ICD-9 diagnoses were syncope and chest pain. Mean LOS was 46.7 hours (range, 0-519 h). Less than 1% of patients were discharged to PAIR. Almost 25% of patients returned to the hospital, either for an emergency department visit or for observation or inpatient stay, within 30 days (Table).
Table
Of the 419 charts reviewed to determine the proportion of patients evaluated by PT, and their subsequent recommendations, 33 were excluded, leaving 386 (92%) for analysis (Figure). There were no significant demographic differences between the patients in the chart review subgroup and the rest of the patients (Appendix). Of the 386 patients whose charts were analyzed, 181 (46.9%) had a PT evaluation, and 17 (4.4%) received a PAIR recommendation (Figure). Of the 17 patients recommended for PAIR, 12 (70.5%) were 65 years old or older, and 1 was discharged to a PAIR facility. Of the 46 patients recommended for home PT, 29 (63%) were discharged home with no services (Table).
PT-evaluated patients had unadjusted mean LOS of 52.2 hours (discharged home with no services), 64.1 hours (home PT or possible PAIR), and 83.1 hours (PAIR) (P = 0.001). With adjustment made for variables previously associated with increased LOS for observation patients, mean LOS for patients recommended for PAIR remained higher than that for patients in the other 2 categories (Table). Patients recommended for PAIR were more likely to return to hospital within 30 days than patients recommended for home PT or possible PAIR and patients discharged home with no services (Table).
Review of CM notes revealed that, of the 17 patients recommended for PAIR, 7 would have accepted PAIR services had they been covered by Medicare, 4 preferred discharge with home health services, and 6 did not provide clear details of patient or family preference.
DISCUSSION
To our knowledge, this is the first study to use chart review to examine the proportion of observation patients who would benefit from PAIR and the relationships among these patients’ rehabilitation needs, dispositions, and outcomes. We tried to be conservative in our estimates by limiting the study population to patients admitted from home. Nevertheless, the potential need for PAIR significantly outweighed the actual use of PAIR on discharge. The study sample was consistent with nationally representative samples of observation patients in terms of proportion of patients admitted from and discharged to facilities7 and the most common ICD-9 diagnoses.9
Physical Therapy Consultations and Observation
Of the 386 patients whose charts were reviewed and analyzed, 17 (4.4%) were evaluated as medically qualifying for and potentially benefiting from PAIR. Although the rate represents a minority of patients, it is 5- to 6-fold higher than the rate of discharge to PAIR, both in our study population and in previous national samples that used administrative data.7 In some cases, the decision not to discharge the patient to PAIR reflected patient and family preference. However, in other cases, patients clearly could have benefited from PAIR and would have gone had it been covered by Medicare. The gap suggests an unmet need for PAIR among a substantial proportion of Medicare beneficiaries for whom the therapy is recommended and wanted.
Efforts to expand coverage for PAIR have been resisted. According to Medicare regulations, beneficiaries qualify for PAIR coverage if they are hospitalized as inpatients for 3 midnights or longer. Days under observation status do not count toward this requirement, even if this status is changed to inpatient.10 The Medicare Payment Advisory Commission (MedPAC) recommendation that time under observation status count toward the Medicare requirement11 has not been accepted,12 in large part because further expansion of PAIR services likely would be unaffordable to Medicare under its payment structure.13 Given our finding that the need for PAIR likely is much higher than previously anticipated, Medicare policy makers should consider broadening access to PAIR while efforts are made to rein in expenditures through payment reform.
One potential area of cost savings is more judicious use of PT evaluation for observation patients, particularly given our finding that the majority of PT consultations resulted in no further recommendations. Efforts to triage PT consultations for appropriateness have had some success, though the literature is scant.14 To improve value for Medicare, healthcare systems, and patients, researchers should rigorously evaluate approaches that maximize appropriate use of PT services.
Hospital Length of Stay
Our cohort’s mean hospital stay was longer than averages reported elsewhere,9 likely reflecting our selection of Medicare patients rather than a general medicine population.6 However, our cohort’s adjusted mean hospital stay was significantly longer for patients recommended for PAIR than for patients without PT needs. That out-of-pocket costs for observation patients increase dramatically as LOS goes past 48 hours6 could have significant financial implications for Medicare beneficiaries.
Return Visits
Almost 25% of our observation patients returned to hospital within 30 days. There was a significant trend toward increased rehospitalization among patients recommended for PAIR than among patients with no PT needs.
Policies related to PAIR for observation patients are rooted in the concern that expanded access to services will contribute to overuse of services and higher healthcare costs.15 However, patients who could have benefited from PAIR but were not covered also were at risk for increased healthcare use and costs. A recent study found that more than one fourth of observation patients with repeat observation stays accrued excessive financial liability.16 Researchers should determine more precisely how the cost of coverage for PAIR placement on an index observation admission compares with the cost of subsequent healthcare use potentially related to insufficient supportive care at home.
Study Limitations
Our results must be interpreted within the context of study limitations. First is the small sample size, particularly the subset of patients selected for detailed manual chart review. We were limited in our ability to calculate sample size prospectively because we were unaware of prior work that described the association between PT recommendation and outcomes among observation patients. However, post hoc analysis estimated that a sample size of 181 patients would have been needed to determine a statistically significant difference in 30-day hospital revisit between patients recommended for PAIR and patients with no PT needs with 80% power, which we achieved. Although there are significant limitations to post hoc sample size estimation, we consider our work hypothesis-generating and hope it will lead to larger studies.
We could not account for the potential bias of the physical therapists, whose evaluations could have been influenced by knowledge of patients’ observation status. Our findings could have underestimated the proportion of patients who otherwise would have been recommended for PAIR. Alternatively, therapists could have inaccurately assessed and overstated the need for PAIR. Although we could not account for the therapists’ accuracy and biases, their assessments provided crucial information beyond what was previously obtained from administrative data alone.7,9
Hospital revisits were only accounted for within our hospital system—another potential source of underestimated findings. A significant proportion of patients recommended for home PT were discharged without services, which is counterintuitive, as Medicare covers home nursing services for observation patients. This finding most likely reflects administrative error but probably merits further evaluation.
Last, causality cannot be inferred from the results of a retrospective observational study.
CONCLUSION
As our study results suggest, there is an unmet need for PAIR services for Medicare observation patients, and LOS and subsequent use may be increased among patients recommended for PAIR. Our estimates are conservative and may underestimate the true need for services within this population. Our findings bolster MedPAC recommendations to amend the policies for Medicare coverage of PAIR services for observation patients.
Acknowledgment
The authors thank Paul Kolm, PhD, for statistical support.
Disclosures
Dr. Schwartz reports receiving personal fees from the Agency for Health Research and Quality, Bayer, the Blue Cross Blue Shield Association, Pfizer, and Takeda, all outside the submitted work. Dr. Hicks is supported by an Institutional Development Award from the National Institute of General Medical Sciences of the National Institutes of Health (grant U54-GM104941; principal investigator Stuart Binder-Macleod, PT, PhD, FAPTA). The other authors have nothing to report.
1. Ortman JM, Velkoff VA, Hogan H. An Aging Nation: The Older Population in the United States (Current Population Reports, P25-1140). Washington, DC: US Census Bureau; 2014. https://www.census.gov/prod/2014pubs/p25-1140.pdf. Published May 2014. Accessed January 1, 2016. 2. Carter C, Garrett B, Wissoker D. The Need to Reform Medicare’s Payments to Skilled Nursing Facilities Is as Strong as Ever. Washington, DC: Medicare Payment Advisory Commission & Urban Institute; 2015. http://www.urban.org/sites/default/files/publication/39036/2000072-The-Need-to-Reform-Medicare-Payments-to-SNF.pdf. Published January 2015. Accessed January 1, 2016. 3. Cassidy A. The two-midnight rule (Health Policy Brief). HealthAffairs website. http://healthaffairs.org/healthpolicybriefs/brief_pdfs/healthpolicybrief_133.pdf. Published January 22, 2015. Accessed January 1, 2016. 4. Sheehy AM, Caponi B, Gangireddy S, et al. Observation and inpatient status: clinical impact of the 2-midnight rule. J Hosp Med. 2014;9(4):203-209. PubMed 5. Wright S. Memorandum report: hospitals’ use of observation stays and short inpatient stays for Medicare beneficiaries (OEI-02-12-00040). Washington, DC: US Dept of Health and Human Services, Office of Inspector General; 2013. https://oig.hhs.gov/oei/reports/oei-02-12-00040.pdf. Published July 29, 2013. Accessed January 1, 2016. 6. Hockenberry JM, Mutter R, Barrett M, Parlato J, Ross MA. Factors associated with prolonged observation services stays and the impact of long stays on patient cost. Health Serv Res. 2014;49(3):893-909. PubMed 7. Feng Z, Jung HY, Wright B, Mor V. The origin and disposition of Medicare observation stays. Med Care. 2014;52(9):796-800. PubMed 8. Ross MA, Hockenberry JM, Mutter R, Barrett M, Wheatley M, Pitts SR. Protocol-driven emergency department observation units offer savings, shorter stays, and reduced admissions. Health Aff. 2013;32(12):2149-2156. PubMed 9. Sheehy AM, Graf B, Gangireddy S, et al. Hospitalized but not admitted: characteristics of patients with “observation status” at an academic medical center. JAMA Intern Med. 2013;173(21):1991-1998. PubMed 10. Centers for Medicare & Medicaid Services. Medicare & Your Hospital Benefits. https://www.medicare.gov/Pubs/pdf/11408.pdf. CMS Product 11408. Published 2014. Revised March 2016. Accessed February 6, 2017. 11. Medicare Payment Advisory Commission. Hospital short-stay policy issues. In: Report to the Congress: Medicare and the Health Care Delivery System. Washington, DC: Medicare Payment Advisory Commission; 2015:173-204. http://www.medpac.gov/docs/default-source/reports/chapter-7-hospital-short-stay-policy-issues-june-2015-report-.pdf. Published June 2015. Accessed January 1, 2016. 12. Centers for Medicare & Medicaid Services (CMS), HHS. Medicare program: hospital outpatient prospective payment and ambulatory surgical center payment systems and quality reporting programs; short inpatient hospital stays; transition for certain Medicare-dependent, small rural hospitals under the hospital inpatient prospective payment system; provider administrative appeals and judicial review. Final rule with comment period; final rule. Fed Regist. 2015;80(219):70297-70607. PubMed 13. Medicare Payment Advisory Commission. Skilled nursing facility services. In: Report to the Congress: Medicare Payment Policy. Washington, DC: Medicare Payment Advisory Commission; 2015:181-209. http://www.medpac.gov/docs/default-source/reports/chapter-8-skilled-nursing-facility-services-march-2015-report-.pdf. Published March 2015. Accessed January 1, 2016. 14. Hobbs JA, Boysen JF, McGarry KA, Thompson JM, Nordrum JT. Development of a unique triage system for acute care physical therapy and occupational therapy services: an administrative case report. Phys Ther. 2010;90(10):1519-1529. PubMed 15. Centers for Medicare & Medicaid Services (CMS), HHS. Medicare program; hospital inpatient prospective payment systems for acute care hospitals and the long-term care hospital prospective payment system and fiscal year 2014 rates; quality reporting requirements for specific providers; hospital conditions of participation; payment policies related to patient status. Final rules. Fed Regist. 2013;78(160):50495-51040. PubMed 16. Kangovi S, Cafardi SG, Smith RA, Kulkarni R, Grande D. Patient financial responsibility for observation care. J Hosp Med. 2015;10(11):718-723. PubMed
As the US population ages and becomes increasingly frail, the need for rehabilitation rises. By 2030, an estimated 20% of the population will be 65 years old or older, and almost 10% will be over 75.1 About 20% of hospitalized Medicare patients receive subsequent care in postacute inpatient rehabilitation (PAIR) facilities, accounting for $31 billion in Medicare expenditures in 2014.2 Although the need for rehabilitation will continue to rise, Medicare policy restricts access to it.
Under Medicare policy, PAIR services are covered for certain hospitalized patients but not others. Hospitalized patients are either inpatients, who are billed under Medicare Part A, or outpatients, billed under Part B. When hospital length of stay (LOS) is anticipated to be less than 2 midnights, patients are admitted as outpatients under the term observation status; when longer stays are expected, patients are admitted as inpatients.3 This recently implemented time-based distinction has been criticized as arbitrary, and as potentially shifting many patients from inpatient to outpatient (observation) status.4
The distinction between inpatient and observation status has significant consequences for posthospital care. Medicare Part A covers care in skilled nursing facilities (SNFs) and acute inpatient rehabilitation facilities (IRFs); after hospitalization, inpatients have access to either, without copay. As observation patients are covered under Medicare Part B, they are technically not covered for either service after their hospital stay. IRFs sometimes accept patients from ambulatory and nonacute settings; observation patients may be accepted in rare circumstances, but they pay the Part A deductible ($1288 in 2016) to have the services covered by Medicare. SNF services are never covered for observation patients, and access to this care requires an average out-of-pocket payment of more than $10,503 per beneficiary for a typical SNF stay.5 Given that about 70% of Medicare patients fall below 300% of the federal poverty line,6 the out-of-pocket costs for PAIR services for observation patients can be prohibitive.
Although only 0.75% of community-dwelling Medicare observation patients are discharged to PAIR facilities,7 it is unclear if the need for this care is higher but remains unmet secondary to cost concerns of Medicare beneficiaries. Also unclear is whether observation patients who would benefit from this care but do not receive it end up with poorer health outcomes and therefore use more healthcare services.
The purpose of this study was to estimate the proportion of Medicare observation patients who are admitted from home and receive a recommendation for placement in a PAIR facility, and to determine the ultimate disposition of such patients. We also sought to evaluate the association between recommendation for PAIR placement, LOS, and 30-day hospital revisit rate.
METHODS
The Institutional Review Board of Christiana Care Health System (CCHS) approved this study.
Sample and Design
This was an observational study of community-dwelling Medicare patients admitted under observation status to Delaware’s CCHS, which consists of a 907-bed regional tertiary-care facility in Newark and a 241-bed community hospital in Wilmington. The study period was January 1 to December 31, 2013. We limited our sample to patients treated by hospitalists on hospital wards, as this care constitutes 80% of the care provided to observation patients at CCHS and the majority of care nationally.8 As neither SNF care nor IRF care is covered under Medicare Part B, and both would result in high out-of-pocket costs for Medicare observation patients, we combined them into a single variable, PAIR.
All data were obtained from institutional electronic medical record and administrative data systems. Study inclusion criteria were Medicare as primary insurance, admission to hospital from home, and care received at either CCHS facility. Exclusion criteria were admission from PAIR facility, long-term care facility, assisted-living facility, or inpatient psychiatric facility; death; discharge against medical advice (AMA) or to hospice, non-SNF, or inpatient psychiatric facility; and discovery (during review of case management [CM] notes) of erroneous listing of Medicare as primary insurance, or of inpatient admission (within 30 days before index observation stay) that qualified for PAIR coverage under Medicare Part A.
We reviewed the medical charts of a representative (~30%) sample of the cohort and examined physical therapy (PT) and CM notes to determine the proportions of patients with recommendations for home with no services, home-based PT, possible PAIR, and PAIR. Charts were sorted by medical record number and were reviewed in consecutive order. We coded a patient as having a recommendation for possible PAIR if the PT notes indicated the patient may benefit from PAIR but could have home PT if PAIR placement was not possible. CM notes were also reviewed for evidence of patient or family preference regarding PAIR placement. All questions about PT and CM recommendations were resolved by consensus.
Measures
For the total study sample, we calculated descriptive statistics and frequencies for demographic and administrative variables, including age, sex, race (Caucasian, African American, other), ethnicity (Hispanic/non-Hispanic), ICD-9 (International Classification of Diseases, Ninth Revision) primary diagnosis code, LOS (in hours) for index observation admission, discharge disposition (home with no services, home PT, possible PAIR, PAIR), and 30-day hospital revisit (emergency department, observation, inpatient admission). We used χ2 test, Student t test, and analysis of variance (ANOVA) to test for statistically significant differences in characteristics between the chart review subgroup and the rest of the sample and between the groups with different disposition recommendations from PT notes.
For the chart review subgroup, we used ANOVA to calculate the unadjusted association between PT recommendation and LOS. We then adjusted for potential confounders, using multivariable linear regression with PT recommendation as a predictor and LOS as the outcome, controlling for variables previously associated with increased LOS among observation patients (primary diagnosis category, age, sex).6 We also adjusted for hospitalist group to account for potential variability in care delivery. As LOS was not normally distributed, we calculated the fourth root of LOS, which resulted in a more normal distribution, and used the transformed values in the regression model. We then calculated predicted values from the regression and back-transformed these to obtain adjusted mean values for LOS.
Figure
RESULTS
Of the 1417 unique patients who had Medicare as primary insurance and were admitted under observation status to a hospitalist service during the study period (2013), 94 were excluded (Figure). Of the remaining 1323 patients, the majority were 65 years old or older, female, white, and non-Hispanic. The most common ICD-9 diagnoses were syncope and chest pain. Mean LOS was 46.7 hours (range, 0-519 h). Less than 1% of patients were discharged to PAIR. Almost 25% of patients returned to the hospital, either for an emergency department visit or for observation or inpatient stay, within 30 days (Table).
Table
Of the 419 charts reviewed to determine the proportion of patients evaluated by PT, and their subsequent recommendations, 33 were excluded, leaving 386 (92%) for analysis (Figure). There were no significant demographic differences between the patients in the chart review subgroup and the rest of the patients (Appendix). Of the 386 patients whose charts were analyzed, 181 (46.9%) had a PT evaluation, and 17 (4.4%) received a PAIR recommendation (Figure). Of the 17 patients recommended for PAIR, 12 (70.5%) were 65 years old or older, and 1 was discharged to a PAIR facility. Of the 46 patients recommended for home PT, 29 (63%) were discharged home with no services (Table).
PT-evaluated patients had unadjusted mean LOS of 52.2 hours (discharged home with no services), 64.1 hours (home PT or possible PAIR), and 83.1 hours (PAIR) (P = 0.001). With adjustment made for variables previously associated with increased LOS for observation patients, mean LOS for patients recommended for PAIR remained higher than that for patients in the other 2 categories (Table). Patients recommended for PAIR were more likely to return to hospital within 30 days than patients recommended for home PT or possible PAIR and patients discharged home with no services (Table).
Review of CM notes revealed that, of the 17 patients recommended for PAIR, 7 would have accepted PAIR services had they been covered by Medicare, 4 preferred discharge with home health services, and 6 did not provide clear details of patient or family preference.
DISCUSSION
To our knowledge, this is the first study to use chart review to examine the proportion of observation patients who would benefit from PAIR and the relationships among these patients’ rehabilitation needs, dispositions, and outcomes. We tried to be conservative in our estimates by limiting the study population to patients admitted from home. Nevertheless, the potential need for PAIR significantly outweighed the actual use of PAIR on discharge. The study sample was consistent with nationally representative samples of observation patients in terms of proportion of patients admitted from and discharged to facilities7 and the most common ICD-9 diagnoses.9
Physical Therapy Consultations and Observation
Of the 386 patients whose charts were reviewed and analyzed, 17 (4.4%) were evaluated as medically qualifying for and potentially benefiting from PAIR. Although the rate represents a minority of patients, it is 5- to 6-fold higher than the rate of discharge to PAIR, both in our study population and in previous national samples that used administrative data.7 In some cases, the decision not to discharge the patient to PAIR reflected patient and family preference. However, in other cases, patients clearly could have benefited from PAIR and would have gone had it been covered by Medicare. The gap suggests an unmet need for PAIR among a substantial proportion of Medicare beneficiaries for whom the therapy is recommended and wanted.
Efforts to expand coverage for PAIR have been resisted. According to Medicare regulations, beneficiaries qualify for PAIR coverage if they are hospitalized as inpatients for 3 midnights or longer. Days under observation status do not count toward this requirement, even if this status is changed to inpatient.10 The Medicare Payment Advisory Commission (MedPAC) recommendation that time under observation status count toward the Medicare requirement11 has not been accepted,12 in large part because further expansion of PAIR services likely would be unaffordable to Medicare under its payment structure.13 Given our finding that the need for PAIR likely is much higher than previously anticipated, Medicare policy makers should consider broadening access to PAIR while efforts are made to rein in expenditures through payment reform.
One potential area of cost savings is more judicious use of PT evaluation for observation patients, particularly given our finding that the majority of PT consultations resulted in no further recommendations. Efforts to triage PT consultations for appropriateness have had some success, though the literature is scant.14 To improve value for Medicare, healthcare systems, and patients, researchers should rigorously evaluate approaches that maximize appropriate use of PT services.
Hospital Length of Stay
Our cohort’s mean hospital stay was longer than averages reported elsewhere,9 likely reflecting our selection of Medicare patients rather than a general medicine population.6 However, our cohort’s adjusted mean hospital stay was significantly longer for patients recommended for PAIR than for patients without PT needs. That out-of-pocket costs for observation patients increase dramatically as LOS goes past 48 hours6 could have significant financial implications for Medicare beneficiaries.
Return Visits
Almost 25% of our observation patients returned to hospital within 30 days. There was a significant trend toward increased rehospitalization among patients recommended for PAIR than among patients with no PT needs.
Policies related to PAIR for observation patients are rooted in the concern that expanded access to services will contribute to overuse of services and higher healthcare costs.15 However, patients who could have benefited from PAIR but were not covered also were at risk for increased healthcare use and costs. A recent study found that more than one fourth of observation patients with repeat observation stays accrued excessive financial liability.16 Researchers should determine more precisely how the cost of coverage for PAIR placement on an index observation admission compares with the cost of subsequent healthcare use potentially related to insufficient supportive care at home.
Study Limitations
Our results must be interpreted within the context of study limitations. First is the small sample size, particularly the subset of patients selected for detailed manual chart review. We were limited in our ability to calculate sample size prospectively because we were unaware of prior work that described the association between PT recommendation and outcomes among observation patients. However, post hoc analysis estimated that a sample size of 181 patients would have been needed to determine a statistically significant difference in 30-day hospital revisit between patients recommended for PAIR and patients with no PT needs with 80% power, which we achieved. Although there are significant limitations to post hoc sample size estimation, we consider our work hypothesis-generating and hope it will lead to larger studies.
We could not account for the potential bias of the physical therapists, whose evaluations could have been influenced by knowledge of patients’ observation status. Our findings could have underestimated the proportion of patients who otherwise would have been recommended for PAIR. Alternatively, therapists could have inaccurately assessed and overstated the need for PAIR. Although we could not account for the therapists’ accuracy and biases, their assessments provided crucial information beyond what was previously obtained from administrative data alone.7,9
Hospital revisits were only accounted for within our hospital system—another potential source of underestimated findings. A significant proportion of patients recommended for home PT were discharged without services, which is counterintuitive, as Medicare covers home nursing services for observation patients. This finding most likely reflects administrative error but probably merits further evaluation.
Last, causality cannot be inferred from the results of a retrospective observational study.
CONCLUSION
As our study results suggest, there is an unmet need for PAIR services for Medicare observation patients, and LOS and subsequent use may be increased among patients recommended for PAIR. Our estimates are conservative and may underestimate the true need for services within this population. Our findings bolster MedPAC recommendations to amend the policies for Medicare coverage of PAIR services for observation patients.
Acknowledgment
The authors thank Paul Kolm, PhD, for statistical support.
Disclosures
Dr. Schwartz reports receiving personal fees from the Agency for Health Research and Quality, Bayer, the Blue Cross Blue Shield Association, Pfizer, and Takeda, all outside the submitted work. Dr. Hicks is supported by an Institutional Development Award from the National Institute of General Medical Sciences of the National Institutes of Health (grant U54-GM104941; principal investigator Stuart Binder-Macleod, PT, PhD, FAPTA). The other authors have nothing to report.
As the US population ages and becomes increasingly frail, the need for rehabilitation rises. By 2030, an estimated 20% of the population will be 65 years old or older, and almost 10% will be over 75.1 About 20% of hospitalized Medicare patients receive subsequent care in postacute inpatient rehabilitation (PAIR) facilities, accounting for $31 billion in Medicare expenditures in 2014.2 Although the need for rehabilitation will continue to rise, Medicare policy restricts access to it.
Under Medicare policy, PAIR services are covered for certain hospitalized patients but not others. Hospitalized patients are either inpatients, who are billed under Medicare Part A, or outpatients, billed under Part B. When hospital length of stay (LOS) is anticipated to be less than 2 midnights, patients are admitted as outpatients under the term observation status; when longer stays are expected, patients are admitted as inpatients.3 This recently implemented time-based distinction has been criticized as arbitrary, and as potentially shifting many patients from inpatient to outpatient (observation) status.4
The distinction between inpatient and observation status has significant consequences for posthospital care. Medicare Part A covers care in skilled nursing facilities (SNFs) and acute inpatient rehabilitation facilities (IRFs); after hospitalization, inpatients have access to either, without copay. As observation patients are covered under Medicare Part B, they are technically not covered for either service after their hospital stay. IRFs sometimes accept patients from ambulatory and nonacute settings; observation patients may be accepted in rare circumstances, but they pay the Part A deductible ($1288 in 2016) to have the services covered by Medicare. SNF services are never covered for observation patients, and access to this care requires an average out-of-pocket payment of more than $10,503 per beneficiary for a typical SNF stay.5 Given that about 70% of Medicare patients fall below 300% of the federal poverty line,6 the out-of-pocket costs for PAIR services for observation patients can be prohibitive.
Although only 0.75% of community-dwelling Medicare observation patients are discharged to PAIR facilities,7 it is unclear if the need for this care is higher but remains unmet secondary to cost concerns of Medicare beneficiaries. Also unclear is whether observation patients who would benefit from this care but do not receive it end up with poorer health outcomes and therefore use more healthcare services.
The purpose of this study was to estimate the proportion of Medicare observation patients who are admitted from home and receive a recommendation for placement in a PAIR facility, and to determine the ultimate disposition of such patients. We also sought to evaluate the association between recommendation for PAIR placement, LOS, and 30-day hospital revisit rate.
METHODS
The Institutional Review Board of Christiana Care Health System (CCHS) approved this study.
Sample and Design
This was an observational study of community-dwelling Medicare patients admitted under observation status to Delaware’s CCHS, which consists of a 907-bed regional tertiary-care facility in Newark and a 241-bed community hospital in Wilmington. The study period was January 1 to December 31, 2013. We limited our sample to patients treated by hospitalists on hospital wards, as this care constitutes 80% of the care provided to observation patients at CCHS and the majority of care nationally.8 As neither SNF care nor IRF care is covered under Medicare Part B, and both would result in high out-of-pocket costs for Medicare observation patients, we combined them into a single variable, PAIR.
All data were obtained from institutional electronic medical record and administrative data systems. Study inclusion criteria were Medicare as primary insurance, admission to hospital from home, and care received at either CCHS facility. Exclusion criteria were admission from PAIR facility, long-term care facility, assisted-living facility, or inpatient psychiatric facility; death; discharge against medical advice (AMA) or to hospice, non-SNF, or inpatient psychiatric facility; and discovery (during review of case management [CM] notes) of erroneous listing of Medicare as primary insurance, or of inpatient admission (within 30 days before index observation stay) that qualified for PAIR coverage under Medicare Part A.
We reviewed the medical charts of a representative (~30%) sample of the cohort and examined physical therapy (PT) and CM notes to determine the proportions of patients with recommendations for home with no services, home-based PT, possible PAIR, and PAIR. Charts were sorted by medical record number and were reviewed in consecutive order. We coded a patient as having a recommendation for possible PAIR if the PT notes indicated the patient may benefit from PAIR but could have home PT if PAIR placement was not possible. CM notes were also reviewed for evidence of patient or family preference regarding PAIR placement. All questions about PT and CM recommendations were resolved by consensus.
Measures
For the total study sample, we calculated descriptive statistics and frequencies for demographic and administrative variables, including age, sex, race (Caucasian, African American, other), ethnicity (Hispanic/non-Hispanic), ICD-9 (International Classification of Diseases, Ninth Revision) primary diagnosis code, LOS (in hours) for index observation admission, discharge disposition (home with no services, home PT, possible PAIR, PAIR), and 30-day hospital revisit (emergency department, observation, inpatient admission). We used χ2 test, Student t test, and analysis of variance (ANOVA) to test for statistically significant differences in characteristics between the chart review subgroup and the rest of the sample and between the groups with different disposition recommendations from PT notes.
For the chart review subgroup, we used ANOVA to calculate the unadjusted association between PT recommendation and LOS. We then adjusted for potential confounders, using multivariable linear regression with PT recommendation as a predictor and LOS as the outcome, controlling for variables previously associated with increased LOS among observation patients (primary diagnosis category, age, sex).6 We also adjusted for hospitalist group to account for potential variability in care delivery. As LOS was not normally distributed, we calculated the fourth root of LOS, which resulted in a more normal distribution, and used the transformed values in the regression model. We then calculated predicted values from the regression and back-transformed these to obtain adjusted mean values for LOS.
Figure
RESULTS
Of the 1417 unique patients who had Medicare as primary insurance and were admitted under observation status to a hospitalist service during the study period (2013), 94 were excluded (Figure). Of the remaining 1323 patients, the majority were 65 years old or older, female, white, and non-Hispanic. The most common ICD-9 diagnoses were syncope and chest pain. Mean LOS was 46.7 hours (range, 0-519 h). Less than 1% of patients were discharged to PAIR. Almost 25% of patients returned to the hospital, either for an emergency department visit or for observation or inpatient stay, within 30 days (Table).
Table
Of the 419 charts reviewed to determine the proportion of patients evaluated by PT, and their subsequent recommendations, 33 were excluded, leaving 386 (92%) for analysis (Figure). There were no significant demographic differences between the patients in the chart review subgroup and the rest of the patients (Appendix). Of the 386 patients whose charts were analyzed, 181 (46.9%) had a PT evaluation, and 17 (4.4%) received a PAIR recommendation (Figure). Of the 17 patients recommended for PAIR, 12 (70.5%) were 65 years old or older, and 1 was discharged to a PAIR facility. Of the 46 patients recommended for home PT, 29 (63%) were discharged home with no services (Table).
PT-evaluated patients had unadjusted mean LOS of 52.2 hours (discharged home with no services), 64.1 hours (home PT or possible PAIR), and 83.1 hours (PAIR) (P = 0.001). With adjustment made for variables previously associated with increased LOS for observation patients, mean LOS for patients recommended for PAIR remained higher than that for patients in the other 2 categories (Table). Patients recommended for PAIR were more likely to return to hospital within 30 days than patients recommended for home PT or possible PAIR and patients discharged home with no services (Table).
Review of CM notes revealed that, of the 17 patients recommended for PAIR, 7 would have accepted PAIR services had they been covered by Medicare, 4 preferred discharge with home health services, and 6 did not provide clear details of patient or family preference.
DISCUSSION
To our knowledge, this is the first study to use chart review to examine the proportion of observation patients who would benefit from PAIR and the relationships among these patients’ rehabilitation needs, dispositions, and outcomes. We tried to be conservative in our estimates by limiting the study population to patients admitted from home. Nevertheless, the potential need for PAIR significantly outweighed the actual use of PAIR on discharge. The study sample was consistent with nationally representative samples of observation patients in terms of proportion of patients admitted from and discharged to facilities7 and the most common ICD-9 diagnoses.9
Physical Therapy Consultations and Observation
Of the 386 patients whose charts were reviewed and analyzed, 17 (4.4%) were evaluated as medically qualifying for and potentially benefiting from PAIR. Although the rate represents a minority of patients, it is 5- to 6-fold higher than the rate of discharge to PAIR, both in our study population and in previous national samples that used administrative data.7 In some cases, the decision not to discharge the patient to PAIR reflected patient and family preference. However, in other cases, patients clearly could have benefited from PAIR and would have gone had it been covered by Medicare. The gap suggests an unmet need for PAIR among a substantial proportion of Medicare beneficiaries for whom the therapy is recommended and wanted.
Efforts to expand coverage for PAIR have been resisted. According to Medicare regulations, beneficiaries qualify for PAIR coverage if they are hospitalized as inpatients for 3 midnights or longer. Days under observation status do not count toward this requirement, even if this status is changed to inpatient.10 The Medicare Payment Advisory Commission (MedPAC) recommendation that time under observation status count toward the Medicare requirement11 has not been accepted,12 in large part because further expansion of PAIR services likely would be unaffordable to Medicare under its payment structure.13 Given our finding that the need for PAIR likely is much higher than previously anticipated, Medicare policy makers should consider broadening access to PAIR while efforts are made to rein in expenditures through payment reform.
One potential area of cost savings is more judicious use of PT evaluation for observation patients, particularly given our finding that the majority of PT consultations resulted in no further recommendations. Efforts to triage PT consultations for appropriateness have had some success, though the literature is scant.14 To improve value for Medicare, healthcare systems, and patients, researchers should rigorously evaluate approaches that maximize appropriate use of PT services.
Hospital Length of Stay
Our cohort’s mean hospital stay was longer than averages reported elsewhere,9 likely reflecting our selection of Medicare patients rather than a general medicine population.6 However, our cohort’s adjusted mean hospital stay was significantly longer for patients recommended for PAIR than for patients without PT needs. That out-of-pocket costs for observation patients increase dramatically as LOS goes past 48 hours6 could have significant financial implications for Medicare beneficiaries.
Return Visits
Almost 25% of our observation patients returned to hospital within 30 days. There was a significant trend toward increased rehospitalization among patients recommended for PAIR than among patients with no PT needs.
Policies related to PAIR for observation patients are rooted in the concern that expanded access to services will contribute to overuse of services and higher healthcare costs.15 However, patients who could have benefited from PAIR but were not covered also were at risk for increased healthcare use and costs. A recent study found that more than one fourth of observation patients with repeat observation stays accrued excessive financial liability.16 Researchers should determine more precisely how the cost of coverage for PAIR placement on an index observation admission compares with the cost of subsequent healthcare use potentially related to insufficient supportive care at home.
Study Limitations
Our results must be interpreted within the context of study limitations. First is the small sample size, particularly the subset of patients selected for detailed manual chart review. We were limited in our ability to calculate sample size prospectively because we were unaware of prior work that described the association between PT recommendation and outcomes among observation patients. However, post hoc analysis estimated that a sample size of 181 patients would have been needed to determine a statistically significant difference in 30-day hospital revisit between patients recommended for PAIR and patients with no PT needs with 80% power, which we achieved. Although there are significant limitations to post hoc sample size estimation, we consider our work hypothesis-generating and hope it will lead to larger studies.
We could not account for the potential bias of the physical therapists, whose evaluations could have been influenced by knowledge of patients’ observation status. Our findings could have underestimated the proportion of patients who otherwise would have been recommended for PAIR. Alternatively, therapists could have inaccurately assessed and overstated the need for PAIR. Although we could not account for the therapists’ accuracy and biases, their assessments provided crucial information beyond what was previously obtained from administrative data alone.7,9
Hospital revisits were only accounted for within our hospital system—another potential source of underestimated findings. A significant proportion of patients recommended for home PT were discharged without services, which is counterintuitive, as Medicare covers home nursing services for observation patients. This finding most likely reflects administrative error but probably merits further evaluation.
Last, causality cannot be inferred from the results of a retrospective observational study.
CONCLUSION
As our study results suggest, there is an unmet need for PAIR services for Medicare observation patients, and LOS and subsequent use may be increased among patients recommended for PAIR. Our estimates are conservative and may underestimate the true need for services within this population. Our findings bolster MedPAC recommendations to amend the policies for Medicare coverage of PAIR services for observation patients.
Acknowledgment
The authors thank Paul Kolm, PhD, for statistical support.
Disclosures
Dr. Schwartz reports receiving personal fees from the Agency for Health Research and Quality, Bayer, the Blue Cross Blue Shield Association, Pfizer, and Takeda, all outside the submitted work. Dr. Hicks is supported by an Institutional Development Award from the National Institute of General Medical Sciences of the National Institutes of Health (grant U54-GM104941; principal investigator Stuart Binder-Macleod, PT, PhD, FAPTA). The other authors have nothing to report.
References
1. Ortman JM, Velkoff VA, Hogan H. An Aging Nation: The Older Population in the United States (Current Population Reports, P25-1140). Washington, DC: US Census Bureau; 2014. https://www.census.gov/prod/2014pubs/p25-1140.pdf. Published May 2014. Accessed January 1, 2016. 2. Carter C, Garrett B, Wissoker D. The Need to Reform Medicare’s Payments to Skilled Nursing Facilities Is as Strong as Ever. Washington, DC: Medicare Payment Advisory Commission & Urban Institute; 2015. http://www.urban.org/sites/default/files/publication/39036/2000072-The-Need-to-Reform-Medicare-Payments-to-SNF.pdf. Published January 2015. Accessed January 1, 2016. 3. Cassidy A. The two-midnight rule (Health Policy Brief). HealthAffairs website. http://healthaffairs.org/healthpolicybriefs/brief_pdfs/healthpolicybrief_133.pdf. Published January 22, 2015. Accessed January 1, 2016. 4. Sheehy AM, Caponi B, Gangireddy S, et al. Observation and inpatient status: clinical impact of the 2-midnight rule. J Hosp Med. 2014;9(4):203-209. PubMed 5. Wright S. Memorandum report: hospitals’ use of observation stays and short inpatient stays for Medicare beneficiaries (OEI-02-12-00040). Washington, DC: US Dept of Health and Human Services, Office of Inspector General; 2013. https://oig.hhs.gov/oei/reports/oei-02-12-00040.pdf. Published July 29, 2013. Accessed January 1, 2016. 6. Hockenberry JM, Mutter R, Barrett M, Parlato J, Ross MA. Factors associated with prolonged observation services stays and the impact of long stays on patient cost. Health Serv Res. 2014;49(3):893-909. PubMed 7. Feng Z, Jung HY, Wright B, Mor V. The origin and disposition of Medicare observation stays. Med Care. 2014;52(9):796-800. PubMed 8. Ross MA, Hockenberry JM, Mutter R, Barrett M, Wheatley M, Pitts SR. Protocol-driven emergency department observation units offer savings, shorter stays, and reduced admissions. Health Aff. 2013;32(12):2149-2156. PubMed 9. Sheehy AM, Graf B, Gangireddy S, et al. Hospitalized but not admitted: characteristics of patients with “observation status” at an academic medical center. JAMA Intern Med. 2013;173(21):1991-1998. PubMed 10. Centers for Medicare & Medicaid Services. Medicare & Your Hospital Benefits. https://www.medicare.gov/Pubs/pdf/11408.pdf. CMS Product 11408. Published 2014. Revised March 2016. Accessed February 6, 2017. 11. Medicare Payment Advisory Commission. Hospital short-stay policy issues. In: Report to the Congress: Medicare and the Health Care Delivery System. Washington, DC: Medicare Payment Advisory Commission; 2015:173-204. http://www.medpac.gov/docs/default-source/reports/chapter-7-hospital-short-stay-policy-issues-june-2015-report-.pdf. Published June 2015. Accessed January 1, 2016. 12. Centers for Medicare & Medicaid Services (CMS), HHS. Medicare program: hospital outpatient prospective payment and ambulatory surgical center payment systems and quality reporting programs; short inpatient hospital stays; transition for certain Medicare-dependent, small rural hospitals under the hospital inpatient prospective payment system; provider administrative appeals and judicial review. Final rule with comment period; final rule. Fed Regist. 2015;80(219):70297-70607. PubMed 13. Medicare Payment Advisory Commission. Skilled nursing facility services. In: Report to the Congress: Medicare Payment Policy. Washington, DC: Medicare Payment Advisory Commission; 2015:181-209. http://www.medpac.gov/docs/default-source/reports/chapter-8-skilled-nursing-facility-services-march-2015-report-.pdf. Published March 2015. Accessed January 1, 2016. 14. Hobbs JA, Boysen JF, McGarry KA, Thompson JM, Nordrum JT. Development of a unique triage system for acute care physical therapy and occupational therapy services: an administrative case report. Phys Ther. 2010;90(10):1519-1529. PubMed 15. Centers for Medicare & Medicaid Services (CMS), HHS. Medicare program; hospital inpatient prospective payment systems for acute care hospitals and the long-term care hospital prospective payment system and fiscal year 2014 rates; quality reporting requirements for specific providers; hospital conditions of participation; payment policies related to patient status. Final rules. Fed Regist. 2013;78(160):50495-51040. PubMed 16. Kangovi S, Cafardi SG, Smith RA, Kulkarni R, Grande D. Patient financial responsibility for observation care. J Hosp Med. 2015;10(11):718-723. PubMed
References
1. Ortman JM, Velkoff VA, Hogan H. An Aging Nation: The Older Population in the United States (Current Population Reports, P25-1140). Washington, DC: US Census Bureau; 2014. https://www.census.gov/prod/2014pubs/p25-1140.pdf. Published May 2014. Accessed January 1, 2016. 2. Carter C, Garrett B, Wissoker D. The Need to Reform Medicare’s Payments to Skilled Nursing Facilities Is as Strong as Ever. Washington, DC: Medicare Payment Advisory Commission & Urban Institute; 2015. http://www.urban.org/sites/default/files/publication/39036/2000072-The-Need-to-Reform-Medicare-Payments-to-SNF.pdf. Published January 2015. Accessed January 1, 2016. 3. Cassidy A. The two-midnight rule (Health Policy Brief). HealthAffairs website. http://healthaffairs.org/healthpolicybriefs/brief_pdfs/healthpolicybrief_133.pdf. Published January 22, 2015. Accessed January 1, 2016. 4. Sheehy AM, Caponi B, Gangireddy S, et al. Observation and inpatient status: clinical impact of the 2-midnight rule. J Hosp Med. 2014;9(4):203-209. PubMed 5. Wright S. Memorandum report: hospitals’ use of observation stays and short inpatient stays for Medicare beneficiaries (OEI-02-12-00040). Washington, DC: US Dept of Health and Human Services, Office of Inspector General; 2013. https://oig.hhs.gov/oei/reports/oei-02-12-00040.pdf. Published July 29, 2013. Accessed January 1, 2016. 6. Hockenberry JM, Mutter R, Barrett M, Parlato J, Ross MA. Factors associated with prolonged observation services stays and the impact of long stays on patient cost. Health Serv Res. 2014;49(3):893-909. PubMed 7. Feng Z, Jung HY, Wright B, Mor V. The origin and disposition of Medicare observation stays. Med Care. 2014;52(9):796-800. PubMed 8. Ross MA, Hockenberry JM, Mutter R, Barrett M, Wheatley M, Pitts SR. Protocol-driven emergency department observation units offer savings, shorter stays, and reduced admissions. Health Aff. 2013;32(12):2149-2156. PubMed 9. Sheehy AM, Graf B, Gangireddy S, et al. Hospitalized but not admitted: characteristics of patients with “observation status” at an academic medical center. JAMA Intern Med. 2013;173(21):1991-1998. PubMed 10. Centers for Medicare & Medicaid Services. Medicare & Your Hospital Benefits. https://www.medicare.gov/Pubs/pdf/11408.pdf. CMS Product 11408. Published 2014. Revised March 2016. Accessed February 6, 2017. 11. Medicare Payment Advisory Commission. Hospital short-stay policy issues. In: Report to the Congress: Medicare and the Health Care Delivery System. Washington, DC: Medicare Payment Advisory Commission; 2015:173-204. http://www.medpac.gov/docs/default-source/reports/chapter-7-hospital-short-stay-policy-issues-june-2015-report-.pdf. Published June 2015. Accessed January 1, 2016. 12. Centers for Medicare & Medicaid Services (CMS), HHS. Medicare program: hospital outpatient prospective payment and ambulatory surgical center payment systems and quality reporting programs; short inpatient hospital stays; transition for certain Medicare-dependent, small rural hospitals under the hospital inpatient prospective payment system; provider administrative appeals and judicial review. Final rule with comment period; final rule. Fed Regist. 2015;80(219):70297-70607. PubMed 13. Medicare Payment Advisory Commission. Skilled nursing facility services. In: Report to the Congress: Medicare Payment Policy. Washington, DC: Medicare Payment Advisory Commission; 2015:181-209. http://www.medpac.gov/docs/default-source/reports/chapter-8-skilled-nursing-facility-services-march-2015-report-.pdf. Published March 2015. Accessed January 1, 2016. 14. Hobbs JA, Boysen JF, McGarry KA, Thompson JM, Nordrum JT. Development of a unique triage system for acute care physical therapy and occupational therapy services: an administrative case report. Phys Ther. 2010;90(10):1519-1529. PubMed 15. Centers for Medicare & Medicaid Services (CMS), HHS. Medicare program; hospital inpatient prospective payment systems for acute care hospitals and the long-term care hospital prospective payment system and fiscal year 2014 rates; quality reporting requirements for specific providers; hospital conditions of participation; payment policies related to patient status. Final rules. Fed Regist. 2013;78(160):50495-51040. PubMed 16. Kangovi S, Cafardi SG, Smith RA, Kulkarni R, Grande D. Patient financial responsibility for observation care. J Hosp Med. 2015;10(11):718-723. PubMed
Address for correspondence and reprint requests: Jennifer N. Goldstein, MD, MSc, Department of Medicine, Christiana Care Health System, Suite 2E70, Ammon Education Building, 4755 Ogletown-Stanton Rd, Newark, DE 19713; Telephone: 302-733-6591; Fax: 302-733-6082; E-mail: jgoldstein@christianacare.org
The field of hospital medicine (HM) is rapidly expanding in the areas of clinical medicine, administration, and quality improvement (QI).1 Emerging with this growth is a gap in the traditional internal medicine (IM) training and skills needed to be effective in HM.1,2 These skills include clinical and nonclinical aptitudes, such as process improvement, health care economics, and leadership.1-3 However, resident education on these topics must compete with other required curricular content in IM residency training.2,4 Few IM residencies offer focused HM training that emphasizes key components of successful HM careers.3,5
Within the past decade, designated HM tracks within IM residency programs have been proposed as a potential solution. Initially, calls for such tracks focused on gaps in the clinical competencies required of hospitalists.1 Tracks have since evolved to also include skills required to drive high-value care, process improvement, and scholarship. Designated HM tracks address these areas through greater breadth of curricula, additional time for reflection, participation in group projects, and active application to clinical care.4 We conducted a study to identify themes that could inform the ongoing evolution of dedicated HM tracks.
METHODS
Programs were initially identified through communication among professional networks. The phrases hospital medicine residency track and internal medicine residency hospitalist track were used in broader Google searches, as there is no database of such tracks. Searches were performed quarterly during the 2015–2016 academic year. The top 20 hits were manually filtered to identify tracks affiliated with major academic centers. IM residency program websites provided basic information for programs with tracks. We excluded tracks focused entirely on QI6 because, though a crucial part of HM, QI training alone is probably insufficient for preparing residents for success as hospitalists on residency completion. Similarly, IM residencies with stand-alone HM clinical rotations without longitudinal HM curricula were excluded.
Semistructured interviews with track directors were conducted by e-mail or telephone for all tracks except one, the details of which are published.7 We tabulated data and reviewed qualitative information to identify themes among the different tracks. As this study did not involve human participants, Institutional Review Board approval was not needed.
RESULTS
We identified 11 HM residency training programs at major academic centers across the United States: Cleveland Clinic, Stanford University, Tulane University, University of California Davis, University of California Irvine, University of Colorado, University of Kentucky, University of Minnesota, University of New Mexico, Virginia Commonwealth University, and Wake Forest University (Table 1). We reviewed the websites of about 10 other programs, but none suggested existence of a track. Additional programs contacted reported no current track.
Table 1
Track Participants and Structure
HM tracks mainly target third-year residents (Table 1). Some extend into the second year of residency, and 4 have opportunities for intern involvement, including a separate match number at Colorado. Tracks accept up to 12 residents per class. Two programs, at Colorado and Virginia, are part of IM programs in which all residents belong to a track (eg, HM, primary care, research).
HM track structures vary widely and are heavily influenced by the content delivery platforms of their IM residency programs. Several HM track directors emphasized the importance of fitting into existing educational frameworks to ensure access to residents and to minimize the burden of participation. Four programs deliver the bulk of their nonclinical content in dedicated blocks; 6 others use brief recurring sessions to deliver smaller aliquots longitudinally (Table 1). The number of protected hours for content delivery ranges from 10 to more than 40 annually. All tracks use multiple content delivery modes, including didactic sessions and journal clubs. Four tracks employ panel discussions to explore career options within HM. Several also use online platforms, including discussions, readings, and modules.
Quality Improvement
The vast majority of curricula prominently feature experiential QI project involvement (Table 2). These mentored longitudinal projects allow applied delivery of content, such as QI methods and management skills. Four tracks use material from the Institute for Healthcare Improvement.8 Several also offer dedicated QI rotations that immerse residents in ongoing QI efforts.
Table 2
Institutional partnerships support these initiatives at several sites. The Minnesota track is a joint venture of the university and Regions Hospital, a nonprofit community hospital. The Virginia track positions HM residents to lead university-wide interdisciplinary QI teams. For project support, the Colorado and Kentucky tracks partner with local QI resources—the Institute for Healthcare Quality, Safety, and Efficiency at Colorado and the Office of Value and Innovation in Healthcare Delivery at Kentucky.
Health Care Economics and Value
Many programs leverage the rapidly growing emphasis on health care “value” as an opportunity for synergy between IM programs and HM tracks. Examples include involving residents in efforts to improve documentation or didactic instruction on topics such as health care finance. The New Mexico and Wake Forest tracks offer elective rotations on health care economics. Several track directors mentioned successfully expanding curricula on health care value from the HM track into IM residency programs at large, providing a measurable service to the residency programs while ensuring content delivery and freeing up additional time for track activities.
Scholarship and Career Development
Most programs provide targeted career development for residents. Six tracks provide sessions on job procurement skills, such as curriculum vitae preparation and interviewing (Table 2). Many also provide content on venues for disseminating scholarly activity. The Colorado, Kentucky, New Mexico, and Tulane programs feature content on abstract and poster creation. Leadership development is addressed in several tracks through dedicated track activities or participation in discrete, outside-track events. Specifically, Colorado offers a leadership track for residents interested in hospital administration, Cleveland has a leadership journal club, Wake Forest enrolls HM residents in leadership training available through the university, and Minnesota sends residents to the Society of Hospital Medicine’s Leadership Academy (Table 2).
Clinical Rotations
Almost all tracks include a clinical rotation, typically pairing residents directly with hospitalist attendings to encourage autonomy and mentorship. Several also offer elective rotations in various disciplines within HM (Table 2). The Kentucky and Virginia tracks incorporate working with advanced practice providers into their practicums. The Cleveland, Minnesota, Tulane, and Virginia tracks offer HM rotations in community hospitals or postacute settings.
HM rotations also pair clinical experiences with didactic education on relevant topics (eg, billing and coding). The Cleveland, Minnesota, and Virginia tracks developed clinical rotations reflecting the common 7-on and 7-off schedule with nonclinical obligations, such as seminars linking specific content to clinical experiences, during nonclinical time.
DISCUSSION
Our investigation into the current state of HM training found that HM track curricula focus largely on QI, health care economics, and professional development. This focus likely developed in response to hospitalists’ increasing engagement in related endeavors. HM tracks have dynamic and variable structures, reflecting an evolving field and the need to fit into existing IM residency program structures. Similarly, the content covered in HM tracks is tightly linked to perceived opportunities within IM residency curricula. The heterogeneity of content suggests the breadth and ambiguity of necessary competencies for aspiring hospitalists. One of the 11 tracks has not had any residents enroll within the past few years—a testament to the continued effort necessary to sustain such tracks, including curricular updates and recruiting. Conversely, many programs now share track content with the larger IM residency program, suggesting HM tracks may be near the forefront of medical education in some areas.
Our study had several limitations. As we are unaware of any databases of HM tracks, we discussed tracks with professional contacts, performed Internet searches, and reviewed IM residency program websites. Our search, however, was not exhaustive; despite our best efforts, we may have missed or mischaracterized some track offerings. Nevertheless, we think that our analysis represents the first thorough compilation of HM tracks and that it will be useful to institutions seeking to create or enhance HM-specific training.
As the field continues to evolve, we are optimistic about the future of HM training. We suspect that HM residency training tracks will continue to expand. More work is needed so these tracks can adjust to the changing HM and IM residency program landscapes and supply well-trained physicians for the HM workforce.
Acknowledgment
The authors thank track directors Alpesh Amin, David Gugliotti, Rick Hilger, Karnjit Johl, Nasir Majeed, Georgia McIntosh, Charles Pizanis, and Jeff Wiese for making this study possible.
Disclosure
Nothing to report.
References
1. Glasheen JJ, Siegal EM, Epstein K, Kutner J, Prochazka AV. Fulfilling the promise of hospital medicine: tailoring internal medicine training to address hospitalists’ needs [published correction appears in J Gen Intern Med. 2008;23(11):1931]. J Gen Intern Med. 2008;23(7):1110-1115. PubMed 2. Arora V, Guardiano S, Donaldson D, Storch I, Hemstreet P. Closing the gap between internal medicine training and practice: recommendations from recent graduates. Am J Med. 2005;118(6):680-685. PubMed 3. Glasheen JJ, Goldenberg J, Nelson JR. Achieving hospital medicine’s promise through internal medicine residency redesign. Mt Sinai J Med. 2008;75(5):436-441. PubMed 4. Wiese J. Residency training: beginning with the end in mind. J Gen Intern Med. 2008;23(7):1122-1123. PubMed 5. Glasheen JJ, Epstein KR, Siegal E, Kutner JS, Prochazka AV. The spectrum of community-based hospitalist practice: a call to tailor internal medicine residency training. Arch Intern Med. 2007;167(7):727-728. PubMed 6. Patel N, Brennan PJ, Metlay J, Bellini L, Shannon RP, Myers JS. Building the pipeline: the creation of a residency training pathway for future physician leaders in health care quality. Acad Med. 2015;90(2):185-190. PubMed 7. Kumar A, Smeraglio A, Witteles R, et al. A resident-created hospitalist curriculum for internal medicine housestaff. J Hosp Med. 2016;11(9):646-649. PubMed 8. Institute for Healthcare Improvement website. http://www.ihi.org. Accessed December 15, 2015.
The field of hospital medicine (HM) is rapidly expanding in the areas of clinical medicine, administration, and quality improvement (QI).1 Emerging with this growth is a gap in the traditional internal medicine (IM) training and skills needed to be effective in HM.1,2 These skills include clinical and nonclinical aptitudes, such as process improvement, health care economics, and leadership.1-3 However, resident education on these topics must compete with other required curricular content in IM residency training.2,4 Few IM residencies offer focused HM training that emphasizes key components of successful HM careers.3,5
Within the past decade, designated HM tracks within IM residency programs have been proposed as a potential solution. Initially, calls for such tracks focused on gaps in the clinical competencies required of hospitalists.1 Tracks have since evolved to also include skills required to drive high-value care, process improvement, and scholarship. Designated HM tracks address these areas through greater breadth of curricula, additional time for reflection, participation in group projects, and active application to clinical care.4 We conducted a study to identify themes that could inform the ongoing evolution of dedicated HM tracks.
METHODS
Programs were initially identified through communication among professional networks. The phrases hospital medicine residency track and internal medicine residency hospitalist track were used in broader Google searches, as there is no database of such tracks. Searches were performed quarterly during the 2015–2016 academic year. The top 20 hits were manually filtered to identify tracks affiliated with major academic centers. IM residency program websites provided basic information for programs with tracks. We excluded tracks focused entirely on QI6 because, though a crucial part of HM, QI training alone is probably insufficient for preparing residents for success as hospitalists on residency completion. Similarly, IM residencies with stand-alone HM clinical rotations without longitudinal HM curricula were excluded.
Semistructured interviews with track directors were conducted by e-mail or telephone for all tracks except one, the details of which are published.7 We tabulated data and reviewed qualitative information to identify themes among the different tracks. As this study did not involve human participants, Institutional Review Board approval was not needed.
RESULTS
We identified 11 HM residency training programs at major academic centers across the United States: Cleveland Clinic, Stanford University, Tulane University, University of California Davis, University of California Irvine, University of Colorado, University of Kentucky, University of Minnesota, University of New Mexico, Virginia Commonwealth University, and Wake Forest University (Table 1). We reviewed the websites of about 10 other programs, but none suggested existence of a track. Additional programs contacted reported no current track.
Table 1
Track Participants and Structure
HM tracks mainly target third-year residents (Table 1). Some extend into the second year of residency, and 4 have opportunities for intern involvement, including a separate match number at Colorado. Tracks accept up to 12 residents per class. Two programs, at Colorado and Virginia, are part of IM programs in which all residents belong to a track (eg, HM, primary care, research).
HM track structures vary widely and are heavily influenced by the content delivery platforms of their IM residency programs. Several HM track directors emphasized the importance of fitting into existing educational frameworks to ensure access to residents and to minimize the burden of participation. Four programs deliver the bulk of their nonclinical content in dedicated blocks; 6 others use brief recurring sessions to deliver smaller aliquots longitudinally (Table 1). The number of protected hours for content delivery ranges from 10 to more than 40 annually. All tracks use multiple content delivery modes, including didactic sessions and journal clubs. Four tracks employ panel discussions to explore career options within HM. Several also use online platforms, including discussions, readings, and modules.
Quality Improvement
The vast majority of curricula prominently feature experiential QI project involvement (Table 2). These mentored longitudinal projects allow applied delivery of content, such as QI methods and management skills. Four tracks use material from the Institute for Healthcare Improvement.8 Several also offer dedicated QI rotations that immerse residents in ongoing QI efforts.
Table 2
Institutional partnerships support these initiatives at several sites. The Minnesota track is a joint venture of the university and Regions Hospital, a nonprofit community hospital. The Virginia track positions HM residents to lead university-wide interdisciplinary QI teams. For project support, the Colorado and Kentucky tracks partner with local QI resources—the Institute for Healthcare Quality, Safety, and Efficiency at Colorado and the Office of Value and Innovation in Healthcare Delivery at Kentucky.
Health Care Economics and Value
Many programs leverage the rapidly growing emphasis on health care “value” as an opportunity for synergy between IM programs and HM tracks. Examples include involving residents in efforts to improve documentation or didactic instruction on topics such as health care finance. The New Mexico and Wake Forest tracks offer elective rotations on health care economics. Several track directors mentioned successfully expanding curricula on health care value from the HM track into IM residency programs at large, providing a measurable service to the residency programs while ensuring content delivery and freeing up additional time for track activities.
Scholarship and Career Development
Most programs provide targeted career development for residents. Six tracks provide sessions on job procurement skills, such as curriculum vitae preparation and interviewing (Table 2). Many also provide content on venues for disseminating scholarly activity. The Colorado, Kentucky, New Mexico, and Tulane programs feature content on abstract and poster creation. Leadership development is addressed in several tracks through dedicated track activities or participation in discrete, outside-track events. Specifically, Colorado offers a leadership track for residents interested in hospital administration, Cleveland has a leadership journal club, Wake Forest enrolls HM residents in leadership training available through the university, and Minnesota sends residents to the Society of Hospital Medicine’s Leadership Academy (Table 2).
Clinical Rotations
Almost all tracks include a clinical rotation, typically pairing residents directly with hospitalist attendings to encourage autonomy and mentorship. Several also offer elective rotations in various disciplines within HM (Table 2). The Kentucky and Virginia tracks incorporate working with advanced practice providers into their practicums. The Cleveland, Minnesota, Tulane, and Virginia tracks offer HM rotations in community hospitals or postacute settings.
HM rotations also pair clinical experiences with didactic education on relevant topics (eg, billing and coding). The Cleveland, Minnesota, and Virginia tracks developed clinical rotations reflecting the common 7-on and 7-off schedule with nonclinical obligations, such as seminars linking specific content to clinical experiences, during nonclinical time.
DISCUSSION
Our investigation into the current state of HM training found that HM track curricula focus largely on QI, health care economics, and professional development. This focus likely developed in response to hospitalists’ increasing engagement in related endeavors. HM tracks have dynamic and variable structures, reflecting an evolving field and the need to fit into existing IM residency program structures. Similarly, the content covered in HM tracks is tightly linked to perceived opportunities within IM residency curricula. The heterogeneity of content suggests the breadth and ambiguity of necessary competencies for aspiring hospitalists. One of the 11 tracks has not had any residents enroll within the past few years—a testament to the continued effort necessary to sustain such tracks, including curricular updates and recruiting. Conversely, many programs now share track content with the larger IM residency program, suggesting HM tracks may be near the forefront of medical education in some areas.
Our study had several limitations. As we are unaware of any databases of HM tracks, we discussed tracks with professional contacts, performed Internet searches, and reviewed IM residency program websites. Our search, however, was not exhaustive; despite our best efforts, we may have missed or mischaracterized some track offerings. Nevertheless, we think that our analysis represents the first thorough compilation of HM tracks and that it will be useful to institutions seeking to create or enhance HM-specific training.
As the field continues to evolve, we are optimistic about the future of HM training. We suspect that HM residency training tracks will continue to expand. More work is needed so these tracks can adjust to the changing HM and IM residency program landscapes and supply well-trained physicians for the HM workforce.
Acknowledgment
The authors thank track directors Alpesh Amin, David Gugliotti, Rick Hilger, Karnjit Johl, Nasir Majeed, Georgia McIntosh, Charles Pizanis, and Jeff Wiese for making this study possible.
Disclosure
Nothing to report.
The field of hospital medicine (HM) is rapidly expanding in the areas of clinical medicine, administration, and quality improvement (QI).1 Emerging with this growth is a gap in the traditional internal medicine (IM) training and skills needed to be effective in HM.1,2 These skills include clinical and nonclinical aptitudes, such as process improvement, health care economics, and leadership.1-3 However, resident education on these topics must compete with other required curricular content in IM residency training.2,4 Few IM residencies offer focused HM training that emphasizes key components of successful HM careers.3,5
Within the past decade, designated HM tracks within IM residency programs have been proposed as a potential solution. Initially, calls for such tracks focused on gaps in the clinical competencies required of hospitalists.1 Tracks have since evolved to also include skills required to drive high-value care, process improvement, and scholarship. Designated HM tracks address these areas through greater breadth of curricula, additional time for reflection, participation in group projects, and active application to clinical care.4 We conducted a study to identify themes that could inform the ongoing evolution of dedicated HM tracks.
METHODS
Programs were initially identified through communication among professional networks. The phrases hospital medicine residency track and internal medicine residency hospitalist track were used in broader Google searches, as there is no database of such tracks. Searches were performed quarterly during the 2015–2016 academic year. The top 20 hits were manually filtered to identify tracks affiliated with major academic centers. IM residency program websites provided basic information for programs with tracks. We excluded tracks focused entirely on QI6 because, though a crucial part of HM, QI training alone is probably insufficient for preparing residents for success as hospitalists on residency completion. Similarly, IM residencies with stand-alone HM clinical rotations without longitudinal HM curricula were excluded.
Semistructured interviews with track directors were conducted by e-mail or telephone for all tracks except one, the details of which are published.7 We tabulated data and reviewed qualitative information to identify themes among the different tracks. As this study did not involve human participants, Institutional Review Board approval was not needed.
RESULTS
We identified 11 HM residency training programs at major academic centers across the United States: Cleveland Clinic, Stanford University, Tulane University, University of California Davis, University of California Irvine, University of Colorado, University of Kentucky, University of Minnesota, University of New Mexico, Virginia Commonwealth University, and Wake Forest University (Table 1). We reviewed the websites of about 10 other programs, but none suggested existence of a track. Additional programs contacted reported no current track.
Table 1
Track Participants and Structure
HM tracks mainly target third-year residents (Table 1). Some extend into the second year of residency, and 4 have opportunities for intern involvement, including a separate match number at Colorado. Tracks accept up to 12 residents per class. Two programs, at Colorado and Virginia, are part of IM programs in which all residents belong to a track (eg, HM, primary care, research).
HM track structures vary widely and are heavily influenced by the content delivery platforms of their IM residency programs. Several HM track directors emphasized the importance of fitting into existing educational frameworks to ensure access to residents and to minimize the burden of participation. Four programs deliver the bulk of their nonclinical content in dedicated blocks; 6 others use brief recurring sessions to deliver smaller aliquots longitudinally (Table 1). The number of protected hours for content delivery ranges from 10 to more than 40 annually. All tracks use multiple content delivery modes, including didactic sessions and journal clubs. Four tracks employ panel discussions to explore career options within HM. Several also use online platforms, including discussions, readings, and modules.
Quality Improvement
The vast majority of curricula prominently feature experiential QI project involvement (Table 2). These mentored longitudinal projects allow applied delivery of content, such as QI methods and management skills. Four tracks use material from the Institute for Healthcare Improvement.8 Several also offer dedicated QI rotations that immerse residents in ongoing QI efforts.
Table 2
Institutional partnerships support these initiatives at several sites. The Minnesota track is a joint venture of the university and Regions Hospital, a nonprofit community hospital. The Virginia track positions HM residents to lead university-wide interdisciplinary QI teams. For project support, the Colorado and Kentucky tracks partner with local QI resources—the Institute for Healthcare Quality, Safety, and Efficiency at Colorado and the Office of Value and Innovation in Healthcare Delivery at Kentucky.
Health Care Economics and Value
Many programs leverage the rapidly growing emphasis on health care “value” as an opportunity for synergy between IM programs and HM tracks. Examples include involving residents in efforts to improve documentation or didactic instruction on topics such as health care finance. The New Mexico and Wake Forest tracks offer elective rotations on health care economics. Several track directors mentioned successfully expanding curricula on health care value from the HM track into IM residency programs at large, providing a measurable service to the residency programs while ensuring content delivery and freeing up additional time for track activities.
Scholarship and Career Development
Most programs provide targeted career development for residents. Six tracks provide sessions on job procurement skills, such as curriculum vitae preparation and interviewing (Table 2). Many also provide content on venues for disseminating scholarly activity. The Colorado, Kentucky, New Mexico, and Tulane programs feature content on abstract and poster creation. Leadership development is addressed in several tracks through dedicated track activities or participation in discrete, outside-track events. Specifically, Colorado offers a leadership track for residents interested in hospital administration, Cleveland has a leadership journal club, Wake Forest enrolls HM residents in leadership training available through the university, and Minnesota sends residents to the Society of Hospital Medicine’s Leadership Academy (Table 2).
Clinical Rotations
Almost all tracks include a clinical rotation, typically pairing residents directly with hospitalist attendings to encourage autonomy and mentorship. Several also offer elective rotations in various disciplines within HM (Table 2). The Kentucky and Virginia tracks incorporate working with advanced practice providers into their practicums. The Cleveland, Minnesota, Tulane, and Virginia tracks offer HM rotations in community hospitals or postacute settings.
HM rotations also pair clinical experiences with didactic education on relevant topics (eg, billing and coding). The Cleveland, Minnesota, and Virginia tracks developed clinical rotations reflecting the common 7-on and 7-off schedule with nonclinical obligations, such as seminars linking specific content to clinical experiences, during nonclinical time.
DISCUSSION
Our investigation into the current state of HM training found that HM track curricula focus largely on QI, health care economics, and professional development. This focus likely developed in response to hospitalists’ increasing engagement in related endeavors. HM tracks have dynamic and variable structures, reflecting an evolving field and the need to fit into existing IM residency program structures. Similarly, the content covered in HM tracks is tightly linked to perceived opportunities within IM residency curricula. The heterogeneity of content suggests the breadth and ambiguity of necessary competencies for aspiring hospitalists. One of the 11 tracks has not had any residents enroll within the past few years—a testament to the continued effort necessary to sustain such tracks, including curricular updates and recruiting. Conversely, many programs now share track content with the larger IM residency program, suggesting HM tracks may be near the forefront of medical education in some areas.
Our study had several limitations. As we are unaware of any databases of HM tracks, we discussed tracks with professional contacts, performed Internet searches, and reviewed IM residency program websites. Our search, however, was not exhaustive; despite our best efforts, we may have missed or mischaracterized some track offerings. Nevertheless, we think that our analysis represents the first thorough compilation of HM tracks and that it will be useful to institutions seeking to create or enhance HM-specific training.
As the field continues to evolve, we are optimistic about the future of HM training. We suspect that HM residency training tracks will continue to expand. More work is needed so these tracks can adjust to the changing HM and IM residency program landscapes and supply well-trained physicians for the HM workforce.
Acknowledgment
The authors thank track directors Alpesh Amin, David Gugliotti, Rick Hilger, Karnjit Johl, Nasir Majeed, Georgia McIntosh, Charles Pizanis, and Jeff Wiese for making this study possible.
Disclosure
Nothing to report.
References
1. Glasheen JJ, Siegal EM, Epstein K, Kutner J, Prochazka AV. Fulfilling the promise of hospital medicine: tailoring internal medicine training to address hospitalists’ needs [published correction appears in J Gen Intern Med. 2008;23(11):1931]. J Gen Intern Med. 2008;23(7):1110-1115. PubMed 2. Arora V, Guardiano S, Donaldson D, Storch I, Hemstreet P. Closing the gap between internal medicine training and practice: recommendations from recent graduates. Am J Med. 2005;118(6):680-685. PubMed 3. Glasheen JJ, Goldenberg J, Nelson JR. Achieving hospital medicine’s promise through internal medicine residency redesign. Mt Sinai J Med. 2008;75(5):436-441. PubMed 4. Wiese J. Residency training: beginning with the end in mind. J Gen Intern Med. 2008;23(7):1122-1123. PubMed 5. Glasheen JJ, Epstein KR, Siegal E, Kutner JS, Prochazka AV. The spectrum of community-based hospitalist practice: a call to tailor internal medicine residency training. Arch Intern Med. 2007;167(7):727-728. PubMed 6. Patel N, Brennan PJ, Metlay J, Bellini L, Shannon RP, Myers JS. Building the pipeline: the creation of a residency training pathway for future physician leaders in health care quality. Acad Med. 2015;90(2):185-190. PubMed 7. Kumar A, Smeraglio A, Witteles R, et al. A resident-created hospitalist curriculum for internal medicine housestaff. J Hosp Med. 2016;11(9):646-649. PubMed 8. Institute for Healthcare Improvement website. http://www.ihi.org. Accessed December 15, 2015.
References
1. Glasheen JJ, Siegal EM, Epstein K, Kutner J, Prochazka AV. Fulfilling the promise of hospital medicine: tailoring internal medicine training to address hospitalists’ needs [published correction appears in J Gen Intern Med. 2008;23(11):1931]. J Gen Intern Med. 2008;23(7):1110-1115. PubMed 2. Arora V, Guardiano S, Donaldson D, Storch I, Hemstreet P. Closing the gap between internal medicine training and practice: recommendations from recent graduates. Am J Med. 2005;118(6):680-685. PubMed 3. Glasheen JJ, Goldenberg J, Nelson JR. Achieving hospital medicine’s promise through internal medicine residency redesign. Mt Sinai J Med. 2008;75(5):436-441. PubMed 4. Wiese J. Residency training: beginning with the end in mind. J Gen Intern Med. 2008;23(7):1122-1123. PubMed 5. Glasheen JJ, Epstein KR, Siegal E, Kutner JS, Prochazka AV. The spectrum of community-based hospitalist practice: a call to tailor internal medicine residency training. Arch Intern Med. 2007;167(7):727-728. PubMed 6. Patel N, Brennan PJ, Metlay J, Bellini L, Shannon RP, Myers JS. Building the pipeline: the creation of a residency training pathway for future physician leaders in health care quality. Acad Med. 2015;90(2):185-190. PubMed 7. Kumar A, Smeraglio A, Witteles R, et al. A resident-created hospitalist curriculum for internal medicine housestaff. J Hosp Med. 2016;11(9):646-649. PubMed 8. Institute for Healthcare Improvement website. http://www.ihi.org. Accessed December 15, 2015.
Address for correspondence and reprint requests: Joseph R. Sweigart, MD, Internal Medicine, Albert B. Chandler Hospital, University of Kentucky, 800 Rose St, MN602, Lexington, KY 40536-0294; Telephone: 859-323-6047; Fax: 859-257-3873; E-mail: joseph.sweigart@uky.edu
Content Gating
Gated (full article locked unless allowed per User)
Psychiatric conditions affect 1 in 5 children,1,2 and having a comorbid psychiatric condition is associated with worse outcomes in children hospitalized for medical or surgical indications.3-7 Although little is known about interventions for improving outcomes for hospitalized children with psychiatric conditions,8 several interventions that integrate medical and psychiatric care are known to improve ambulatory patient outcomes.9-14 The success of initiatives that test whether integrated medical and psychiatric care models can improve pediatric hospital outcomes depends on reliable identification of comorbid psychiatric conditions and family and clinician having a shared understanding of a patient’s psychiatric diagnoses.
Mental health care system fragmentation, stigma, and privacy issues15-20 may contribute to clinical teams and families having disparate views of psychiatric comorbidities. Evidence suggests that hospital clinicians caring for pediatric medical and surgical inpatients are often unaware of a psychiatric condition that has been diagnosed or managed in the ambulatory setting,3,6 even in cases in which the patient and family are aware of the diagnosis. Conversely, for other patients, clinicians may be aware of a psychiatric diagnosis, but patient and family may not share that understanding or reliably report a psychiatric diagnosis.21-23 Although hospitalization may not be the ideal setting for identifying a new psychiatric diagnosis, given the short-term relationship between patient and clinical care team, addressing and managing a psychiatric comorbidity that is known to family or clinician are important elements of patient-centered hospital care.
The success of interventions in improving hospital outcomes for hospitalized children with psychiatric comorbidity depends on patients, families, and clinicians having a shared understanding of which patients have psychiatric conditions, and on accurate estimates of the scope of the population in need of psychiatric care during pediatric hospitalization.
We conducted a study to compare estimates of point prevalence of psychiatric comorbidity identified by family report (FR) or clinician documentation (CD) and to determine the degree of FR–CD agreement regarding the presence of psychiatric comorbidity in hospitalized children.
METHODS
We estimated point prevalence and determined FR–CD agreement regarding diagnosed psychiatric comorbidities in a cross-sectional sample of pediatric medical and surgical hospitalizations at Children’s Hospital of Philadelphia (CHOP). CHOP is a free-standing 535-bed children’s hospital that serves as a community hospital for the city of Philadelphia; a regional referral center for eastern Pennsylvania, Delaware, and southern New Jersey; and a national and international quaternary referral center. This study was approved by CHOP’s institutional review board.
Patients eligible for inclusion in the study were 4 to 21 years old and hospitalized for a medical or surgical indication. Patients were ineligible if they were hospitalized for a primary psychiatric indication, were medically unstable (eg, received end-of-life care or escalating interventions for a life-threatening condition), had significant cognitive impairment precluding communication (eg, history of severe hypoxic-ischemic encephalopathy), or did not speak English (pertains to consenting parent, guardian, or patient).
The cross-sectional patient sample was selected using a point prevalence recruitment strategy. All eligible patients on each of CHOP’s 20 inpatient medical, surgical, and critical care units were approached for study participation on 2 dates between July 2015 and March 2016. To avoid enrolling the same patient multiple times for a single hospitalization, we separated recruitment dates on each unit by at least 3 months. A goal sample size of 100 to 150 patients was selected to provide precision sufficient to achieve a confidence interval (CI) of 10% around an estimate of the point prevalence of any mental health condition.
To obtain family report of prior psychiatric diagnoses, we interviewed patients and/or their parents during the hospitalization. For 18- to 21-year-old patients, the adolescent patient completed the interview. For patients under 18 years old, parents completed the interview, and for 14- to 17-year-old adolescents,either the parent, the patient, or both could complete the interview. Adolescents were asked to complete the interview confidentially without a parent present. The structured interview included questions derived from the National Survey of Children’s Health24 and the Services Assessment for Children and Adolescents22 to report the patient’s active psychiatric conditions. Interviewees reported whether the patient had ever been diagnosed with any psychiatric disorder, whether the condition was ongoing in the year prior to hospitalization, and whether the patient received any mental health services in clinical settings or school in the 12 months prior to hospitalization.
For CD, we identified a psychiatric diagnosis associated with the index hospitalization if a psychiatric diagnosis was noted in the patient’s admission note, discharge summary, or hospital problem list, or if an International Classification of Diseases (ICD) code for a psychiatric diagnosis was submitted for billing for the index hospitalization. The Healthcare Cost and Utilization Project condition classification system was used to sort psychiatric condition codes25-27 into 5 categories: attention-deficit/hyperactivity disorder (ADHD), anxiety disorders, depression, disruptive behavior disorders, and autism spectrum disorders. A residual category of other, less common psychiatric conditions included eating disorders, attachment disorders, and bipolar disorder.
For each condition category, we determined the point prevalence of having a psychiatric diagnosis identified by FR and having a diagnosis identified by CD. We used McNemar tests to compare point prevalence estimates, the Clopper-Pearson method to calculate CIs around the estimates,28 and Cohen κ statistics to estimate FR–CD agreement regarding psychiatric diagnoses, grouping patients by type of psychiatric diagnosis and by clinical and demographic characteristics. All statistical tests were 2-sided, and P < 0.05 was used for statistical significance. All statistical analyses were performed with Stata Version 13.1 (StataCorp, College Station, Texas).
RESULTS
Of 640 patients hospitalized on study recruitment dates, 411 were ineligible for the study (282 were <4 or >21 years old, 42 were not English speakers, 37 had cognitive impairment, 30 were not medically stable, and 20 were admitted for a primary psychiatric diagnosis). Of the 229 eligible patients, 119 (52%) enrolled. Included patients were 57% female; 9% Hispanic; and 35% black, 55% white, and 15% other race. Forty-eight percent of the enrollees had Medicaid (48%), and 52% had private health insurance. Mean age was 12.3 years. Of enrolled patients, 38% were admitted to subspecialty medical services. Enrollee demographics were representative of hospital-level demographics for the study-eligible population; there were no significant differences in age, sex, race, ethnicity, payer type, or hospital service admission type between enrollees and patients who declined to participate (all Ps > 0.05). Table 1 lists demographic and clinical characteristics of the complete study sample and of the groups with FR- or CD-identified psychiatric diagnosis.
Table 1
Of 119 enrollees, 26 (22%; 95% CI, 15%-30%) had at least 1 FR-identified comorbid psychiatric diagnosis, and 30 (25%; 95% CI, 17%-33%) had at least 1 CD-identified diagnosis. In 13 cases, adolescents (age, 14-17 years) and their parents both completed the structured interview; there were no discrepancies between interview results.
In total, 39 of 119 patients (33%, 95% CI: 24-42%) had either a family-reported or clinician-documented psychiatric diagnosis at the time of hospitalization. For 17 of 119 patients (14%; 95% CI: 9-22%), family-report and clinician-documentation both identified the patient as having a comorbid psychiatric diagnosis. For 9 of 119 patients (8%; 95% CI: 4-14%) families reported a psychiatric diagnosis, but clinicians did not document one. Conversely, for another 13 of 119 patients (11%; 95% CI: 6-18%), a clinician documented a psychiatric diagnosis but the family did not report one. The Figure shows the point prevalence of family-reported psychiatric diagnoses and clinician-documented psychiatric diagnoses for 5 common psychiatric condition categories.
Figure
The most common psychiatric conditions reported by families or documented by clinicians were ADHD (n=16, 13%), anxiety (n=19, 16%), and depression (n=15, 13%). Estimates of the point prevalence of psychiatric conditions were similar when using either family report or clinician documentation alone to identify patients with psychiatric conditions. However, when family report and clinician documentation were used together to identify patients with psychiatric comorbidity, the estimated point prevalence of psychiatric comorbidity was significantly higher for depression, anxiety, and autism. Table 2 displays the point prevalence of patient- and parent-reported psychiatric diagnoses and clinician-documented psychiatric diagnoses.
Table 2
Although point prevalence estimates were similar for FR- and CD-identified comorbid psychiatric conditions, FR–CD agreement was modest. It was fair for any psychiatric diagnosis (κ = .49; 95% CI, .30-.67), highest for ADHD (κ = .79; 95% CI, .61-.96), and fair or poor for other psychiatric conditions (κ range, .11-.48). Table 3 lists the FR–CD agreement data for psychiatric diagnoses for hospitalized children and adolescents.
Table 3
We compared the distribution of FR and CD psychiatric diagnoses with FR use of mental health services. Of the 119 patients, 47 (39%; 95% CI, 31%-49%) had used mental health services within the year before hospitalization. Of these 47 patients, 15 (32%; 95% CI, 19%-47%) had a psychiatric diagnosis identified by both FR and CD, 6 (13%; 95% CI, 5%-26%) had a diagnosis identified only by FR, 8 (17%; 95% CI, 8%-30%) had a diagnosis identified only by CD, and 18 (38%; 95% CI, 25%-54%) had no FR- or CD-identified diagnosis. For 5 (38%; 95% CI, 14%-68%) of the 13 patients with a CD-only diagnosis, the family reported no use of mental health services within the year before hospitalization.
DISCUSSION
At a tertiary-care children’s hospital, we found high point prevalence of comorbid psychiatric conditions and low agreement between FR- and CD-identified psychiatric conditions. Estimates of the prevalence of psychiatric comorbidity among pediatric medical and surgical inpatients were similar for FR- and CD-identified psychiatric conditions, though each method missed about one third of the cases identified by the other method. FR only and CD only each identified about 1 in 4 or 5 hospitalized children and adolescents with a psychiatric comorbidity. When FR and CD were combined, a comorbid psychiatric diagnosis was identified in about 1 in 3 medical and surgical inpatients aged 4 to 21 years. FR–CD agreement was substantial only for ADHD and was fair to slight for most other psychiatric conditions, including autism, depression, anxiety, and disruptive behavior disorders (eg, conduct disorder, oppositional defiant disorder).
Our finding that psychiatric conditions were more commonly reported by families and documented by clinicians for white patients is consistent with a large body of evidence showing that racial or ethnic minority patients experience more stigma related to mental health diagnoses and use mental health services less.29-33 Families were more likely to report use of mental health services than a known mental health diagnosis. This finding may reflect families’ willingness to use services even if they do not understand or experience stigma related to psychiatric diagnoses. Alternatively, use of mental health services without a diagnosis may reflect clinicians’ willingness to refer a child for services when the child is perceived to have an impairment even in the absence of a clear psychiatric diagnosis.
The low FR–CD agreement regarding psychiatric conditions in hospitalized children and adolescents raises 3 issues for pediatric hospital care. First, earlier studies likely underestimated the prevalence of these conditions. A 2014 study of a national sample found that 13% of children hospitalized for a physical health condition had psychiatric comorbidity.25 That study and other large-scale studies showing a high and increasing prevalence of primary psychiatric conditions in hospitalized children and adolescents have relied on administrative data derived from clinician-documented diagnoses.25-27 Our study findings suggest that reliance on administrative data could result in underestimation of the prevalence of psychiatric comorbidity in hospitalized children by as much as 40%. Pediatric hospitals are reporting a shortage of pediatric mental health specialists.34 Augmenting estimates of the prevalence of psychiatric comorbidity in hospitalized children with reports from other sources, including families or outpatient administrative records, may aid health systems in allocating mental health resources for pediatric inpatients.
The second issue is that the present data suggest that families and clinicians do not share the same information about a child’s psychiatric diagnoses when the child is hospitalized for a medical condition or surgical procedure. Low FR–CD agreement regarding psychiatric diagnoses suggests families and clinical teams are not always “on the same page” about psychiatric needs during hospitalization. Implications of this finding are relevant to inpatient and ambulatory care settings. In cases in which a clinician recognizes a psychiatric condition but the family does not, the family may not seek outpatient treatment. In the present study, one third of patients with a psychiatric diagnosis identified by CD but not FR were not engaged in ambulatory treatment for the condition. Conversely, a psychiatric diagnosis identified by FR but not CD suggests clinical teams lack the skills and knowledge needed to elicit information about psychiatric conditions and their potential relevance to inpatient care. As a result, clinicians may miss opportunities to provide interventions that may improve physical or mental health outcomes. For example, clinical teams with information about a patient’s anxiety disorder may be better able to provide brief interventions to prevent medical treatments from triggering anxiety symptoms and to mitigate the risk for traumatic stress symptoms related to the hospitalization.
The third issue is that anxiety disorders were most likely to be the subject of FR–CD disagreement. This finding identifies children with anxiety disorders as a priority population for research into differences between families and clinicians in understanding patients’ psychiatric diagnoses. Our findings suggest families and clinicians have different views of patients’ anxiety symptoms. Anxiety disorders are a risk factor for worse outcomes in children with chronic physical conditions,3,35-37 and acute hospitalization is associated with posthospital anxiety symptoms.38,39 Thus, anxiety disorders are particularly relevant to hospital care and are a priority for research on the differences between families’ and clinicians’ perspectives on children’s psychiatric diagnoses.
Our findings should be interpreted in the context of study limitations. First, because of resource limitations, we did not obtain psychiatric diagnostic evaluations or records to confirm FR- and CD-identified psychiatric diagnoses. Although this lack of clinical confirmation could have resulted in misclassification bias, the risk of bias was no higher than in many other studies that have successfully used hospital records21,25 and family reports to identify psychiatric comorbidity.40 Second, because the study included only English-speaking patients and families, results cannot be generalized to non-English-speaking populations. Third, this was a single-center study, conducted in a free-standing tertiary-care children’s hospital. Sample size was small, particularly for estimating the prevalence of individual psychiatric conditions. Patient characteristics and clinical practice patterns may differ at other types of hospitals. Larger multicenter studies are warranted. Despite these limitations, our results provide important new information that can further our understanding of the epidemiology of psychiatric conditions in hospitalized children. This information should interest clinical teams caring for children with comorbid physical and mental health conditions.
CONCLUSIONS
Low FR–CD agreement regarding hospitalized children’s psychiatric comorbidities suggests that patients and their families and clinicians do not always share the same information about these comorbidities, and that the prevalence of psychiatric comorbidity in hospitalized children is likely underestimated. To allocate adequate resources for these children, health systems may need to obtain information from multiple sources. Furthermore, we need to better our understanding of strategies for communicating about hospitalized children’s psychiatric conditions so that we can develop interventions to improve hospital outcomes for this vulnerable population.
Disclosures
The direct costs of this project were funded by an internal pilot grant from the Center for Pediatric Health Disparities at Children’s Hospital of Philadelphia. Dr. Doupnik was supported by Ruth L. Kirschstein National Research Service Award T32-HP010026-11, funded by the National Institutes of Health. The sponsors had no role in study design; collection, analysis, or interpretation of data; manuscript writing; or deciding to submit this article for publication.
References
1. Perou R, Bitsko RH, Blumberg SJ, et al; Centers for Disease Control and Prevention (CDC). Mental health surveillance among children—United States, 2005-2011. MMWR Suppl. 2013;62(2):1-35. PubMed 2. Merikangas KR, Nakamura EF, Kessler RC. Epidemiology of mental disorders in children and adolescents. Dialogues Clin Neurosci. 2009;11(1):7-20. PubMed 3. Doupnik SK, Mitra N, Feudtner C, Marcus SC. The influence of comorbid mood and anxiety disorders on outcomes of pediatric patients hospitalized for pneumonia. Hosp Pediatr. 2016;6(3):135-142. PubMed 4. Snell C, Fernandes S, Bujoreanu IS, Garcia G. Depression, illness severity, and healthcare utilization in cystic fibrosis. Pediatr Pulmonol. 2014;49(12):1177-1181. PubMed 5. Garrison MM, Katon WJ, Richardson LP. The impact of psychiatric comorbidities on readmissions for diabetes in youth. Diabetes Care. 2005;28(9):2150-2154. PubMed 6. Myrvik MP, Campbell AD, Davis MM, Butcher JL. Impact of psychiatric diagnoses on hospital length of stay in children with sickle cell anemia. Pediatr Blood Cancer. 2012;58(2):239-243. PubMed 7. Myrvik MP, Burks LM, Hoffman RG, Dasgupta M, Panepinto JA. Mental health disorders influence admission rates for pain in children with sickle cell disease. Pediatr Blood Cancer. 2013;60(7):1211-1214. PubMed 8. Bujoreanu S, White MT, Gerber B, Ibeziako P. Effect of timing of psychiatry consultation on length of pediatric hospitalization and hospital charges. Hosp Pediatr. 2015;5(5):269-275. PubMed 9. Archer J, Bower P, Gilbody S, et al. Collaborative care for depression and anxiety problems. Cochrane Database Syst Rev. 2012;10:CD006525. PubMed 10. Aupont O, Doerfler L, Connor DF, Stille C, Tisminetzky M, McLaughlin TJ. A collaborative care model to improve access to pediatric mental health services. Adm Policy Ment Health. 2013;40(4):264-273. PubMed 11. Kolko DJ, Campo J, Kilbourne AM, Hart J, Sakolsky D, Wisniewski S. Collaborative care outcomes for pediatric behavioral health problems: a cluster randomized trial. Pediatrics. 2014;133(4):e981-e992. PubMed 12. Richardson L, McCauley E, Katon W. Collaborative care for adolescent depression: a pilot study. Gen Hosp Psychiatry. 2009;31(1):36-45. PubMed 13. Richardson LP, Ludman E, McCauley E, et al. Collaborative care for adolescents with depression in primary care: a randomized clinical trial. JAMA. 2014;312(8):809-816. PubMed 14. Huffman JC, Mastromauro CA, Beach SR, et al. Collaborative care for depression and anxiety disorders in patients with recent cardiac events: the Management of Sadness and Anxiety in Cardiology (MOSAIC) randomized clinical trial. JAMA Intern Med. 2014;174(6):927-935. PubMed 15. Henderson C, Noblett J, Parke H, et al. Mental health-related stigma in health care and mental health-care settings. Lancet Psychiatry. 2014;1(6):467-482. PubMed 16. Pescosolido BA, Perry BL, Martin JK, McLeod JD, Jensen PS. Stigmatizing attitudes and beliefs about treatment and psychiatric medications for children with mental illness. Psychiatr Serv. 2007;58(5):613-618. PubMed 17. Pescosolido BA. Culture, children, and mental health treatment: special section on the National Stigma Study–Children. Psychiatr Serv. 2007;58(5):611-612. PubMed 18. Britt TW, Greene-Shortridge TM, Brink S, et al. Perceived stigma and barriers to care for psychological treatment: implications for reactions to stressors in different contexts. J Soc Clin Psychol. 2008;27(4):317-335. 19. Thornicroft G, Mehta N, Clement S, et al. Evidence for effective interventions to reduce mental-health-related stigma and discrimination. Lancet. 2016;387(10023):1123-1132. PubMed 20. Health Insurance Portability and Accountability Act of 1996. Washington, DC: US Government Publishing Office; 1996. https://www.gpo.gov/fdsys/pkg/CRPT-104hrpt736/pdf/CRPT-104hrpt736.pdf. Accessed January 21, 2017. 21. Frayne SM, Miller DR, Sharkansky EJ, et al. Using administrative data to identify mental illness: what approach is best? Am J Med Qual. 2010;25(1):42-50. PubMed 22. Horwitz SM, Hoagwood K, Stiffman AR, et al. Reliability of the Services Assessment for Children and Adolescents. Psychiatr Serv. 2001;52(8):1088-1094. PubMed 23. Ascher BH, Farmer EM, Burns BJ, Angold A. The Child and Adolescent Services Assessment (CASA): description and psychometrics. J Emot Behav Disord. 1996;4(1):12-20. 24. Data Resource Center for Child & Adolescent Health, Child and Adolescent Health Measurement Initiative, Oregon Health & Science University. National Survey of Children’s Health (NSCH), 2011/12. http://childhealthdata.org/docs/drc/2011-12-guide-to-topics-questions-draft.pdf?sfvrsn=4. Accessed January 21, 2017. 25. Bardach NS, Coker TR, Zima BT, et al. Common and costly hospitalizations for pediatric mental health disorders. Pediatrics. 2014;133(4):602-609. PubMed 26. Torio CM, Encinosa W, Berdahl T, McCormick MC, Simpson LA. Annual report on health care for children and youth in the United States: national estimates of cost, utilization and expenditures for children with mental health conditions. Acad Pediatr. 2015;15(1):19-35. PubMed 27. Zima BT, Rodean J, Hall M, Bardach NS, Coker TR, Berry JG. Ten year national trends in pediatric hospitalizations by psychiatric complexity. Abstract presented at: 62nd Annual Meeting of the American Academy of Child and Adolescent Psychiatry; October 2015; San Antonio, TX. 28. Clopper CJ, Pearson ES. The use of confidence or fiducial limits illustrated in the case of the binomial. Biometrika. 1934;26:404-413. 29. Wu CH, Erickson SR, Piette JD, Balkrishnan R. Mental health resource utilization and health care costs associated with race and comorbid anxiety among Medicaid enrollees with major depressive disorder. J Nat Med Assoc. 2012;104(1-2):78-88. PubMed 30. Mapelli E, Black T, Doan Q. Trends in pediatric emergency department utilization for mental health-related visits. J Pediatr. 2015;167(4):905-910. PubMed 31. Berdahl T, Owens PL, Dougherty D, McCormick MC, Pylypchuk Y, Simpson LA. Annual report on health care for children and youth in the United States: racial/ethnic and socioeconomic disparities in children’s health care quality. Acad Pediatr. 2010;10(2):95-118. PubMed 32. Flores G; Committee on Pediatric Research. Technical report—racial and ethnic disparities in the health and health care of children. Pediatrics. 2010;125(4):e979-e1020. PubMed 33. Turner EA, Jensen-Doss A, Heffer RW. Ethnicity as a moderator of how parents’ attitudes and perceived stigma influence intentions to seek child mental health services. Cultur Divers Ethnic Minor Psychol. 2015;21(4):613-618. PubMed 34. Shaw RJ, Wamboldt M, Bursch B, Stuber M. Practice patterns in pediatric consultation-liaison psychiatry: a national survey. Psychosomatics. 2006;47(1):43-49. PubMed 35. Benton TD, Ifeagwu JA, Smith-Whitley K. Anxiety and depression in children and adolescents with sickle cell disease. Curr Psychiatry Rep. 2007;9(2):114-121. PubMed 36. Chavira DA, Garland AF, Daley S, Hough R. The impact of medical comorbidity on mental health and functional health outcomes among children with anxiety disorders. J Dev Behav Pediatr. 2008;29(5):394-402. PubMed 37. Knight A, Weiss P, Morales K, et al. Depression and anxiety and their association with healthcare utilization in pediatric lupus and mixed connective tissue disease patients: a cross-sectional study. Pediatr Rheumatol Online J. 2014;12:42. PubMed 38. Marsac ML, Kassam-Adams N, Delahanty DL, F. Widaman K, Barakat LP. Posttraumatic stress following acute medical trauma in children: a proposed model of bio-psycho-social processes during the peri-trauma period. Clin Child Fam Psychol Rev. 2014;17(4):399-411. PubMed 39. Marsac ML, Hildenbrand AK, Kohser KL, Winston FK, Li Y, Kassam-Adams N. Preventing posttraumatic stress following pediatric injury: a randomized controlled trial of a web-based psycho-educational intervention for parents. J Pediatr Psychol. 2013;38(10):1101-1111. PubMed 40. Petersen MC, Kube DA, Whitaker TM, Graff JC, Palmer FB. Prevalence of developmental and behavioral disorders in a pediatric hospital. Pediatrics. 2009;123(3):e490-e495. PubMed
Psychiatric conditions affect 1 in 5 children,1,2 and having a comorbid psychiatric condition is associated with worse outcomes in children hospitalized for medical or surgical indications.3-7 Although little is known about interventions for improving outcomes for hospitalized children with psychiatric conditions,8 several interventions that integrate medical and psychiatric care are known to improve ambulatory patient outcomes.9-14 The success of initiatives that test whether integrated medical and psychiatric care models can improve pediatric hospital outcomes depends on reliable identification of comorbid psychiatric conditions and family and clinician having a shared understanding of a patient’s psychiatric diagnoses.
Mental health care system fragmentation, stigma, and privacy issues15-20 may contribute to clinical teams and families having disparate views of psychiatric comorbidities. Evidence suggests that hospital clinicians caring for pediatric medical and surgical inpatients are often unaware of a psychiatric condition that has been diagnosed or managed in the ambulatory setting,3,6 even in cases in which the patient and family are aware of the diagnosis. Conversely, for other patients, clinicians may be aware of a psychiatric diagnosis, but patient and family may not share that understanding or reliably report a psychiatric diagnosis.21-23 Although hospitalization may not be the ideal setting for identifying a new psychiatric diagnosis, given the short-term relationship between patient and clinical care team, addressing and managing a psychiatric comorbidity that is known to family or clinician are important elements of patient-centered hospital care.
The success of interventions in improving hospital outcomes for hospitalized children with psychiatric comorbidity depends on patients, families, and clinicians having a shared understanding of which patients have psychiatric conditions, and on accurate estimates of the scope of the population in need of psychiatric care during pediatric hospitalization.
We conducted a study to compare estimates of point prevalence of psychiatric comorbidity identified by family report (FR) or clinician documentation (CD) and to determine the degree of FR–CD agreement regarding the presence of psychiatric comorbidity in hospitalized children.
METHODS
We estimated point prevalence and determined FR–CD agreement regarding diagnosed psychiatric comorbidities in a cross-sectional sample of pediatric medical and surgical hospitalizations at Children’s Hospital of Philadelphia (CHOP). CHOP is a free-standing 535-bed children’s hospital that serves as a community hospital for the city of Philadelphia; a regional referral center for eastern Pennsylvania, Delaware, and southern New Jersey; and a national and international quaternary referral center. This study was approved by CHOP’s institutional review board.
Patients eligible for inclusion in the study were 4 to 21 years old and hospitalized for a medical or surgical indication. Patients were ineligible if they were hospitalized for a primary psychiatric indication, were medically unstable (eg, received end-of-life care or escalating interventions for a life-threatening condition), had significant cognitive impairment precluding communication (eg, history of severe hypoxic-ischemic encephalopathy), or did not speak English (pertains to consenting parent, guardian, or patient).
The cross-sectional patient sample was selected using a point prevalence recruitment strategy. All eligible patients on each of CHOP’s 20 inpatient medical, surgical, and critical care units were approached for study participation on 2 dates between July 2015 and March 2016. To avoid enrolling the same patient multiple times for a single hospitalization, we separated recruitment dates on each unit by at least 3 months. A goal sample size of 100 to 150 patients was selected to provide precision sufficient to achieve a confidence interval (CI) of 10% around an estimate of the point prevalence of any mental health condition.
To obtain family report of prior psychiatric diagnoses, we interviewed patients and/or their parents during the hospitalization. For 18- to 21-year-old patients, the adolescent patient completed the interview. For patients under 18 years old, parents completed the interview, and for 14- to 17-year-old adolescents,either the parent, the patient, or both could complete the interview. Adolescents were asked to complete the interview confidentially without a parent present. The structured interview included questions derived from the National Survey of Children’s Health24 and the Services Assessment for Children and Adolescents22 to report the patient’s active psychiatric conditions. Interviewees reported whether the patient had ever been diagnosed with any psychiatric disorder, whether the condition was ongoing in the year prior to hospitalization, and whether the patient received any mental health services in clinical settings or school in the 12 months prior to hospitalization.
For CD, we identified a psychiatric diagnosis associated with the index hospitalization if a psychiatric diagnosis was noted in the patient’s admission note, discharge summary, or hospital problem list, or if an International Classification of Diseases (ICD) code for a psychiatric diagnosis was submitted for billing for the index hospitalization. The Healthcare Cost and Utilization Project condition classification system was used to sort psychiatric condition codes25-27 into 5 categories: attention-deficit/hyperactivity disorder (ADHD), anxiety disorders, depression, disruptive behavior disorders, and autism spectrum disorders. A residual category of other, less common psychiatric conditions included eating disorders, attachment disorders, and bipolar disorder.
For each condition category, we determined the point prevalence of having a psychiatric diagnosis identified by FR and having a diagnosis identified by CD. We used McNemar tests to compare point prevalence estimates, the Clopper-Pearson method to calculate CIs around the estimates,28 and Cohen κ statistics to estimate FR–CD agreement regarding psychiatric diagnoses, grouping patients by type of psychiatric diagnosis and by clinical and demographic characteristics. All statistical tests were 2-sided, and P < 0.05 was used for statistical significance. All statistical analyses were performed with Stata Version 13.1 (StataCorp, College Station, Texas).
RESULTS
Of 640 patients hospitalized on study recruitment dates, 411 were ineligible for the study (282 were <4 or >21 years old, 42 were not English speakers, 37 had cognitive impairment, 30 were not medically stable, and 20 were admitted for a primary psychiatric diagnosis). Of the 229 eligible patients, 119 (52%) enrolled. Included patients were 57% female; 9% Hispanic; and 35% black, 55% white, and 15% other race. Forty-eight percent of the enrollees had Medicaid (48%), and 52% had private health insurance. Mean age was 12.3 years. Of enrolled patients, 38% were admitted to subspecialty medical services. Enrollee demographics were representative of hospital-level demographics for the study-eligible population; there were no significant differences in age, sex, race, ethnicity, payer type, or hospital service admission type between enrollees and patients who declined to participate (all Ps > 0.05). Table 1 lists demographic and clinical characteristics of the complete study sample and of the groups with FR- or CD-identified psychiatric diagnosis.
Table 1
Of 119 enrollees, 26 (22%; 95% CI, 15%-30%) had at least 1 FR-identified comorbid psychiatric diagnosis, and 30 (25%; 95% CI, 17%-33%) had at least 1 CD-identified diagnosis. In 13 cases, adolescents (age, 14-17 years) and their parents both completed the structured interview; there were no discrepancies between interview results.
In total, 39 of 119 patients (33%, 95% CI: 24-42%) had either a family-reported or clinician-documented psychiatric diagnosis at the time of hospitalization. For 17 of 119 patients (14%; 95% CI: 9-22%), family-report and clinician-documentation both identified the patient as having a comorbid psychiatric diagnosis. For 9 of 119 patients (8%; 95% CI: 4-14%) families reported a psychiatric diagnosis, but clinicians did not document one. Conversely, for another 13 of 119 patients (11%; 95% CI: 6-18%), a clinician documented a psychiatric diagnosis but the family did not report one. The Figure shows the point prevalence of family-reported psychiatric diagnoses and clinician-documented psychiatric diagnoses for 5 common psychiatric condition categories.
Figure
The most common psychiatric conditions reported by families or documented by clinicians were ADHD (n=16, 13%), anxiety (n=19, 16%), and depression (n=15, 13%). Estimates of the point prevalence of psychiatric conditions were similar when using either family report or clinician documentation alone to identify patients with psychiatric conditions. However, when family report and clinician documentation were used together to identify patients with psychiatric comorbidity, the estimated point prevalence of psychiatric comorbidity was significantly higher for depression, anxiety, and autism. Table 2 displays the point prevalence of patient- and parent-reported psychiatric diagnoses and clinician-documented psychiatric diagnoses.
Table 2
Although point prevalence estimates were similar for FR- and CD-identified comorbid psychiatric conditions, FR–CD agreement was modest. It was fair for any psychiatric diagnosis (κ = .49; 95% CI, .30-.67), highest for ADHD (κ = .79; 95% CI, .61-.96), and fair or poor for other psychiatric conditions (κ range, .11-.48). Table 3 lists the FR–CD agreement data for psychiatric diagnoses for hospitalized children and adolescents.
Table 3
We compared the distribution of FR and CD psychiatric diagnoses with FR use of mental health services. Of the 119 patients, 47 (39%; 95% CI, 31%-49%) had used mental health services within the year before hospitalization. Of these 47 patients, 15 (32%; 95% CI, 19%-47%) had a psychiatric diagnosis identified by both FR and CD, 6 (13%; 95% CI, 5%-26%) had a diagnosis identified only by FR, 8 (17%; 95% CI, 8%-30%) had a diagnosis identified only by CD, and 18 (38%; 95% CI, 25%-54%) had no FR- or CD-identified diagnosis. For 5 (38%; 95% CI, 14%-68%) of the 13 patients with a CD-only diagnosis, the family reported no use of mental health services within the year before hospitalization.
DISCUSSION
At a tertiary-care children’s hospital, we found high point prevalence of comorbid psychiatric conditions and low agreement between FR- and CD-identified psychiatric conditions. Estimates of the prevalence of psychiatric comorbidity among pediatric medical and surgical inpatients were similar for FR- and CD-identified psychiatric conditions, though each method missed about one third of the cases identified by the other method. FR only and CD only each identified about 1 in 4 or 5 hospitalized children and adolescents with a psychiatric comorbidity. When FR and CD were combined, a comorbid psychiatric diagnosis was identified in about 1 in 3 medical and surgical inpatients aged 4 to 21 years. FR–CD agreement was substantial only for ADHD and was fair to slight for most other psychiatric conditions, including autism, depression, anxiety, and disruptive behavior disorders (eg, conduct disorder, oppositional defiant disorder).
Our finding that psychiatric conditions were more commonly reported by families and documented by clinicians for white patients is consistent with a large body of evidence showing that racial or ethnic minority patients experience more stigma related to mental health diagnoses and use mental health services less.29-33 Families were more likely to report use of mental health services than a known mental health diagnosis. This finding may reflect families’ willingness to use services even if they do not understand or experience stigma related to psychiatric diagnoses. Alternatively, use of mental health services without a diagnosis may reflect clinicians’ willingness to refer a child for services when the child is perceived to have an impairment even in the absence of a clear psychiatric diagnosis.
The low FR–CD agreement regarding psychiatric conditions in hospitalized children and adolescents raises 3 issues for pediatric hospital care. First, earlier studies likely underestimated the prevalence of these conditions. A 2014 study of a national sample found that 13% of children hospitalized for a physical health condition had psychiatric comorbidity.25 That study and other large-scale studies showing a high and increasing prevalence of primary psychiatric conditions in hospitalized children and adolescents have relied on administrative data derived from clinician-documented diagnoses.25-27 Our study findings suggest that reliance on administrative data could result in underestimation of the prevalence of psychiatric comorbidity in hospitalized children by as much as 40%. Pediatric hospitals are reporting a shortage of pediatric mental health specialists.34 Augmenting estimates of the prevalence of psychiatric comorbidity in hospitalized children with reports from other sources, including families or outpatient administrative records, may aid health systems in allocating mental health resources for pediatric inpatients.
The second issue is that the present data suggest that families and clinicians do not share the same information about a child’s psychiatric diagnoses when the child is hospitalized for a medical condition or surgical procedure. Low FR–CD agreement regarding psychiatric diagnoses suggests families and clinical teams are not always “on the same page” about psychiatric needs during hospitalization. Implications of this finding are relevant to inpatient and ambulatory care settings. In cases in which a clinician recognizes a psychiatric condition but the family does not, the family may not seek outpatient treatment. In the present study, one third of patients with a psychiatric diagnosis identified by CD but not FR were not engaged in ambulatory treatment for the condition. Conversely, a psychiatric diagnosis identified by FR but not CD suggests clinical teams lack the skills and knowledge needed to elicit information about psychiatric conditions and their potential relevance to inpatient care. As a result, clinicians may miss opportunities to provide interventions that may improve physical or mental health outcomes. For example, clinical teams with information about a patient’s anxiety disorder may be better able to provide brief interventions to prevent medical treatments from triggering anxiety symptoms and to mitigate the risk for traumatic stress symptoms related to the hospitalization.
The third issue is that anxiety disorders were most likely to be the subject of FR–CD disagreement. This finding identifies children with anxiety disorders as a priority population for research into differences between families and clinicians in understanding patients’ psychiatric diagnoses. Our findings suggest families and clinicians have different views of patients’ anxiety symptoms. Anxiety disorders are a risk factor for worse outcomes in children with chronic physical conditions,3,35-37 and acute hospitalization is associated with posthospital anxiety symptoms.38,39 Thus, anxiety disorders are particularly relevant to hospital care and are a priority for research on the differences between families’ and clinicians’ perspectives on children’s psychiatric diagnoses.
Our findings should be interpreted in the context of study limitations. First, because of resource limitations, we did not obtain psychiatric diagnostic evaluations or records to confirm FR- and CD-identified psychiatric diagnoses. Although this lack of clinical confirmation could have resulted in misclassification bias, the risk of bias was no higher than in many other studies that have successfully used hospital records21,25 and family reports to identify psychiatric comorbidity.40 Second, because the study included only English-speaking patients and families, results cannot be generalized to non-English-speaking populations. Third, this was a single-center study, conducted in a free-standing tertiary-care children’s hospital. Sample size was small, particularly for estimating the prevalence of individual psychiatric conditions. Patient characteristics and clinical practice patterns may differ at other types of hospitals. Larger multicenter studies are warranted. Despite these limitations, our results provide important new information that can further our understanding of the epidemiology of psychiatric conditions in hospitalized children. This information should interest clinical teams caring for children with comorbid physical and mental health conditions.
CONCLUSIONS
Low FR–CD agreement regarding hospitalized children’s psychiatric comorbidities suggests that patients and their families and clinicians do not always share the same information about these comorbidities, and that the prevalence of psychiatric comorbidity in hospitalized children is likely underestimated. To allocate adequate resources for these children, health systems may need to obtain information from multiple sources. Furthermore, we need to better our understanding of strategies for communicating about hospitalized children’s psychiatric conditions so that we can develop interventions to improve hospital outcomes for this vulnerable population.
Disclosures
The direct costs of this project were funded by an internal pilot grant from the Center for Pediatric Health Disparities at Children’s Hospital of Philadelphia. Dr. Doupnik was supported by Ruth L. Kirschstein National Research Service Award T32-HP010026-11, funded by the National Institutes of Health. The sponsors had no role in study design; collection, analysis, or interpretation of data; manuscript writing; or deciding to submit this article for publication.
Psychiatric conditions affect 1 in 5 children,1,2 and having a comorbid psychiatric condition is associated with worse outcomes in children hospitalized for medical or surgical indications.3-7 Although little is known about interventions for improving outcomes for hospitalized children with psychiatric conditions,8 several interventions that integrate medical and psychiatric care are known to improve ambulatory patient outcomes.9-14 The success of initiatives that test whether integrated medical and psychiatric care models can improve pediatric hospital outcomes depends on reliable identification of comorbid psychiatric conditions and family and clinician having a shared understanding of a patient’s psychiatric diagnoses.
Mental health care system fragmentation, stigma, and privacy issues15-20 may contribute to clinical teams and families having disparate views of psychiatric comorbidities. Evidence suggests that hospital clinicians caring for pediatric medical and surgical inpatients are often unaware of a psychiatric condition that has been diagnosed or managed in the ambulatory setting,3,6 even in cases in which the patient and family are aware of the diagnosis. Conversely, for other patients, clinicians may be aware of a psychiatric diagnosis, but patient and family may not share that understanding or reliably report a psychiatric diagnosis.21-23 Although hospitalization may not be the ideal setting for identifying a new psychiatric diagnosis, given the short-term relationship between patient and clinical care team, addressing and managing a psychiatric comorbidity that is known to family or clinician are important elements of patient-centered hospital care.
The success of interventions in improving hospital outcomes for hospitalized children with psychiatric comorbidity depends on patients, families, and clinicians having a shared understanding of which patients have psychiatric conditions, and on accurate estimates of the scope of the population in need of psychiatric care during pediatric hospitalization.
We conducted a study to compare estimates of point prevalence of psychiatric comorbidity identified by family report (FR) or clinician documentation (CD) and to determine the degree of FR–CD agreement regarding the presence of psychiatric comorbidity in hospitalized children.
METHODS
We estimated point prevalence and determined FR–CD agreement regarding diagnosed psychiatric comorbidities in a cross-sectional sample of pediatric medical and surgical hospitalizations at Children’s Hospital of Philadelphia (CHOP). CHOP is a free-standing 535-bed children’s hospital that serves as a community hospital for the city of Philadelphia; a regional referral center for eastern Pennsylvania, Delaware, and southern New Jersey; and a national and international quaternary referral center. This study was approved by CHOP’s institutional review board.
Patients eligible for inclusion in the study were 4 to 21 years old and hospitalized for a medical or surgical indication. Patients were ineligible if they were hospitalized for a primary psychiatric indication, were medically unstable (eg, received end-of-life care or escalating interventions for a life-threatening condition), had significant cognitive impairment precluding communication (eg, history of severe hypoxic-ischemic encephalopathy), or did not speak English (pertains to consenting parent, guardian, or patient).
The cross-sectional patient sample was selected using a point prevalence recruitment strategy. All eligible patients on each of CHOP’s 20 inpatient medical, surgical, and critical care units were approached for study participation on 2 dates between July 2015 and March 2016. To avoid enrolling the same patient multiple times for a single hospitalization, we separated recruitment dates on each unit by at least 3 months. A goal sample size of 100 to 150 patients was selected to provide precision sufficient to achieve a confidence interval (CI) of 10% around an estimate of the point prevalence of any mental health condition.
To obtain family report of prior psychiatric diagnoses, we interviewed patients and/or their parents during the hospitalization. For 18- to 21-year-old patients, the adolescent patient completed the interview. For patients under 18 years old, parents completed the interview, and for 14- to 17-year-old adolescents,either the parent, the patient, or both could complete the interview. Adolescents were asked to complete the interview confidentially without a parent present. The structured interview included questions derived from the National Survey of Children’s Health24 and the Services Assessment for Children and Adolescents22 to report the patient’s active psychiatric conditions. Interviewees reported whether the patient had ever been diagnosed with any psychiatric disorder, whether the condition was ongoing in the year prior to hospitalization, and whether the patient received any mental health services in clinical settings or school in the 12 months prior to hospitalization.
For CD, we identified a psychiatric diagnosis associated with the index hospitalization if a psychiatric diagnosis was noted in the patient’s admission note, discharge summary, or hospital problem list, or if an International Classification of Diseases (ICD) code for a psychiatric diagnosis was submitted for billing for the index hospitalization. The Healthcare Cost and Utilization Project condition classification system was used to sort psychiatric condition codes25-27 into 5 categories: attention-deficit/hyperactivity disorder (ADHD), anxiety disorders, depression, disruptive behavior disorders, and autism spectrum disorders. A residual category of other, less common psychiatric conditions included eating disorders, attachment disorders, and bipolar disorder.
For each condition category, we determined the point prevalence of having a psychiatric diagnosis identified by FR and having a diagnosis identified by CD. We used McNemar tests to compare point prevalence estimates, the Clopper-Pearson method to calculate CIs around the estimates,28 and Cohen κ statistics to estimate FR–CD agreement regarding psychiatric diagnoses, grouping patients by type of psychiatric diagnosis and by clinical and demographic characteristics. All statistical tests were 2-sided, and P < 0.05 was used for statistical significance. All statistical analyses were performed with Stata Version 13.1 (StataCorp, College Station, Texas).
RESULTS
Of 640 patients hospitalized on study recruitment dates, 411 were ineligible for the study (282 were <4 or >21 years old, 42 were not English speakers, 37 had cognitive impairment, 30 were not medically stable, and 20 were admitted for a primary psychiatric diagnosis). Of the 229 eligible patients, 119 (52%) enrolled. Included patients were 57% female; 9% Hispanic; and 35% black, 55% white, and 15% other race. Forty-eight percent of the enrollees had Medicaid (48%), and 52% had private health insurance. Mean age was 12.3 years. Of enrolled patients, 38% were admitted to subspecialty medical services. Enrollee demographics were representative of hospital-level demographics for the study-eligible population; there were no significant differences in age, sex, race, ethnicity, payer type, or hospital service admission type between enrollees and patients who declined to participate (all Ps > 0.05). Table 1 lists demographic and clinical characteristics of the complete study sample and of the groups with FR- or CD-identified psychiatric diagnosis.
Table 1
Of 119 enrollees, 26 (22%; 95% CI, 15%-30%) had at least 1 FR-identified comorbid psychiatric diagnosis, and 30 (25%; 95% CI, 17%-33%) had at least 1 CD-identified diagnosis. In 13 cases, adolescents (age, 14-17 years) and their parents both completed the structured interview; there were no discrepancies between interview results.
In total, 39 of 119 patients (33%, 95% CI: 24-42%) had either a family-reported or clinician-documented psychiatric diagnosis at the time of hospitalization. For 17 of 119 patients (14%; 95% CI: 9-22%), family-report and clinician-documentation both identified the patient as having a comorbid psychiatric diagnosis. For 9 of 119 patients (8%; 95% CI: 4-14%) families reported a psychiatric diagnosis, but clinicians did not document one. Conversely, for another 13 of 119 patients (11%; 95% CI: 6-18%), a clinician documented a psychiatric diagnosis but the family did not report one. The Figure shows the point prevalence of family-reported psychiatric diagnoses and clinician-documented psychiatric diagnoses for 5 common psychiatric condition categories.
Figure
The most common psychiatric conditions reported by families or documented by clinicians were ADHD (n=16, 13%), anxiety (n=19, 16%), and depression (n=15, 13%). Estimates of the point prevalence of psychiatric conditions were similar when using either family report or clinician documentation alone to identify patients with psychiatric conditions. However, when family report and clinician documentation were used together to identify patients with psychiatric comorbidity, the estimated point prevalence of psychiatric comorbidity was significantly higher for depression, anxiety, and autism. Table 2 displays the point prevalence of patient- and parent-reported psychiatric diagnoses and clinician-documented psychiatric diagnoses.
Table 2
Although point prevalence estimates were similar for FR- and CD-identified comorbid psychiatric conditions, FR–CD agreement was modest. It was fair for any psychiatric diagnosis (κ = .49; 95% CI, .30-.67), highest for ADHD (κ = .79; 95% CI, .61-.96), and fair or poor for other psychiatric conditions (κ range, .11-.48). Table 3 lists the FR–CD agreement data for psychiatric diagnoses for hospitalized children and adolescents.
Table 3
We compared the distribution of FR and CD psychiatric diagnoses with FR use of mental health services. Of the 119 patients, 47 (39%; 95% CI, 31%-49%) had used mental health services within the year before hospitalization. Of these 47 patients, 15 (32%; 95% CI, 19%-47%) had a psychiatric diagnosis identified by both FR and CD, 6 (13%; 95% CI, 5%-26%) had a diagnosis identified only by FR, 8 (17%; 95% CI, 8%-30%) had a diagnosis identified only by CD, and 18 (38%; 95% CI, 25%-54%) had no FR- or CD-identified diagnosis. For 5 (38%; 95% CI, 14%-68%) of the 13 patients with a CD-only diagnosis, the family reported no use of mental health services within the year before hospitalization.
DISCUSSION
At a tertiary-care children’s hospital, we found high point prevalence of comorbid psychiatric conditions and low agreement between FR- and CD-identified psychiatric conditions. Estimates of the prevalence of psychiatric comorbidity among pediatric medical and surgical inpatients were similar for FR- and CD-identified psychiatric conditions, though each method missed about one third of the cases identified by the other method. FR only and CD only each identified about 1 in 4 or 5 hospitalized children and adolescents with a psychiatric comorbidity. When FR and CD were combined, a comorbid psychiatric diagnosis was identified in about 1 in 3 medical and surgical inpatients aged 4 to 21 years. FR–CD agreement was substantial only for ADHD and was fair to slight for most other psychiatric conditions, including autism, depression, anxiety, and disruptive behavior disorders (eg, conduct disorder, oppositional defiant disorder).
Our finding that psychiatric conditions were more commonly reported by families and documented by clinicians for white patients is consistent with a large body of evidence showing that racial or ethnic minority patients experience more stigma related to mental health diagnoses and use mental health services less.29-33 Families were more likely to report use of mental health services than a known mental health diagnosis. This finding may reflect families’ willingness to use services even if they do not understand or experience stigma related to psychiatric diagnoses. Alternatively, use of mental health services without a diagnosis may reflect clinicians’ willingness to refer a child for services when the child is perceived to have an impairment even in the absence of a clear psychiatric diagnosis.
The low FR–CD agreement regarding psychiatric conditions in hospitalized children and adolescents raises 3 issues for pediatric hospital care. First, earlier studies likely underestimated the prevalence of these conditions. A 2014 study of a national sample found that 13% of children hospitalized for a physical health condition had psychiatric comorbidity.25 That study and other large-scale studies showing a high and increasing prevalence of primary psychiatric conditions in hospitalized children and adolescents have relied on administrative data derived from clinician-documented diagnoses.25-27 Our study findings suggest that reliance on administrative data could result in underestimation of the prevalence of psychiatric comorbidity in hospitalized children by as much as 40%. Pediatric hospitals are reporting a shortage of pediatric mental health specialists.34 Augmenting estimates of the prevalence of psychiatric comorbidity in hospitalized children with reports from other sources, including families or outpatient administrative records, may aid health systems in allocating mental health resources for pediatric inpatients.
The second issue is that the present data suggest that families and clinicians do not share the same information about a child’s psychiatric diagnoses when the child is hospitalized for a medical condition or surgical procedure. Low FR–CD agreement regarding psychiatric diagnoses suggests families and clinical teams are not always “on the same page” about psychiatric needs during hospitalization. Implications of this finding are relevant to inpatient and ambulatory care settings. In cases in which a clinician recognizes a psychiatric condition but the family does not, the family may not seek outpatient treatment. In the present study, one third of patients with a psychiatric diagnosis identified by CD but not FR were not engaged in ambulatory treatment for the condition. Conversely, a psychiatric diagnosis identified by FR but not CD suggests clinical teams lack the skills and knowledge needed to elicit information about psychiatric conditions and their potential relevance to inpatient care. As a result, clinicians may miss opportunities to provide interventions that may improve physical or mental health outcomes. For example, clinical teams with information about a patient’s anxiety disorder may be better able to provide brief interventions to prevent medical treatments from triggering anxiety symptoms and to mitigate the risk for traumatic stress symptoms related to the hospitalization.
The third issue is that anxiety disorders were most likely to be the subject of FR–CD disagreement. This finding identifies children with anxiety disorders as a priority population for research into differences between families and clinicians in understanding patients’ psychiatric diagnoses. Our findings suggest families and clinicians have different views of patients’ anxiety symptoms. Anxiety disorders are a risk factor for worse outcomes in children with chronic physical conditions,3,35-37 and acute hospitalization is associated with posthospital anxiety symptoms.38,39 Thus, anxiety disorders are particularly relevant to hospital care and are a priority for research on the differences between families’ and clinicians’ perspectives on children’s psychiatric diagnoses.
Our findings should be interpreted in the context of study limitations. First, because of resource limitations, we did not obtain psychiatric diagnostic evaluations or records to confirm FR- and CD-identified psychiatric diagnoses. Although this lack of clinical confirmation could have resulted in misclassification bias, the risk of bias was no higher than in many other studies that have successfully used hospital records21,25 and family reports to identify psychiatric comorbidity.40 Second, because the study included only English-speaking patients and families, results cannot be generalized to non-English-speaking populations. Third, this was a single-center study, conducted in a free-standing tertiary-care children’s hospital. Sample size was small, particularly for estimating the prevalence of individual psychiatric conditions. Patient characteristics and clinical practice patterns may differ at other types of hospitals. Larger multicenter studies are warranted. Despite these limitations, our results provide important new information that can further our understanding of the epidemiology of psychiatric conditions in hospitalized children. This information should interest clinical teams caring for children with comorbid physical and mental health conditions.
CONCLUSIONS
Low FR–CD agreement regarding hospitalized children’s psychiatric comorbidities suggests that patients and their families and clinicians do not always share the same information about these comorbidities, and that the prevalence of psychiatric comorbidity in hospitalized children is likely underestimated. To allocate adequate resources for these children, health systems may need to obtain information from multiple sources. Furthermore, we need to better our understanding of strategies for communicating about hospitalized children’s psychiatric conditions so that we can develop interventions to improve hospital outcomes for this vulnerable population.
Disclosures
The direct costs of this project were funded by an internal pilot grant from the Center for Pediatric Health Disparities at Children’s Hospital of Philadelphia. Dr. Doupnik was supported by Ruth L. Kirschstein National Research Service Award T32-HP010026-11, funded by the National Institutes of Health. The sponsors had no role in study design; collection, analysis, or interpretation of data; manuscript writing; or deciding to submit this article for publication.
References
1. Perou R, Bitsko RH, Blumberg SJ, et al; Centers for Disease Control and Prevention (CDC). Mental health surveillance among children—United States, 2005-2011. MMWR Suppl. 2013;62(2):1-35. PubMed 2. Merikangas KR, Nakamura EF, Kessler RC. Epidemiology of mental disorders in children and adolescents. Dialogues Clin Neurosci. 2009;11(1):7-20. PubMed 3. Doupnik SK, Mitra N, Feudtner C, Marcus SC. The influence of comorbid mood and anxiety disorders on outcomes of pediatric patients hospitalized for pneumonia. Hosp Pediatr. 2016;6(3):135-142. PubMed 4. Snell C, Fernandes S, Bujoreanu IS, Garcia G. Depression, illness severity, and healthcare utilization in cystic fibrosis. Pediatr Pulmonol. 2014;49(12):1177-1181. PubMed 5. Garrison MM, Katon WJ, Richardson LP. The impact of psychiatric comorbidities on readmissions for diabetes in youth. Diabetes Care. 2005;28(9):2150-2154. PubMed 6. Myrvik MP, Campbell AD, Davis MM, Butcher JL. Impact of psychiatric diagnoses on hospital length of stay in children with sickle cell anemia. Pediatr Blood Cancer. 2012;58(2):239-243. PubMed 7. Myrvik MP, Burks LM, Hoffman RG, Dasgupta M, Panepinto JA. Mental health disorders influence admission rates for pain in children with sickle cell disease. Pediatr Blood Cancer. 2013;60(7):1211-1214. PubMed 8. Bujoreanu S, White MT, Gerber B, Ibeziako P. Effect of timing of psychiatry consultation on length of pediatric hospitalization and hospital charges. Hosp Pediatr. 2015;5(5):269-275. PubMed 9. Archer J, Bower P, Gilbody S, et al. Collaborative care for depression and anxiety problems. Cochrane Database Syst Rev. 2012;10:CD006525. PubMed 10. Aupont O, Doerfler L, Connor DF, Stille C, Tisminetzky M, McLaughlin TJ. A collaborative care model to improve access to pediatric mental health services. Adm Policy Ment Health. 2013;40(4):264-273. PubMed 11. Kolko DJ, Campo J, Kilbourne AM, Hart J, Sakolsky D, Wisniewski S. Collaborative care outcomes for pediatric behavioral health problems: a cluster randomized trial. Pediatrics. 2014;133(4):e981-e992. PubMed 12. Richardson L, McCauley E, Katon W. Collaborative care for adolescent depression: a pilot study. Gen Hosp Psychiatry. 2009;31(1):36-45. PubMed 13. Richardson LP, Ludman E, McCauley E, et al. Collaborative care for adolescents with depression in primary care: a randomized clinical trial. JAMA. 2014;312(8):809-816. PubMed 14. Huffman JC, Mastromauro CA, Beach SR, et al. Collaborative care for depression and anxiety disorders in patients with recent cardiac events: the Management of Sadness and Anxiety in Cardiology (MOSAIC) randomized clinical trial. JAMA Intern Med. 2014;174(6):927-935. PubMed 15. Henderson C, Noblett J, Parke H, et al. Mental health-related stigma in health care and mental health-care settings. Lancet Psychiatry. 2014;1(6):467-482. PubMed 16. Pescosolido BA, Perry BL, Martin JK, McLeod JD, Jensen PS. Stigmatizing attitudes and beliefs about treatment and psychiatric medications for children with mental illness. Psychiatr Serv. 2007;58(5):613-618. PubMed 17. Pescosolido BA. Culture, children, and mental health treatment: special section on the National Stigma Study–Children. Psychiatr Serv. 2007;58(5):611-612. PubMed 18. Britt TW, Greene-Shortridge TM, Brink S, et al. Perceived stigma and barriers to care for psychological treatment: implications for reactions to stressors in different contexts. J Soc Clin Psychol. 2008;27(4):317-335. 19. Thornicroft G, Mehta N, Clement S, et al. Evidence for effective interventions to reduce mental-health-related stigma and discrimination. Lancet. 2016;387(10023):1123-1132. PubMed 20. Health Insurance Portability and Accountability Act of 1996. Washington, DC: US Government Publishing Office; 1996. https://www.gpo.gov/fdsys/pkg/CRPT-104hrpt736/pdf/CRPT-104hrpt736.pdf. Accessed January 21, 2017. 21. Frayne SM, Miller DR, Sharkansky EJ, et al. Using administrative data to identify mental illness: what approach is best? Am J Med Qual. 2010;25(1):42-50. PubMed 22. Horwitz SM, Hoagwood K, Stiffman AR, et al. Reliability of the Services Assessment for Children and Adolescents. Psychiatr Serv. 2001;52(8):1088-1094. PubMed 23. Ascher BH, Farmer EM, Burns BJ, Angold A. The Child and Adolescent Services Assessment (CASA): description and psychometrics. J Emot Behav Disord. 1996;4(1):12-20. 24. Data Resource Center for Child & Adolescent Health, Child and Adolescent Health Measurement Initiative, Oregon Health & Science University. National Survey of Children’s Health (NSCH), 2011/12. http://childhealthdata.org/docs/drc/2011-12-guide-to-topics-questions-draft.pdf?sfvrsn=4. Accessed January 21, 2017. 25. Bardach NS, Coker TR, Zima BT, et al. Common and costly hospitalizations for pediatric mental health disorders. Pediatrics. 2014;133(4):602-609. PubMed 26. Torio CM, Encinosa W, Berdahl T, McCormick MC, Simpson LA. Annual report on health care for children and youth in the United States: national estimates of cost, utilization and expenditures for children with mental health conditions. Acad Pediatr. 2015;15(1):19-35. PubMed 27. Zima BT, Rodean J, Hall M, Bardach NS, Coker TR, Berry JG. Ten year national trends in pediatric hospitalizations by psychiatric complexity. Abstract presented at: 62nd Annual Meeting of the American Academy of Child and Adolescent Psychiatry; October 2015; San Antonio, TX. 28. Clopper CJ, Pearson ES. The use of confidence or fiducial limits illustrated in the case of the binomial. Biometrika. 1934;26:404-413. 29. Wu CH, Erickson SR, Piette JD, Balkrishnan R. Mental health resource utilization and health care costs associated with race and comorbid anxiety among Medicaid enrollees with major depressive disorder. J Nat Med Assoc. 2012;104(1-2):78-88. PubMed 30. Mapelli E, Black T, Doan Q. Trends in pediatric emergency department utilization for mental health-related visits. J Pediatr. 2015;167(4):905-910. PubMed 31. Berdahl T, Owens PL, Dougherty D, McCormick MC, Pylypchuk Y, Simpson LA. Annual report on health care for children and youth in the United States: racial/ethnic and socioeconomic disparities in children’s health care quality. Acad Pediatr. 2010;10(2):95-118. PubMed 32. Flores G; Committee on Pediatric Research. Technical report—racial and ethnic disparities in the health and health care of children. Pediatrics. 2010;125(4):e979-e1020. PubMed 33. Turner EA, Jensen-Doss A, Heffer RW. Ethnicity as a moderator of how parents’ attitudes and perceived stigma influence intentions to seek child mental health services. Cultur Divers Ethnic Minor Psychol. 2015;21(4):613-618. PubMed 34. Shaw RJ, Wamboldt M, Bursch B, Stuber M. Practice patterns in pediatric consultation-liaison psychiatry: a national survey. Psychosomatics. 2006;47(1):43-49. PubMed 35. Benton TD, Ifeagwu JA, Smith-Whitley K. Anxiety and depression in children and adolescents with sickle cell disease. Curr Psychiatry Rep. 2007;9(2):114-121. PubMed 36. Chavira DA, Garland AF, Daley S, Hough R. The impact of medical comorbidity on mental health and functional health outcomes among children with anxiety disorders. J Dev Behav Pediatr. 2008;29(5):394-402. PubMed 37. Knight A, Weiss P, Morales K, et al. Depression and anxiety and their association with healthcare utilization in pediatric lupus and mixed connective tissue disease patients: a cross-sectional study. Pediatr Rheumatol Online J. 2014;12:42. PubMed 38. Marsac ML, Kassam-Adams N, Delahanty DL, F. Widaman K, Barakat LP. Posttraumatic stress following acute medical trauma in children: a proposed model of bio-psycho-social processes during the peri-trauma period. Clin Child Fam Psychol Rev. 2014;17(4):399-411. PubMed 39. Marsac ML, Hildenbrand AK, Kohser KL, Winston FK, Li Y, Kassam-Adams N. Preventing posttraumatic stress following pediatric injury: a randomized controlled trial of a web-based psycho-educational intervention for parents. J Pediatr Psychol. 2013;38(10):1101-1111. PubMed 40. Petersen MC, Kube DA, Whitaker TM, Graff JC, Palmer FB. Prevalence of developmental and behavioral disorders in a pediatric hospital. Pediatrics. 2009;123(3):e490-e495. PubMed
References
1. Perou R, Bitsko RH, Blumberg SJ, et al; Centers for Disease Control and Prevention (CDC). Mental health surveillance among children—United States, 2005-2011. MMWR Suppl. 2013;62(2):1-35. PubMed 2. Merikangas KR, Nakamura EF, Kessler RC. Epidemiology of mental disorders in children and adolescents. Dialogues Clin Neurosci. 2009;11(1):7-20. PubMed 3. Doupnik SK, Mitra N, Feudtner C, Marcus SC. The influence of comorbid mood and anxiety disorders on outcomes of pediatric patients hospitalized for pneumonia. Hosp Pediatr. 2016;6(3):135-142. PubMed 4. Snell C, Fernandes S, Bujoreanu IS, Garcia G. Depression, illness severity, and healthcare utilization in cystic fibrosis. Pediatr Pulmonol. 2014;49(12):1177-1181. PubMed 5. Garrison MM, Katon WJ, Richardson LP. The impact of psychiatric comorbidities on readmissions for diabetes in youth. Diabetes Care. 2005;28(9):2150-2154. PubMed 6. Myrvik MP, Campbell AD, Davis MM, Butcher JL. Impact of psychiatric diagnoses on hospital length of stay in children with sickle cell anemia. Pediatr Blood Cancer. 2012;58(2):239-243. PubMed 7. Myrvik MP, Burks LM, Hoffman RG, Dasgupta M, Panepinto JA. Mental health disorders influence admission rates for pain in children with sickle cell disease. Pediatr Blood Cancer. 2013;60(7):1211-1214. PubMed 8. Bujoreanu S, White MT, Gerber B, Ibeziako P. Effect of timing of psychiatry consultation on length of pediatric hospitalization and hospital charges. Hosp Pediatr. 2015;5(5):269-275. PubMed 9. Archer J, Bower P, Gilbody S, et al. Collaborative care for depression and anxiety problems. Cochrane Database Syst Rev. 2012;10:CD006525. PubMed 10. Aupont O, Doerfler L, Connor DF, Stille C, Tisminetzky M, McLaughlin TJ. A collaborative care model to improve access to pediatric mental health services. Adm Policy Ment Health. 2013;40(4):264-273. PubMed 11. Kolko DJ, Campo J, Kilbourne AM, Hart J, Sakolsky D, Wisniewski S. Collaborative care outcomes for pediatric behavioral health problems: a cluster randomized trial. Pediatrics. 2014;133(4):e981-e992. PubMed 12. Richardson L, McCauley E, Katon W. Collaborative care for adolescent depression: a pilot study. Gen Hosp Psychiatry. 2009;31(1):36-45. PubMed 13. Richardson LP, Ludman E, McCauley E, et al. Collaborative care for adolescents with depression in primary care: a randomized clinical trial. JAMA. 2014;312(8):809-816. PubMed 14. Huffman JC, Mastromauro CA, Beach SR, et al. Collaborative care for depression and anxiety disorders in patients with recent cardiac events: the Management of Sadness and Anxiety in Cardiology (MOSAIC) randomized clinical trial. JAMA Intern Med. 2014;174(6):927-935. PubMed 15. Henderson C, Noblett J, Parke H, et al. Mental health-related stigma in health care and mental health-care settings. Lancet Psychiatry. 2014;1(6):467-482. PubMed 16. Pescosolido BA, Perry BL, Martin JK, McLeod JD, Jensen PS. Stigmatizing attitudes and beliefs about treatment and psychiatric medications for children with mental illness. Psychiatr Serv. 2007;58(5):613-618. PubMed 17. Pescosolido BA. Culture, children, and mental health treatment: special section on the National Stigma Study–Children. Psychiatr Serv. 2007;58(5):611-612. PubMed 18. Britt TW, Greene-Shortridge TM, Brink S, et al. Perceived stigma and barriers to care for psychological treatment: implications for reactions to stressors in different contexts. J Soc Clin Psychol. 2008;27(4):317-335. 19. Thornicroft G, Mehta N, Clement S, et al. Evidence for effective interventions to reduce mental-health-related stigma and discrimination. Lancet. 2016;387(10023):1123-1132. PubMed 20. Health Insurance Portability and Accountability Act of 1996. Washington, DC: US Government Publishing Office; 1996. https://www.gpo.gov/fdsys/pkg/CRPT-104hrpt736/pdf/CRPT-104hrpt736.pdf. Accessed January 21, 2017. 21. Frayne SM, Miller DR, Sharkansky EJ, et al. Using administrative data to identify mental illness: what approach is best? Am J Med Qual. 2010;25(1):42-50. PubMed 22. Horwitz SM, Hoagwood K, Stiffman AR, et al. Reliability of the Services Assessment for Children and Adolescents. Psychiatr Serv. 2001;52(8):1088-1094. PubMed 23. Ascher BH, Farmer EM, Burns BJ, Angold A. The Child and Adolescent Services Assessment (CASA): description and psychometrics. J Emot Behav Disord. 1996;4(1):12-20. 24. Data Resource Center for Child & Adolescent Health, Child and Adolescent Health Measurement Initiative, Oregon Health & Science University. National Survey of Children’s Health (NSCH), 2011/12. http://childhealthdata.org/docs/drc/2011-12-guide-to-topics-questions-draft.pdf?sfvrsn=4. Accessed January 21, 2017. 25. Bardach NS, Coker TR, Zima BT, et al. Common and costly hospitalizations for pediatric mental health disorders. Pediatrics. 2014;133(4):602-609. PubMed 26. Torio CM, Encinosa W, Berdahl T, McCormick MC, Simpson LA. Annual report on health care for children and youth in the United States: national estimates of cost, utilization and expenditures for children with mental health conditions. Acad Pediatr. 2015;15(1):19-35. PubMed 27. Zima BT, Rodean J, Hall M, Bardach NS, Coker TR, Berry JG. Ten year national trends in pediatric hospitalizations by psychiatric complexity. Abstract presented at: 62nd Annual Meeting of the American Academy of Child and Adolescent Psychiatry; October 2015; San Antonio, TX. 28. Clopper CJ, Pearson ES. The use of confidence or fiducial limits illustrated in the case of the binomial. Biometrika. 1934;26:404-413. 29. Wu CH, Erickson SR, Piette JD, Balkrishnan R. Mental health resource utilization and health care costs associated with race and comorbid anxiety among Medicaid enrollees with major depressive disorder. J Nat Med Assoc. 2012;104(1-2):78-88. PubMed 30. Mapelli E, Black T, Doan Q. Trends in pediatric emergency department utilization for mental health-related visits. J Pediatr. 2015;167(4):905-910. PubMed 31. Berdahl T, Owens PL, Dougherty D, McCormick MC, Pylypchuk Y, Simpson LA. Annual report on health care for children and youth in the United States: racial/ethnic and socioeconomic disparities in children’s health care quality. Acad Pediatr. 2010;10(2):95-118. PubMed 32. Flores G; Committee on Pediatric Research. Technical report—racial and ethnic disparities in the health and health care of children. Pediatrics. 2010;125(4):e979-e1020. PubMed 33. Turner EA, Jensen-Doss A, Heffer RW. Ethnicity as a moderator of how parents’ attitudes and perceived stigma influence intentions to seek child mental health services. Cultur Divers Ethnic Minor Psychol. 2015;21(4):613-618. PubMed 34. Shaw RJ, Wamboldt M, Bursch B, Stuber M. Practice patterns in pediatric consultation-liaison psychiatry: a national survey. Psychosomatics. 2006;47(1):43-49. PubMed 35. Benton TD, Ifeagwu JA, Smith-Whitley K. Anxiety and depression in children and adolescents with sickle cell disease. Curr Psychiatry Rep. 2007;9(2):114-121. PubMed 36. Chavira DA, Garland AF, Daley S, Hough R. The impact of medical comorbidity on mental health and functional health outcomes among children with anxiety disorders. J Dev Behav Pediatr. 2008;29(5):394-402. PubMed 37. Knight A, Weiss P, Morales K, et al. Depression and anxiety and their association with healthcare utilization in pediatric lupus and mixed connective tissue disease patients: a cross-sectional study. Pediatr Rheumatol Online J. 2014;12:42. PubMed 38. Marsac ML, Kassam-Adams N, Delahanty DL, F. Widaman K, Barakat LP. Posttraumatic stress following acute medical trauma in children: a proposed model of bio-psycho-social processes during the peri-trauma period. Clin Child Fam Psychol Rev. 2014;17(4):399-411. PubMed 39. Marsac ML, Hildenbrand AK, Kohser KL, Winston FK, Li Y, Kassam-Adams N. Preventing posttraumatic stress following pediatric injury: a randomized controlled trial of a web-based psycho-educational intervention for parents. J Pediatr Psychol. 2013;38(10):1101-1111. PubMed 40. Petersen MC, Kube DA, Whitaker TM, Graff JC, Palmer FB. Prevalence of developmental and behavioral disorders in a pediatric hospital. Pediatrics. 2009;123(3):e490-e495. PubMed
Address for correspondence and reprint requests: Stephanie K. Doupnik, MD, Division of General Pediatrics, Children’s Hospital of Philadelphia, Room 1451, CHOP North, 3401 Civic Center Blvd, Philadelphia, PA 19104; Telephone: 215-590-1000; Fax, 267-426-0380; E-mail: doupniks@chop.edu
Content Gating
Open Access (article Unlocked/Open Access)
Alternative CME
Consolidated Pubs: Do Not Show Source Publication Logo
