Original Research

Patellar height may be important in determining function after total knee arthroplasty (TKA). By altering patellofemoral joint mechanics, patella baja may cause several functional issues after TKA.^1-8 Patella baja leads to decreased range of motion (ROM) affecting both extension and flexion.^5,8,9 Deep flexion can be restricted in TKA patients with patella baja because of tracking limitations associated with an inferiorly displaced patella. As the knee is brought into flexion, the patella can impinge on the anterior aspect of the tibial polyethylene or the tibial tray—presenting a true block to flexion and potentially altering wear.^1,10

Another functional issue with patella baja is loss of strength in the extensor mechanism. The patella serves as a fulcrum for the extensor muscles of the knee. When positioned properly and functioning properly, the patella increases the extensor forces generated. When the patella is positioned in baja, the knee generates decreased extensor mechanism force.^6,7 This can result in a lag, with the patient being unable to fully extend the knee. Extension-dependent activities are impaired. Patients with weak extensor function can experience poor function with stair climbing, rising from a chair, and exiting an automobile. The improper function and scarring of the patella can result in increased anterior knee pain and worse functional outcome scores after TKAs.^3,9

An abnormally positioned patella can either result from or lead to increased scarring in the knee.^9,11 Patellar height is often measured with the Insall-Salvati ratio (ISR), which is the patella tendon length (measurement of the tendon from the tibial tubercle to the inferior pole of the patella) divided by the patellar length (longest measured dimension of the patella) (Figure 1).¹² Patella baja is defined as an ISR of less than 0.8. Other indices that reference off the tibial plateau (Blackburne-Peel ratio, Canton-Deschamps ratio) reflect an elevation of the joint line, or pseudobaja, and are unreliable for analysis of patella baja after TKA.¹³

Postoperative patella baja has been reported in 10% to 34% of primary TKAs.^4,7 Inferior positioning of the patella and scarring can cause intraoperative difficulty with exposure and may complicate outcomes.^9,13 The exposure scar is often larger in TKA revisions for infection compared with primary TKAs.

We conducted a study to compare the incidence of patella baja in noninfected and infected TKA revisions. We hypothesized that, compared with noninfected knees, infected knees treated with nonarticulating spacers would have a higher incidence of patella baja both before and after surgery secondary to more inflammation, immobilization, and related scarring.

Materials and Methods

We conducted a retrospective case–cohort study of 148 consecutive TKA revisions. All TKA revisions were performed between 2003 and 2009 using a mobile-bearing revision system from a single manufacturer. All surgeries were done at a single institution by the 2 senior surgeons. The surgical approach was a standard medial parapatellar approach without patellar eversion. Our institutional review board approved the study and waived the requirement for informed consent, as this was a retrospective study of existing medical records that posed no more than minimal risk to patients.

To properly evaluate patellar height, orthopedic specialty–trained radiologic technicians obtained preoperative and postoperative weight-bearing radiographs using a standardized lateral radiograph in clinic. Two blinded investigators measured ISR radiographically both before surgery (preexplant for septic revisions) and at latest follow-up (postreplant for septic revisions). Patients with inadequate films and/or patellectomies were excluded, along with patients who had less than 6 months of postoperative follow-up.

Ninety-one patients (101 TKAs) met the study inclusion criteria. Two groups of cases were compared: aseptic revisions (n = 67) and septic revisions (n = 34). Reasons for aseptic revisions included implant loosening (24/67, 35.8%), instability (12/67, 17.9%), pain (12/67, 17.9%), lysis (5/67, 7.5%), stiffness (3/67, 4.5%), and malrotation (2/67, 3.0%). Infection was determined by Musculoskeletal Infection Society criteria, as documented by positive aspirations and/or intraoperative tissue cultures taken at prosthesis explantation, elevated white blood cell count in the aspirate, elevated percentage of polymorphonuclear (PMN) cells in the aspirate, gross purulence, presence of chronic draining sinus, or histologic analysis revealing acute inflammation with more than 5 PMN cells per high power field.^14,15

All infected TKAs were treated with 2-stage revisions. The standard of care at our institution through this series was to use a nonarticulating spacer for the treatment of infection. Weight-bearing status varied by extent of bone damage. Six weeks of culture-specific intravenous antibiotics were administered with assistance from an infectious disease consultant. Reimplantation was performed when clinical and laboratory criteria for resolution of infection were met—specifically, when erythrocyte sedimentation rate was less than 30 mm/h, C-reactive protein level was less than 10 mg/L, and aspirates were culture-negative. Mean (range) follow-up was 33.9 (6.2-75.7) months for aseptic revisions and 32.3 (7.5-94.2) months for septic revisions. Radiographic follow-up was performed at each visit, with weight-bearing anteroposterior and posteroanterior views, along with a lateral knee radiograph. At final follow-up, ROM was recorded by the senior attending evaluating the patient.

Categorical variables were statistically analyzed with χ² tests, and continuous variables were analyzed with Student t test, analysis of variance, and univariate analysis of covariance (ANCOVA). Statistical significance was set at P < .05. Intrarater reliability was measured with the intraclass correlation coefficient (ICC). All statistical analysis was performed with Predictive Analytics SoftWare Statistics Version 20.0 (SPSS, Chicago, Illinois).

Results

Ninety-one consecutive patients (43 men, 48 women) were included in this study. Mean (SD) age was 66.4 (10.1) years. Mean (SD) preoperative ISR in septic and aseptic cases was 0.94 (0.25) for men and 1.02 (0.23) for women (P = .10). Mean postoperative ISR in septic and aseptic cases was 0.84 (0.27) for men and 0.99 (0.23) for women (P = .004). There was a sex difference between septic and aseptic revisions. There were 22 men and 36 women in the aseptic group and 21 men and 12 women in the septic group (P = .01). Men were more likely than women to have septic revisions and patella baja. Table 1 compares the patient demographics of the 2 patient populations. Mean (SD) number of surgeries, including irrigation and débridement procedures before reimplantation, was larger for septic revisions, 2.9 (0.9), than for aseptic revisions, 1.4 (0.8) (P < .001).

Infection was the most common reason for revision and accounted for 33.7% (34/101) of all revisions. Noninfectious indications, in declining order of frequency, included loosening (23.8%, 24/101), instability (11.9%, 12/101), pain (11.9%, 12/101), osteolysis (5.0%, 5/101), polyethylene wear (5.0%, 5/101), failed unicompartmental knee (4.0%, 4/101), stiffness (3.0%, 3/101), and patellar problems (2.0%, 2/101) (Table 2). ISR decreased significantly only in infected revisions. It is important to note that there was not a high incidence of stiffness or patellofemoral failure in revision patients before surgery.

Mean (SD) ISR did not differ between groups before surgery, 1.00 (0.25) for aseptic and 0.96 (0.22) for septic (P = .49), but differed significantly after surgery, 0.99 (0.23) for aseptic and 0.77 (0.24) for septic (P < .001) (Figure 2). The univariate ANCOVA also demonstrated a postoperative difference between groups when taking the preoperative ratio into account: 0.99 (0.23) for aseptic and 0.78 (0.24) for septic (P = .005) (Table 3). Before surgery, 22.4% and 23.9% of the aseptic and septic groups, respectively, had patella baja (P = .58). After surgery, 17.6% and 58.8% of the aseptic and septic groups had patella baja (P = .001) (Table 4). The ICC for preoperative ISR was 0.94, and the ICC for postoperative ISR was 0.96, which indicates excellent agreement of measurements between the 2 blinded investigators.

ROM differed between septic and aseptic groups owing to the difference in postoperative flexion. Mean (SD) postoperative extension was 2.2° (5.4°) for the aseptic group and 5.1° (9.8°) for the septic group—not significantly different (P = .13). Mean (SD) postoperative flexion was 110.2° (18.8°) for the aseptic group and 97.2° (29.4°) for the septic group—significantly different (P = .02). The groups differed significantly (P = .02) in mean (SD) ROM: 108.0° (20.7°) for aseptic and 92.2° (34.6°) for septic (Table 1). ROM was also significantly associated with patella baja (P = .04), as patients with ISR of less than 0.8 had mean (SD) postoperative ROM of 95.1° (31.6°), and patients without patella baja had mean (SD) postoperative ROM of 106.8° (23.6°).

For the septic group, mean (SD) time between first and second stages was 13.0 (8.3) weeks (range, 1-44.3 weeks). Mean (SD) timing of spacer placement was not statistically significantly different (P = .90) between patients who had patella baja, 12.9 (8.8) weeks, and patients who did not have patella baja, 13.2 (7.8) weeks.

Discussion    

This study demonstrated that TKAs done for septic reasons resulted in a higher incidence of patella baja and decreased ROM. Incidence of patella baja was higher both before and after revision in septic TKAs than in aseptic TKAs, proving the hypothesis under study. Prerevision incidence was not significantly different, but there was a trend that could not be ignored. This may suggest that there is already an ongoing process in the infected knee that contributes to patella baja; the precise etiology remains unclear and is likely multifactorial. For example, scar formation may be increased in patients with chronic infection, predisposing to patella baja. This assertion is indirectly supported by a recent study from our institution revealing longer average surgical time in septic versus aseptic knee revisions; the difference was thought to reflect increased scar-tissue formation.¹⁶ That study also found that patients who underwent septic revisions had significantly more surgical procedures than patients who underwent aseptic revisions. Repetitive surgeries—specifically, repetitive arthrotomies during irrigation and débridement before reimplantation—lead to increased scar formation, which may contribute to preoperative and postoperative patella baja. This may be reflected in the findings that ROM was decreased in patients in the septic group versus patients in the aseptic group and that ROM was decreased in patients with patella baja. In addition, our study found that male patients were more likely to undergo TKA revision for septic reasons and to develop postoperative patella baja. This finding contrasts with that of a study⁵ that compared preoperative and postoperative ISR in primary TKA and found that women were more likely than men to have patella baja. Although women are more likely to undergo TKA revision,¹⁷ men may be more susceptible to infection and subsequent patella baja.

The higher postoperative rate of patella baja in the septic group became statistically significant even when preoperative incidence was considered. This may have been caused by infection-related scarring and by prolonged immobilization of septic knees with use of nonarticulating antibiotic spacers. By keeping these knees immobile with a nonarticulating spacer for a prolonged period in the healing phase of the infection, scar tissue may mature and form over the time between stages. A comparable example may be high tibial osteotomies, in which a high incidence of patella baja has been partly attributed to prolonged casting.¹¹ Future work comparing the results of articulating and nonarticulating spacers will help to determine if immobilization contributes to patella baja in infected TKAs.

There are several limitations to our study. Patient outcome questionnaires were not used, and they would have allowed for the assessment of physical outcomes and emotional satisfaction by comparing outcomes between patients with and without patella baja and comparing septic and aseptic TKAs. In addition, there was no standard method for quantifying difficulty of revision, which would have enabled us to compare difficulty of revision in patients with patella baja.

Conclusion

This study identified a high rate of patella baja and decreased ROM in TKA revisions, particularly infected revisions treated with a nonarticulating spacer. It is important to determine if there are functional consequences. Further investigation is needed regarding the cause, prevention, and management of this potentially debilitating outcome after revision TKA.

Patellar height may be important in determining function after total knee arthroplasty (TKA). By altering patellofemoral joint mechanics, patella baja may cause several functional issues after TKA.^1-8 Patella baja leads to decreased range of motion (ROM) affecting both extension and flexion.^5,8,9 Deep flexion can be restricted in TKA patients with patella baja because of tracking limitations associated with an inferiorly displaced patella. As the knee is brought into flexion, the patella can impinge on the anterior aspect of the tibial polyethylene or the tibial tray—presenting a true block to flexion and potentially altering wear.^1,10

Another functional issue with patella baja is loss of strength in the extensor mechanism. The patella serves as a fulcrum for the extensor muscles of the knee. When positioned properly and functioning properly, the patella increases the extensor forces generated. When the patella is positioned in baja, the knee generates decreased extensor mechanism force.^6,7 This can result in a lag, with the patient being unable to fully extend the knee. Extension-dependent activities are impaired. Patients with weak extensor function can experience poor function with stair climbing, rising from a chair, and exiting an automobile. The improper function and scarring of the patella can result in increased anterior knee pain and worse functional outcome scores after TKAs.^3,9

An abnormally positioned patella can either result from or lead to increased scarring in the knee.^9,11 Patellar height is often measured with the Insall-Salvati ratio (ISR), which is the patella tendon length (measurement of the tendon from the tibial tubercle to the inferior pole of the patella) divided by the patellar length (longest measured dimension of the patella) (Figure 1).¹² Patella baja is defined as an ISR of less than 0.8. Other indices that reference off the tibial plateau (Blackburne-Peel ratio, Canton-Deschamps ratio) reflect an elevation of the joint line, or pseudobaja, and are unreliable for analysis of patella baja after TKA.¹³

Postoperative patella baja has been reported in 10% to 34% of primary TKAs.^4,7 Inferior positioning of the patella and scarring can cause intraoperative difficulty with exposure and may complicate outcomes.^9,13 The exposure scar is often larger in TKA revisions for infection compared with primary TKAs.

We conducted a study to compare the incidence of patella baja in noninfected and infected TKA revisions. We hypothesized that, compared with noninfected knees, infected knees treated with nonarticulating spacers would have a higher incidence of patella baja both before and after surgery secondary to more inflammation, immobilization, and related scarring.

Materials and Methods

We conducted a retrospective case–cohort study of 148 consecutive TKA revisions. All TKA revisions were performed between 2003 and 2009 using a mobile-bearing revision system from a single manufacturer. All surgeries were done at a single institution by the 2 senior surgeons. The surgical approach was a standard medial parapatellar approach without patellar eversion. Our institutional review board approved the study and waived the requirement for informed consent, as this was a retrospective study of existing medical records that posed no more than minimal risk to patients.

To properly evaluate patellar height, orthopedic specialty–trained radiologic technicians obtained preoperative and postoperative weight-bearing radiographs using a standardized lateral radiograph in clinic. Two blinded investigators measured ISR radiographically both before surgery (preexplant for septic revisions) and at latest follow-up (postreplant for septic revisions). Patients with inadequate films and/or patellectomies were excluded, along with patients who had less than 6 months of postoperative follow-up.

Ninety-one patients (101 TKAs) met the study inclusion criteria. Two groups of cases were compared: aseptic revisions (n = 67) and septic revisions (n = 34). Reasons for aseptic revisions included implant loosening (24/67, 35.8%), instability (12/67, 17.9%), pain (12/67, 17.9%), lysis (5/67, 7.5%), stiffness (3/67, 4.5%), and malrotation (2/67, 3.0%). Infection was determined by Musculoskeletal Infection Society criteria, as documented by positive aspirations and/or intraoperative tissue cultures taken at prosthesis explantation, elevated white blood cell count in the aspirate, elevated percentage of polymorphonuclear (PMN) cells in the aspirate, gross purulence, presence of chronic draining sinus, or histologic analysis revealing acute inflammation with more than 5 PMN cells per high power field.^14,15

All infected TKAs were treated with 2-stage revisions. The standard of care at our institution through this series was to use a nonarticulating spacer for the treatment of infection. Weight-bearing status varied by extent of bone damage. Six weeks of culture-specific intravenous antibiotics were administered with assistance from an infectious disease consultant. Reimplantation was performed when clinical and laboratory criteria for resolution of infection were met—specifically, when erythrocyte sedimentation rate was less than 30 mm/h, C-reactive protein level was less than 10 mg/L, and aspirates were culture-negative. Mean (range) follow-up was 33.9 (6.2-75.7) months for aseptic revisions and 32.3 (7.5-94.2) months for septic revisions. Radiographic follow-up was performed at each visit, with weight-bearing anteroposterior and posteroanterior views, along with a lateral knee radiograph. At final follow-up, ROM was recorded by the senior attending evaluating the patient.

Categorical variables were statistically analyzed with χ² tests, and continuous variables were analyzed with Student t test, analysis of variance, and univariate analysis of covariance (ANCOVA). Statistical significance was set at P < .05. Intrarater reliability was measured with the intraclass correlation coefficient (ICC). All statistical analysis was performed with Predictive Analytics SoftWare Statistics Version 20.0 (SPSS, Chicago, Illinois).

Results

Ninety-one consecutive patients (43 men, 48 women) were included in this study. Mean (SD) age was 66.4 (10.1) years. Mean (SD) preoperative ISR in septic and aseptic cases was 0.94 (0.25) for men and 1.02 (0.23) for women (P = .10). Mean postoperative ISR in septic and aseptic cases was 0.84 (0.27) for men and 0.99 (0.23) for women (P = .004). There was a sex difference between septic and aseptic revisions. There were 22 men and 36 women in the aseptic group and 21 men and 12 women in the septic group (P = .01). Men were more likely than women to have septic revisions and patella baja. Table 1 compares the patient demographics of the 2 patient populations. Mean (SD) number of surgeries, including irrigation and débridement procedures before reimplantation, was larger for septic revisions, 2.9 (0.9), than for aseptic revisions, 1.4 (0.8) (P < .001).

Infection was the most common reason for revision and accounted for 33.7% (34/101) of all revisions. Noninfectious indications, in declining order of frequency, included loosening (23.8%, 24/101), instability (11.9%, 12/101), pain (11.9%, 12/101), osteolysis (5.0%, 5/101), polyethylene wear (5.0%, 5/101), failed unicompartmental knee (4.0%, 4/101), stiffness (3.0%, 3/101), and patellar problems (2.0%, 2/101) (Table 2). ISR decreased significantly only in infected revisions. It is important to note that there was not a high incidence of stiffness or patellofemoral failure in revision patients before surgery.

Mean (SD) ISR did not differ between groups before surgery, 1.00 (0.25) for aseptic and 0.96 (0.22) for septic (P = .49), but differed significantly after surgery, 0.99 (0.23) for aseptic and 0.77 (0.24) for septic (P < .001) (Figure 2). The univariate ANCOVA also demonstrated a postoperative difference between groups when taking the preoperative ratio into account: 0.99 (0.23) for aseptic and 0.78 (0.24) for septic (P = .005) (Table 3). Before surgery, 22.4% and 23.9% of the aseptic and septic groups, respectively, had patella baja (P = .58). After surgery, 17.6% and 58.8% of the aseptic and septic groups had patella baja (P = .001) (Table 4). The ICC for preoperative ISR was 0.94, and the ICC for postoperative ISR was 0.96, which indicates excellent agreement of measurements between the 2 blinded investigators.

ROM differed between septic and aseptic groups owing to the difference in postoperative flexion. Mean (SD) postoperative extension was 2.2° (5.4°) for the aseptic group and 5.1° (9.8°) for the septic group—not significantly different (P = .13). Mean (SD) postoperative flexion was 110.2° (18.8°) for the aseptic group and 97.2° (29.4°) for the septic group—significantly different (P = .02). The groups differed significantly (P = .02) in mean (SD) ROM: 108.0° (20.7°) for aseptic and 92.2° (34.6°) for septic (Table 1). ROM was also significantly associated with patella baja (P = .04), as patients with ISR of less than 0.8 had mean (SD) postoperative ROM of 95.1° (31.6°), and patients without patella baja had mean (SD) postoperative ROM of 106.8° (23.6°).

For the septic group, mean (SD) time between first and second stages was 13.0 (8.3) weeks (range, 1-44.3 weeks). Mean (SD) timing of spacer placement was not statistically significantly different (P = .90) between patients who had patella baja, 12.9 (8.8) weeks, and patients who did not have patella baja, 13.2 (7.8) weeks.

Discussion    

This study demonstrated that TKAs done for septic reasons resulted in a higher incidence of patella baja and decreased ROM. Incidence of patella baja was higher both before and after revision in septic TKAs than in aseptic TKAs, proving the hypothesis under study. Prerevision incidence was not significantly different, but there was a trend that could not be ignored. This may suggest that there is already an ongoing process in the infected knee that contributes to patella baja; the precise etiology remains unclear and is likely multifactorial. For example, scar formation may be increased in patients with chronic infection, predisposing to patella baja. This assertion is indirectly supported by a recent study from our institution revealing longer average surgical time in septic versus aseptic knee revisions; the difference was thought to reflect increased scar-tissue formation.¹⁶ That study also found that patients who underwent septic revisions had significantly more surgical procedures than patients who underwent aseptic revisions. Repetitive surgeries—specifically, repetitive arthrotomies during irrigation and débridement before reimplantation—lead to increased scar formation, which may contribute to preoperative and postoperative patella baja. This may be reflected in the findings that ROM was decreased in patients in the septic group versus patients in the aseptic group and that ROM was decreased in patients with patella baja. In addition, our study found that male patients were more likely to undergo TKA revision for septic reasons and to develop postoperative patella baja. This finding contrasts with that of a study⁵ that compared preoperative and postoperative ISR in primary TKA and found that women were more likely than men to have patella baja. Although women are more likely to undergo TKA revision,¹⁷ men may be more susceptible to infection and subsequent patella baja.

The higher postoperative rate of patella baja in the septic group became statistically significant even when preoperative incidence was considered. This may have been caused by infection-related scarring and by prolonged immobilization of septic knees with use of nonarticulating antibiotic spacers. By keeping these knees immobile with a nonarticulating spacer for a prolonged period in the healing phase of the infection, scar tissue may mature and form over the time between stages. A comparable example may be high tibial osteotomies, in which a high incidence of patella baja has been partly attributed to prolonged casting.¹¹ Future work comparing the results of articulating and nonarticulating spacers will help to determine if immobilization contributes to patella baja in infected TKAs.

There are several limitations to our study. Patient outcome questionnaires were not used, and they would have allowed for the assessment of physical outcomes and emotional satisfaction by comparing outcomes between patients with and without patella baja and comparing septic and aseptic TKAs. In addition, there was no standard method for quantifying difficulty of revision, which would have enabled us to compare difficulty of revision in patients with patella baja.

Conclusion

This study identified a high rate of patella baja and decreased ROM in TKA revisions, particularly infected revisions treated with a nonarticulating spacer. It is important to determine if there are functional consequences. Further investigation is needed regarding the cause, prevention, and management of this potentially debilitating outcome after revision TKA.

Patellar height may be important in determining function after total knee arthroplasty (TKA). By altering patellofemoral joint mechanics, patella baja may cause several functional issues after TKA.^1-8 Patella baja leads to decreased range of motion (ROM) affecting both extension and flexion.^5,8,9 Deep flexion can be restricted in TKA patients with patella baja because of tracking limitations associated with an inferiorly displaced patella. As the knee is brought into flexion, the patella can impinge on the anterior aspect of the tibial polyethylene or the tibial tray—presenting a true block to flexion and potentially altering wear.^1,10

Another functional issue with patella baja is loss of strength in the extensor mechanism. The patella serves as a fulcrum for the extensor muscles of the knee. When positioned properly and functioning properly, the patella increases the extensor forces generated. When the patella is positioned in baja, the knee generates decreased extensor mechanism force.^6,7 This can result in a lag, with the patient being unable to fully extend the knee. Extension-dependent activities are impaired. Patients with weak extensor function can experience poor function with stair climbing, rising from a chair, and exiting an automobile. The improper function and scarring of the patella can result in increased anterior knee pain and worse functional outcome scores after TKAs.^3,9

An abnormally positioned patella can either result from or lead to increased scarring in the knee.^9,11 Patellar height is often measured with the Insall-Salvati ratio (ISR), which is the patella tendon length (measurement of the tendon from the tibial tubercle to the inferior pole of the patella) divided by the patellar length (longest measured dimension of the patella) (Figure 1).¹² Patella baja is defined as an ISR of less than 0.8. Other indices that reference off the tibial plateau (Blackburne-Peel ratio, Canton-Deschamps ratio) reflect an elevation of the joint line, or pseudobaja, and are unreliable for analysis of patella baja after TKA.¹³

Postoperative patella baja has been reported in 10% to 34% of primary TKAs.^4,7 Inferior positioning of the patella and scarring can cause intraoperative difficulty with exposure and may complicate outcomes.^9,13 The exposure scar is often larger in TKA revisions for infection compared with primary TKAs.

We conducted a study to compare the incidence of patella baja in noninfected and infected TKA revisions. We hypothesized that, compared with noninfected knees, infected knees treated with nonarticulating spacers would have a higher incidence of patella baja both before and after surgery secondary to more inflammation, immobilization, and related scarring.

Materials and Methods

We conducted a retrospective case–cohort study of 148 consecutive TKA revisions. All TKA revisions were performed between 2003 and 2009 using a mobile-bearing revision system from a single manufacturer. All surgeries were done at a single institution by the 2 senior surgeons. The surgical approach was a standard medial parapatellar approach without patellar eversion. Our institutional review board approved the study and waived the requirement for informed consent, as this was a retrospective study of existing medical records that posed no more than minimal risk to patients.

To properly evaluate patellar height, orthopedic specialty–trained radiologic technicians obtained preoperative and postoperative weight-bearing radiographs using a standardized lateral radiograph in clinic. Two blinded investigators measured ISR radiographically both before surgery (preexplant for septic revisions) and at latest follow-up (postreplant for septic revisions). Patients with inadequate films and/or patellectomies were excluded, along with patients who had less than 6 months of postoperative follow-up.

Ninety-one patients (101 TKAs) met the study inclusion criteria. Two groups of cases were compared: aseptic revisions (n = 67) and septic revisions (n = 34). Reasons for aseptic revisions included implant loosening (24/67, 35.8%), instability (12/67, 17.9%), pain (12/67, 17.9%), lysis (5/67, 7.5%), stiffness (3/67, 4.5%), and malrotation (2/67, 3.0%). Infection was determined by Musculoskeletal Infection Society criteria, as documented by positive aspirations and/or intraoperative tissue cultures taken at prosthesis explantation, elevated white blood cell count in the aspirate, elevated percentage of polymorphonuclear (PMN) cells in the aspirate, gross purulence, presence of chronic draining sinus, or histologic analysis revealing acute inflammation with more than 5 PMN cells per high power field.^14,15

All infected TKAs were treated with 2-stage revisions. The standard of care at our institution through this series was to use a nonarticulating spacer for the treatment of infection. Weight-bearing status varied by extent of bone damage. Six weeks of culture-specific intravenous antibiotics were administered with assistance from an infectious disease consultant. Reimplantation was performed when clinical and laboratory criteria for resolution of infection were met—specifically, when erythrocyte sedimentation rate was less than 30 mm/h, C-reactive protein level was less than 10 mg/L, and aspirates were culture-negative. Mean (range) follow-up was 33.9 (6.2-75.7) months for aseptic revisions and 32.3 (7.5-94.2) months for septic revisions. Radiographic follow-up was performed at each visit, with weight-bearing anteroposterior and posteroanterior views, along with a lateral knee radiograph. At final follow-up, ROM was recorded by the senior attending evaluating the patient.

Categorical variables were statistically analyzed with χ² tests, and continuous variables were analyzed with Student t test, analysis of variance, and univariate analysis of covariance (ANCOVA). Statistical significance was set at P < .05. Intrarater reliability was measured with the intraclass correlation coefficient (ICC). All statistical analysis was performed with Predictive Analytics SoftWare Statistics Version 20.0 (SPSS, Chicago, Illinois).

Results

Ninety-one consecutive patients (43 men, 48 women) were included in this study. Mean (SD) age was 66.4 (10.1) years. Mean (SD) preoperative ISR in septic and aseptic cases was 0.94 (0.25) for men and 1.02 (0.23) for women (P = .10). Mean postoperative ISR in septic and aseptic cases was 0.84 (0.27) for men and 0.99 (0.23) for women (P = .004). There was a sex difference between septic and aseptic revisions. There were 22 men and 36 women in the aseptic group and 21 men and 12 women in the septic group (P = .01). Men were more likely than women to have septic revisions and patella baja. Table 1 compares the patient demographics of the 2 patient populations. Mean (SD) number of surgeries, including irrigation and débridement procedures before reimplantation, was larger for septic revisions, 2.9 (0.9), than for aseptic revisions, 1.4 (0.8) (P < .001).

Infection was the most common reason for revision and accounted for 33.7% (34/101) of all revisions. Noninfectious indications, in declining order of frequency, included loosening (23.8%, 24/101), instability (11.9%, 12/101), pain (11.9%, 12/101), osteolysis (5.0%, 5/101), polyethylene wear (5.0%, 5/101), failed unicompartmental knee (4.0%, 4/101), stiffness (3.0%, 3/101), and patellar problems (2.0%, 2/101) (Table 2). ISR decreased significantly only in infected revisions. It is important to note that there was not a high incidence of stiffness or patellofemoral failure in revision patients before surgery.

Mean (SD) ISR did not differ between groups before surgery, 1.00 (0.25) for aseptic and 0.96 (0.22) for septic (P = .49), but differed significantly after surgery, 0.99 (0.23) for aseptic and 0.77 (0.24) for septic (P < .001) (Figure 2). The univariate ANCOVA also demonstrated a postoperative difference between groups when taking the preoperative ratio into account: 0.99 (0.23) for aseptic and 0.78 (0.24) for septic (P = .005) (Table 3). Before surgery, 22.4% and 23.9% of the aseptic and septic groups, respectively, had patella baja (P = .58). After surgery, 17.6% and 58.8% of the aseptic and septic groups had patella baja (P = .001) (Table 4). The ICC for preoperative ISR was 0.94, and the ICC for postoperative ISR was 0.96, which indicates excellent agreement of measurements between the 2 blinded investigators.

ROM differed between septic and aseptic groups owing to the difference in postoperative flexion. Mean (SD) postoperative extension was 2.2° (5.4°) for the aseptic group and 5.1° (9.8°) for the septic group—not significantly different (P = .13). Mean (SD) postoperative flexion was 110.2° (18.8°) for the aseptic group and 97.2° (29.4°) for the septic group—significantly different (P = .02). The groups differed significantly (P = .02) in mean (SD) ROM: 108.0° (20.7°) for aseptic and 92.2° (34.6°) for septic (Table 1). ROM was also significantly associated with patella baja (P = .04), as patients with ISR of less than 0.8 had mean (SD) postoperative ROM of 95.1° (31.6°), and patients without patella baja had mean (SD) postoperative ROM of 106.8° (23.6°).

For the septic group, mean (SD) time between first and second stages was 13.0 (8.3) weeks (range, 1-44.3 weeks). Mean (SD) timing of spacer placement was not statistically significantly different (P = .90) between patients who had patella baja, 12.9 (8.8) weeks, and patients who did not have patella baja, 13.2 (7.8) weeks.

Discussion    

This study demonstrated that TKAs done for septic reasons resulted in a higher incidence of patella baja and decreased ROM. Incidence of patella baja was higher both before and after revision in septic TKAs than in aseptic TKAs, proving the hypothesis under study. Prerevision incidence was not significantly different, but there was a trend that could not be ignored. This may suggest that there is already an ongoing process in the infected knee that contributes to patella baja; the precise etiology remains unclear and is likely multifactorial. For example, scar formation may be increased in patients with chronic infection, predisposing to patella baja. This assertion is indirectly supported by a recent study from our institution revealing longer average surgical time in septic versus aseptic knee revisions; the difference was thought to reflect increased scar-tissue formation.¹⁶ That study also found that patients who underwent septic revisions had significantly more surgical procedures than patients who underwent aseptic revisions. Repetitive surgeries—specifically, repetitive arthrotomies during irrigation and débridement before reimplantation—lead to increased scar formation, which may contribute to preoperative and postoperative patella baja. This may be reflected in the findings that ROM was decreased in patients in the septic group versus patients in the aseptic group and that ROM was decreased in patients with patella baja. In addition, our study found that male patients were more likely to undergo TKA revision for septic reasons and to develop postoperative patella baja. This finding contrasts with that of a study⁵ that compared preoperative and postoperative ISR in primary TKA and found that women were more likely than men to have patella baja. Although women are more likely to undergo TKA revision,¹⁷ men may be more susceptible to infection and subsequent patella baja.

The higher postoperative rate of patella baja in the septic group became statistically significant even when preoperative incidence was considered. This may have been caused by infection-related scarring and by prolonged immobilization of septic knees with use of nonarticulating antibiotic spacers. By keeping these knees immobile with a nonarticulating spacer for a prolonged period in the healing phase of the infection, scar tissue may mature and form over the time between stages. A comparable example may be high tibial osteotomies, in which a high incidence of patella baja has been partly attributed to prolonged casting.¹¹ Future work comparing the results of articulating and nonarticulating spacers will help to determine if immobilization contributes to patella baja in infected TKAs.

There are several limitations to our study. Patient outcome questionnaires were not used, and they would have allowed for the assessment of physical outcomes and emotional satisfaction by comparing outcomes between patients with and without patella baja and comparing septic and aseptic TKAs. In addition, there was no standard method for quantifying difficulty of revision, which would have enabled us to compare difficulty of revision in patients with patella baja.

Conclusion

This study identified a high rate of patella baja and decreased ROM in TKA revisions, particularly infected revisions treated with a nonarticulating spacer. It is important to determine if there are functional consequences. Further investigation is needed regarding the cause, prevention, and management of this potentially debilitating outcome after revision TKA.

Heart failure (HF) continues to grow as a significant health problem in the U.S., accounting for 1.1 million hospitalizations annually.¹ About 5.8 million Americans have HF, and 670,000 new cases are diagnosed each year.¹ The prevalence of HF increases with age. Persons aged > 65 years comprise the largest group of patients hospitalized for the condition. Heart failure-related hospitalizations place a major financial burden on patients, caregivers, and the national health care system. In 2010, the estimated cost of health care for HF was $35 billion with hospitalizations accounting for 1% to 2% of the total annual health care costs.^1-3 Furthermore, > 50% of patients with HF are rehospitalized before their first outpatient follow-up.⁴

To ensure patients are ready for discharge, HF guidelines recommend specific interventions for all hospitalized patients. These recommendations include successful transition from IV to oral diuretic therapy as well as the initiation of a ß-blocker and an angiotensin-converting enzyme inhibitor (ACE-I) or an angiotensin receptor blocker (ARB) in stable patients with a left ventricular ejection fraction (LVEF) < 40% and without contraindications. Additionally, patients and their caregivers should receive comprehensive discharge instructions regarding medications, the importance of adherence and regular follow-up, sodium and fluid restriction, weight monitoring, physical activity, and a plan for worsening symptoms. When available, assistance with the hospital-to-home transition should also be provided.^5,6

Heart Failure Measures

Recognizing common factors essential to HF care, the Joint Commission has implemented HF core measures that all U.S. hospitals are required to meet to maintain accreditation status. These guideline-supported measures include receipt of diet, weight, and medication instructions; measured or scheduled assessment of LVEF; ACE-I or an ARB prescribed in patients with LVEF < 40%; and smoking cessation counseling before discharge for all patients with HF.

In addition to HF core measures, 30-day HF readmission rates have also become available to the general public as another hospital quality indicator. In 2009, the Centers for Medicare & Medicaid Services began publicly reporting 30-day HF readmission rates for Medicare patients. A CMS report indicated a 24.8% national 30-day HF readmission rate from July 1, 2007, through June 30, 2010.⁷ Unfortunately, even with the increased quality improvement effort, national HF rehospitalization rates have remained relatively steady in recent years.³

The VA health care system has a growing number of veterans with HF, and it is the leading discharge diagnosis in patients treated at VA hospitals. The number of HF-related hospitalizations at the VA health care system increased from just over 74,000 in fiscal year 2002 to 96,000 in 2009.⁸

To advance the care of veterans with HF and implement best practices, the VA launched the Chronic HF-Quality Enhancement Research Initiative (CHF-QUERI). The major goals of this initiative are to reduce hospitalization rates, increase use of life-prolonging care, empower patients and their caregivers in self-management, and improve appropriateness of HF therapies and tests. As part of its efforts, CHF-QUERI launched the HF Provider Network (HF Network), involving more than 712 VA health care providers (as of July 2014 there were more than 900 providers) committed to improving HF management throughout the entire VA health care system. The HF Network has already put into practice several quality improvement initiatives.

The National Hospital to Home initiative led by the American College of Cardiology and the Institute for Healthcare Improvement was launched throughout the VA system in January 2010.⁹ The main goal of this initiative is to reduce all-cause hospital readmission rates in patients with a discharge diagnosis of HF by improving medication management, early follow-up after discharge, and symptom management.

The Jesse Brown VAMC (JBVAMC) is an active participant of the Hospital to Home initiative, embracing the goals of reducing HF readmission rates and improving the transition of veterans from inpatient to outpatient care. The JBVAMC also has been successfully meeting or exceeding HF core measures except for providing discharge instructions. In May 2011, 91% of patients received discharge instructions, falling just slightly below the 93% target goal. Despite the implementation of HF care improvement initiatives and successful core measure performance, from July 1, 2007, to June 30, 2010, the average HF 30-day readmission rate at JBVAMC was reported to be 28.4%, compared with the national average of 24.8%. Additionally, the average readmission rate for fiscal year 2011 was 31% at JBVAMC, showing a further increase in readmission rates.

The cost of a hospital bed at JBVAMC ranges from about $2,000 to $5,000 per day. According to the American Heart Association’s Get With the Guidelines-HF registry, the mean hospital length of stay for HF in 2009 was 5.5 days.¹ Consequently, HF hospitalizations could potentially cost JBVAMC nearly $7 million annually. Therefore, HF readmissions not only affect patients and caregivers, but also represent a financial burden for JBVAMC. 

METHODS

The purpose of this study was to identify factors contributing to the high HF readmission rates in veterans enrolled at JBVAMC. This study was an Institutional Review Board and VA Research and Development Committee-approved retrospective, electronic chart review of patients with an ICD-9 principal discharge diagnosis code for HF and hospitalization for HF exacerbation anytime between October 1, 2010, and March 1, 2011. A patient chart was reviewed for 6 months after inclusion. A report was generated to identify patients discharged from JBVAMC with a principal discharge diagnosis of HF between October 1, 2010, and March 1, 2011, using the following ICD-9 HF codes: 402.01, 402.11, 402.91, 404.01, 404.03, 404.11, 404.13, 404.91, 404.93, 428.0, 428.1, 428.20, 428.21, 428.22, 428.23, 428.30, 428.31, 428.32, 428.33, 428.40, 428.41, 428.42, 428.43, and 428.9.

Patients were included if aged  ≥ 18 years with one of the ICD-9 HF codes as the principal discharge diagnosis within the study period. Patients were excluded from the study if transferred to or from an outside hospital, discharged without an ICD-9 principal diagnosis code for HF, electively admitted for HF, not treated for HF during hospitalization, left the hospital against medical advice, had chart documentation with comfort measures only, were discharged/transferred to hospice, had active HF medications listed under non-VA medications in the electronic medication profile, or did not receive follow-up at JBVAMC. Study participants were included in the study once, which was classified as their index HF hospitalization, and were followed for 6 months thereafter.

The primary endpoint was the difference in patient characteristics between 2 groups of patients: those readmitted for HF within 30 days of the index hospitalization and those readmitted after 30 days or not at all.

The study had multiple secondary endpoints. One was the difference in patient characteristics between 2 groups of patients: those readmitted for HF within 90 days of the index hospitalization and those readmitted after 90 days or not at all. Another secondary endpoint was the difference in patient characteristics between 2 groups of patients: those with ≥ 2 readmissions for HF within 6 months and those with < 2 HF readmissions within 6 months. Additional secondary endpoints included percentage of patients readmitted for HF within 30 days of the index HF hospitalization, time to readmission if applicable, time to death if applicable, and average number of readmissions per patient within 6 months.

Index data collected included age, gender, ethnicity, prior diagnosis of HF, date of diagnosis, hospitalization for HF within 30 days of the index HF admission, in-hospital cardiac arrest, comorbid conditions, systolic blood pressure (BP), heart rate, respiratory rate, weight, serum sodium, blood urea nitrogen, serum creatinine, hematocrit, and glucose. For this study, comorbid conditions gathered were diabetes mellitus, coronary artery disease, prior percutaneous coronary intervention, aortic stenosis, stroke, chronic obstructive pulmonary disease, and dementia.

Medication profiles were reviewed at the time of admission to determine whether the patient was prescribed an ACE-I/ARB, ß-blocker, diuretic, hydralazine and isosorbide dinitrate, aldosterone antagonist, digoxin, NSAIDs, nonvasoselective calcium channel blocker, and an antiarrhytmic other than amiodarone and dofetilide. Hospitalization data included the most recent LVEF, the number of days on oral diuretic therapy after stopping IV diuretics, the number of days admitted, and documentation of an in-person inpatient dietitian consultation.

Data collected at discharge included diet/weight/medication instructions, weight, BP,  American College of Cardiology/American Heart Association HF stage and New York Heart Association (NYHA) HF functional class, if documented. Discharge medication profiles were assessed for the number of medications (< 9 or ≥ 9), documentation of active prescriptions for an ACE-I/ARB and a ß-blocker (or contraindication documented), diuretic, hydralazine and isosorbide dinitrate, aldosterone antagonist, and digoxin. Other data collected were documentation of a scheduled follow-up appointment with primary care physician, urgent care, chronic HF (CHF) clinic, or cardiologist, and whether the patient was discharged to home, skilled nursing facility, shelter, or homeless. Additionally, if the patient was discharged on a diuretic, the dose was compared with the baseline diuretic. If the diuretic at discharge was different from the home diuretic, equivalent doses were used for comparison with that of the baseline diuretic.

Postdischarge data collection included telephone follow-up within 48 hours of discharge, medication compliance since the initial hospitalization, date of first outpatient follow-up after initial hospital discharge, enrollment in CHF clinic/CHF-PharmD/Care Coordination Home Telehealth (CCHT) program, outpatient dietitian consultations, and date of death if applicable. Medication adherence was defined as ≥ 80% of lowest percentage filled medication of all HF medications, determined by the refill history in the computerized patient record system (CPRS). First outpatient follow-up was defined as a visit in which HF was addressed in the assessment and plan.

If readmitted within the study period, data collection included the date of first nonelective hospital readmission for HF, BP, heart rate, weight, serum digoxin level, serum creatinine, serum potassium, and whether the patient was on a target dose of HF recommended medications (if LVEF < 40% and no contraindication). Heart failure recommended medications for which target doses are established include ACE-I/ARB and ß-blockers. For this study, target doses of ACE-Is were captopril 50 mg 3 times daily, enalapril 10 mg twice daily, fosinopril 40 mg daily, lisinopril 20 mg daily, ramipril 10 mg daily, and trandolapril 4 mg daily. Target doses for ARBs were candesartan 32 mg daily, losartan 50 mg daily, and valsartan 160 mg twice daily. ß-blocker target doses were bisoprolol 10 mg daily, carvedilol 25 mg twice daily (50 mg twice daily if patients’ weight was > 85 kg), and metoprolol succinate 200 mg daily.^5,6 A statistical analysis was not performed on the data.

RESULTS

A total of 137 patient charts were reviewed, and 109 patients were included in the study. Patients were excluded if they transferred to or from an outside hospital (n = 8), had no follow-up at JBVAMC (n = 8), left the hospital against medical advice (n = 4), were electively admitted (n = 4), were not treated for HF (n = 3), or only had comfort measures documented in the chart (n = 1). The patients included were predominantly male (99%) and African American (78%) and had a mean age of 70 years. The majority of the patients had a prior diagnosis of HF (87%) and a history of systolic HF (58%). Most patients were previously prescribed an ACE-I/ARB (83%) and a ß-blocker (76%) at the time of admission (Table 1).

Six patients were readmitted within 30 days of the index hospitalization, whereas 103 patients were readmitted after 30 days or not at all. With respect to secondary endpoints, there were 21 patients readmitted within 90 days of the index hospitalization, whereas 88 patients were readmitted after 90 days or not at all. Additionally, 6 patients were readmitted ≥ 2 times within 6 months of the index hospitalization, whereas 103 patients were readmitted < 2 times within 6 months.

Baseline characteristics seemed similar across the study groups, except a greater percentage of patients readmitted within 30 days of the index HF hospitalization had a prior history of systolic HF and were hospitalized for HF 30 days prior to the index hospitalization (Table 2). In addition, patients readmitted within 30 days tended to receive a shorter duration of oral diuretic therapy after discontinuation of IV diuretics (mean 0.2 days vs 1.1 days). Patients in this group with an LVEF < 40% were less likely to be discharged on an ACE-I/ARB (75% vs 95%) and a ß-blocker (50% vs 85%) than were the patients who were readmitted after 30 days or not at all. These trends continued for patients readmitted within 90 days of the index hospitalization and for those readmitted after 90 days or not at all. The mean length of stay for the index HF hospitalization was about 5 days and was comparable among all study groups.

From the evaluation of postdischarge characteristics, no patients readmitted within 30 days had a follow-up appointment scheduled with the CHF clinic. In comparison with patients readmitted after 30 days or not at all, more patients had follow-up at an urgent care clinic (33% vs 6%) or no follow-up appointment scheduled at the time of discharge (17% vs 2%). Half of all the patients with a scheduled follow-up missed their appointment. Additionally, medication adherence was lower (33% vs 80%), and none of the patients were enrolled in the CHF-PharmD clinic (0% vs 5%). A similar trend continued for the secondary endpoint groups (Table 3). Last, none of the study patients had an outpatient dietitian consultation.

On readmission, the majority of patients readmitted within 30 days were not on a target dose of an ACE-I/ARB (75%), and none were on a target dose of a ß-blocker. The same trend continued for the secondary endpoint groups. None of the study patients had a serum digoxin level > 0.9 ng/mL. However, serum digoxin level was not measured in all readmitted patients prescribed digoxin (Table 4).

In regard to other secondary endpoints, 6 patients (5.5%) were readmitted for HF within 30 days of the index HF hospitalization. The average number of readmissions per patient in 6 months was < 1, mean time to readmission was 85 days (n = 33), and mean time to death was 88 days (n = 5) when applicable.

DISCUSSION

Based on the trends observed in this study, multiple recommendations can be made to further improve the quality of care and reduce HF readmissions at JBVAMC. The medical center physicians currently use a discharge note template, which already includes sections such as HF discharge instructions and follow-up appointments. The template also prompts providers to prescribe an ACE-I in appropriate patients.

When JBVAMC providers are ready to enter discharge notes into the CPRS, they first select the discharge note template from available note template options. The electronic template contains spaces for the provider to enter a patient’s primary reason for hospitalization, date of admission, discharge medication list, specific or suggested dates for follow-up with outpatient provider(s), general diet/weight/medication instructions, a space to answer whether the patient has HF, a space to record NYHA HF class if applicable, and a space to record whether the patient is prescribed or will be prescribed an ACE-I if appropriate, or whether ACE-I is contraindicated. The providers are able to modify and add information to the discharge note template as they see appropriate.

The findings of this study suggest that modifying the existing discharge template to include additional provider prompts in a form of designated spaces asking for specific information may help improve HF care outcomes. If providers are prompted to answer whether an oral diuretic was continued for at least 24 hours after stopping IV diuretics for HF, adherence to the HF guideline-recommended duration of oral diuretic therapy may improve. Additionally, ß-blocker prescribing in appropriate systolic HF patients may increase if providers are prompted. To enhance continuity of care, the discharge note template may be modified to include a section in which the providers can document patients followed by outside providers. This can be done by incorporating a space in the discharge template to enter the patient’s non-VA provider information if applicable and may help further coordinate the care of such patients to ensure that they are not lost.

Furthermore, the discharge template may be modified to include a prompt to place a CHF clinic consult to increase provider awareness about the availability of CHF and CHF-PharmD clinics at JBVAMC. CHF and CHF-PharmD clinics collaborate to provide comprehensive care to HF patients. After an initial evaluation at the CHF clinic, patients are referred to the clinical pharmacist for further medication therapy management when necessary. Currently, the physicians are encouraged to refer HF patients to the CHF clinic after discharge, but not all providers know that such a service is available. The prompt within the discharge note template would provide CHF/CHF-PharmD clinic provider contact information, clinic times, and a link that would take the provider to an appropriate screen for placing the consult.

Limitations

There are several limitations to this study, including its retrospective design and small sample size. Another source of potential study limitation was the initial process for creating a study patient list. The study list was designed to use ICD-9 codes to capture readmissions only for HF and only at JBVAMC. This was achieved by specifying any of the HF ICD-9 codes as the principal discharge diagnosis. However, the providers may not have always used a HF specific ICD-9 code for the principal discharge diagnosis, even if a patient was admitted primarily for HF. The provider may have chosen another principal discharge diagnosis for which the patient received treatment during the hospitalization.

There are multiple ways to obtain HF patient lists, one includes using the diagnosis-related group codes instead of ICD-9 codes. Due to the way the patient list was obtained and an inherent possibility that some patients admitted for HF had a non-HF ICD-9 code recorded as their principal discharge diagnosis, some eligible patients may not have appeared on the generated list. Additionally, this study captured readmission rates for only HF whereas the national HF 30-day readmission rate represents all-cause readmissions for HF patients. This difference may be reflected in the low 30-day readmission rate observed.

Another possible limitation was the timing of the launch of the CHF-PharmD clinic and the initiative for telephone follow-up 48 hours postdischarge. The CHF-PharmD clinic was launched in April 2011, and the initiative for telephone follow-up 48 hours postdischarge began in January 2011. As the start dates fell within the study period, these services may not have been available to all patients. Therefore, the data describing patient enrollment in CHF-PharmD clinic and those who received postdischarge telephone follow-up may not accurately reflect current practice. Last, statistical tests were not used in the study data analysis leaving any differences found open to interpretation. To minimize these limitations, larger prospective studies with statistical analysis capturing all-cause readmissions are necessary to further evaluate patient characteristics that may be contributing to HF readmissions at JBVAMC.

Conclusions

In general, earlier and more frequent readmissions were more common in patients who were converted to oral diuretic therapy for < 24 hours before discharge and were not discharged on an ACE-I/ARB and a b-blocker when appropriate. Additionally, most of the readmitted patients had no follow-up scheduled at discharge, were nonadherent with medications and follow-up appointments, and were not enrolled in the CHF and/or CHF-PharmD clinic. The majority of patients with systolic HF were not at target doses of either the ACE-I/ARB or the ß-blocker when readmitted. Overall, JBVAMC had a low percentage of patients readmitted for HF within 30 days, but there is still room for improvement in reducing HF readmissions.

At the time of discharge, all JBVAMC patients receive printed instructions and recommendations for their care after hospitalization. The patient handout includes the most current medications, diet/weight/medication instructions, and actual or suggested dates for follow-up appointments and/or tests. It may enhance awareness regarding dietician services to patients if the current discharge instruction template can be modified to provide information regarding the outpatient dietitian class. This could include date, time, and location of classes as well as dietician contact information. (See Appendixes 1 and 2.)

When these recommendations have been implemented, further studies will be warranted to assess the impact of the interventions. 

Acknowledgments
The authors thank Ms. Yvette Bloodson for her assistance in generating the initial patient list.

Author disclosures
The authors report no actual or potential conflicts of interest with regard to this article.

Disclaimer
The opinions expressed herein are those of the authors and do not necessarily reflect those of Federal Practitioner, Frontline Medical Communications Inc., the U.S. Government, or any of its agencies. This article may discuss unlabeled or investigational use of certain drugs. Please review the complete prescribing information for specific drugs or drug combinations—including indications, contraindications, warnings, and adverse effects—before administering pharmacologic therapy to patients.

Heart failure (HF) continues to grow as a significant health problem in the U.S., accounting for 1.1 million hospitalizations annually.¹ About 5.8 million Americans have HF, and 670,000 new cases are diagnosed each year.¹ The prevalence of HF increases with age. Persons aged > 65 years comprise the largest group of patients hospitalized for the condition. Heart failure-related hospitalizations place a major financial burden on patients, caregivers, and the national health care system. In 2010, the estimated cost of health care for HF was $35 billion with hospitalizations accounting for 1% to 2% of the total annual health care costs.^1-3 Furthermore, > 50% of patients with HF are rehospitalized before their first outpatient follow-up.⁴

To ensure patients are ready for discharge, HF guidelines recommend specific interventions for all hospitalized patients. These recommendations include successful transition from IV to oral diuretic therapy as well as the initiation of a ß-blocker and an angiotensin-converting enzyme inhibitor (ACE-I) or an angiotensin receptor blocker (ARB) in stable patients with a left ventricular ejection fraction (LVEF) < 40% and without contraindications. Additionally, patients and their caregivers should receive comprehensive discharge instructions regarding medications, the importance of adherence and regular follow-up, sodium and fluid restriction, weight monitoring, physical activity, and a plan for worsening symptoms. When available, assistance with the hospital-to-home transition should also be provided.^5,6

Heart Failure Measures

Recognizing common factors essential to HF care, the Joint Commission has implemented HF core measures that all U.S. hospitals are required to meet to maintain accreditation status. These guideline-supported measures include receipt of diet, weight, and medication instructions; measured or scheduled assessment of LVEF; ACE-I or an ARB prescribed in patients with LVEF < 40%; and smoking cessation counseling before discharge for all patients with HF.

In addition to HF core measures, 30-day HF readmission rates have also become available to the general public as another hospital quality indicator. In 2009, the Centers for Medicare & Medicaid Services began publicly reporting 30-day HF readmission rates for Medicare patients. A CMS report indicated a 24.8% national 30-day HF readmission rate from July 1, 2007, through June 30, 2010.⁷ Unfortunately, even with the increased quality improvement effort, national HF rehospitalization rates have remained relatively steady in recent years.³

The VA health care system has a growing number of veterans with HF, and it is the leading discharge diagnosis in patients treated at VA hospitals. The number of HF-related hospitalizations at the VA health care system increased from just over 74,000 in fiscal year 2002 to 96,000 in 2009.⁸

To advance the care of veterans with HF and implement best practices, the VA launched the Chronic HF-Quality Enhancement Research Initiative (CHF-QUERI). The major goals of this initiative are to reduce hospitalization rates, increase use of life-prolonging care, empower patients and their caregivers in self-management, and improve appropriateness of HF therapies and tests. As part of its efforts, CHF-QUERI launched the HF Provider Network (HF Network), involving more than 712 VA health care providers (as of July 2014 there were more than 900 providers) committed to improving HF management throughout the entire VA health care system. The HF Network has already put into practice several quality improvement initiatives.

The National Hospital to Home initiative led by the American College of Cardiology and the Institute for Healthcare Improvement was launched throughout the VA system in January 2010.⁹ The main goal of this initiative is to reduce all-cause hospital readmission rates in patients with a discharge diagnosis of HF by improving medication management, early follow-up after discharge, and symptom management.

The Jesse Brown VAMC (JBVAMC) is an active participant of the Hospital to Home initiative, embracing the goals of reducing HF readmission rates and improving the transition of veterans from inpatient to outpatient care. The JBVAMC also has been successfully meeting or exceeding HF core measures except for providing discharge instructions. In May 2011, 91% of patients received discharge instructions, falling just slightly below the 93% target goal. Despite the implementation of HF care improvement initiatives and successful core measure performance, from July 1, 2007, to June 30, 2010, the average HF 30-day readmission rate at JBVAMC was reported to be 28.4%, compared with the national average of 24.8%. Additionally, the average readmission rate for fiscal year 2011 was 31% at JBVAMC, showing a further increase in readmission rates.

The cost of a hospital bed at JBVAMC ranges from about $2,000 to $5,000 per day. According to the American Heart Association’s Get With the Guidelines-HF registry, the mean hospital length of stay for HF in 2009 was 5.5 days.¹ Consequently, HF hospitalizations could potentially cost JBVAMC nearly $7 million annually. Therefore, HF readmissions not only affect patients and caregivers, but also represent a financial burden for JBVAMC. 

METHODS

The purpose of this study was to identify factors contributing to the high HF readmission rates in veterans enrolled at JBVAMC. This study was an Institutional Review Board and VA Research and Development Committee-approved retrospective, electronic chart review of patients with an ICD-9 principal discharge diagnosis code for HF and hospitalization for HF exacerbation anytime between October 1, 2010, and March 1, 2011. A patient chart was reviewed for 6 months after inclusion. A report was generated to identify patients discharged from JBVAMC with a principal discharge diagnosis of HF between October 1, 2010, and March 1, 2011, using the following ICD-9 HF codes: 402.01, 402.11, 402.91, 404.01, 404.03, 404.11, 404.13, 404.91, 404.93, 428.0, 428.1, 428.20, 428.21, 428.22, 428.23, 428.30, 428.31, 428.32, 428.33, 428.40, 428.41, 428.42, 428.43, and 428.9.

Patients were included if aged  ≥ 18 years with one of the ICD-9 HF codes as the principal discharge diagnosis within the study period. Patients were excluded from the study if transferred to or from an outside hospital, discharged without an ICD-9 principal diagnosis code for HF, electively admitted for HF, not treated for HF during hospitalization, left the hospital against medical advice, had chart documentation with comfort measures only, were discharged/transferred to hospice, had active HF medications listed under non-VA medications in the electronic medication profile, or did not receive follow-up at JBVAMC. Study participants were included in the study once, which was classified as their index HF hospitalization, and were followed for 6 months thereafter.

The primary endpoint was the difference in patient characteristics between 2 groups of patients: those readmitted for HF within 30 days of the index hospitalization and those readmitted after 30 days or not at all.

The study had multiple secondary endpoints. One was the difference in patient characteristics between 2 groups of patients: those readmitted for HF within 90 days of the index hospitalization and those readmitted after 90 days or not at all. Another secondary endpoint was the difference in patient characteristics between 2 groups of patients: those with ≥ 2 readmissions for HF within 6 months and those with < 2 HF readmissions within 6 months. Additional secondary endpoints included percentage of patients readmitted for HF within 30 days of the index HF hospitalization, time to readmission if applicable, time to death if applicable, and average number of readmissions per patient within 6 months.

Index data collected included age, gender, ethnicity, prior diagnosis of HF, date of diagnosis, hospitalization for HF within 30 days of the index HF admission, in-hospital cardiac arrest, comorbid conditions, systolic blood pressure (BP), heart rate, respiratory rate, weight, serum sodium, blood urea nitrogen, serum creatinine, hematocrit, and glucose. For this study, comorbid conditions gathered were diabetes mellitus, coronary artery disease, prior percutaneous coronary intervention, aortic stenosis, stroke, chronic obstructive pulmonary disease, and dementia.

Medication profiles were reviewed at the time of admission to determine whether the patient was prescribed an ACE-I/ARB, ß-blocker, diuretic, hydralazine and isosorbide dinitrate, aldosterone antagonist, digoxin, NSAIDs, nonvasoselective calcium channel blocker, and an antiarrhytmic other than amiodarone and dofetilide. Hospitalization data included the most recent LVEF, the number of days on oral diuretic therapy after stopping IV diuretics, the number of days admitted, and documentation of an in-person inpatient dietitian consultation.

Data collected at discharge included diet/weight/medication instructions, weight, BP,  American College of Cardiology/American Heart Association HF stage and New York Heart Association (NYHA) HF functional class, if documented. Discharge medication profiles were assessed for the number of medications (< 9 or ≥ 9), documentation of active prescriptions for an ACE-I/ARB and a ß-blocker (or contraindication documented), diuretic, hydralazine and isosorbide dinitrate, aldosterone antagonist, and digoxin. Other data collected were documentation of a scheduled follow-up appointment with primary care physician, urgent care, chronic HF (CHF) clinic, or cardiologist, and whether the patient was discharged to home, skilled nursing facility, shelter, or homeless. Additionally, if the patient was discharged on a diuretic, the dose was compared with the baseline diuretic. If the diuretic at discharge was different from the home diuretic, equivalent doses were used for comparison with that of the baseline diuretic.

Postdischarge data collection included telephone follow-up within 48 hours of discharge, medication compliance since the initial hospitalization, date of first outpatient follow-up after initial hospital discharge, enrollment in CHF clinic/CHF-PharmD/Care Coordination Home Telehealth (CCHT) program, outpatient dietitian consultations, and date of death if applicable. Medication adherence was defined as ≥ 80% of lowest percentage filled medication of all HF medications, determined by the refill history in the computerized patient record system (CPRS). First outpatient follow-up was defined as a visit in which HF was addressed in the assessment and plan.

If readmitted within the study period, data collection included the date of first nonelective hospital readmission for HF, BP, heart rate, weight, serum digoxin level, serum creatinine, serum potassium, and whether the patient was on a target dose of HF recommended medications (if LVEF < 40% and no contraindication). Heart failure recommended medications for which target doses are established include ACE-I/ARB and ß-blockers. For this study, target doses of ACE-Is were captopril 50 mg 3 times daily, enalapril 10 mg twice daily, fosinopril 40 mg daily, lisinopril 20 mg daily, ramipril 10 mg daily, and trandolapril 4 mg daily. Target doses for ARBs were candesartan 32 mg daily, losartan 50 mg daily, and valsartan 160 mg twice daily. ß-blocker target doses were bisoprolol 10 mg daily, carvedilol 25 mg twice daily (50 mg twice daily if patients’ weight was > 85 kg), and metoprolol succinate 200 mg daily.^5,6 A statistical analysis was not performed on the data.

RESULTS

A total of 137 patient charts were reviewed, and 109 patients were included in the study. Patients were excluded if they transferred to or from an outside hospital (n = 8), had no follow-up at JBVAMC (n = 8), left the hospital against medical advice (n = 4), were electively admitted (n = 4), were not treated for HF (n = 3), or only had comfort measures documented in the chart (n = 1). The patients included were predominantly male (99%) and African American (78%) and had a mean age of 70 years. The majority of the patients had a prior diagnosis of HF (87%) and a history of systolic HF (58%). Most patients were previously prescribed an ACE-I/ARB (83%) and a ß-blocker (76%) at the time of admission (Table 1).

Six patients were readmitted within 30 days of the index hospitalization, whereas 103 patients were readmitted after 30 days or not at all. With respect to secondary endpoints, there were 21 patients readmitted within 90 days of the index hospitalization, whereas 88 patients were readmitted after 90 days or not at all. Additionally, 6 patients were readmitted ≥ 2 times within 6 months of the index hospitalization, whereas 103 patients were readmitted < 2 times within 6 months.

Baseline characteristics seemed similar across the study groups, except a greater percentage of patients readmitted within 30 days of the index HF hospitalization had a prior history of systolic HF and were hospitalized for HF 30 days prior to the index hospitalization (Table 2). In addition, patients readmitted within 30 days tended to receive a shorter duration of oral diuretic therapy after discontinuation of IV diuretics (mean 0.2 days vs 1.1 days). Patients in this group with an LVEF < 40% were less likely to be discharged on an ACE-I/ARB (75% vs 95%) and a ß-blocker (50% vs 85%) than were the patients who were readmitted after 30 days or not at all. These trends continued for patients readmitted within 90 days of the index hospitalization and for those readmitted after 90 days or not at all. The mean length of stay for the index HF hospitalization was about 5 days and was comparable among all study groups.

From the evaluation of postdischarge characteristics, no patients readmitted within 30 days had a follow-up appointment scheduled with the CHF clinic. In comparison with patients readmitted after 30 days or not at all, more patients had follow-up at an urgent care clinic (33% vs 6%) or no follow-up appointment scheduled at the time of discharge (17% vs 2%). Half of all the patients with a scheduled follow-up missed their appointment. Additionally, medication adherence was lower (33% vs 80%), and none of the patients were enrolled in the CHF-PharmD clinic (0% vs 5%). A similar trend continued for the secondary endpoint groups (Table 3). Last, none of the study patients had an outpatient dietitian consultation.

On readmission, the majority of patients readmitted within 30 days were not on a target dose of an ACE-I/ARB (75%), and none were on a target dose of a ß-blocker. The same trend continued for the secondary endpoint groups. None of the study patients had a serum digoxin level > 0.9 ng/mL. However, serum digoxin level was not measured in all readmitted patients prescribed digoxin (Table 4).

In regard to other secondary endpoints, 6 patients (5.5%) were readmitted for HF within 30 days of the index HF hospitalization. The average number of readmissions per patient in 6 months was < 1, mean time to readmission was 85 days (n = 33), and mean time to death was 88 days (n = 5) when applicable.

DISCUSSION

Based on the trends observed in this study, multiple recommendations can be made to further improve the quality of care and reduce HF readmissions at JBVAMC. The medical center physicians currently use a discharge note template, which already includes sections such as HF discharge instructions and follow-up appointments. The template also prompts providers to prescribe an ACE-I in appropriate patients.

When JBVAMC providers are ready to enter discharge notes into the CPRS, they first select the discharge note template from available note template options. The electronic template contains spaces for the provider to enter a patient’s primary reason for hospitalization, date of admission, discharge medication list, specific or suggested dates for follow-up with outpatient provider(s), general diet/weight/medication instructions, a space to answer whether the patient has HF, a space to record NYHA HF class if applicable, and a space to record whether the patient is prescribed or will be prescribed an ACE-I if appropriate, or whether ACE-I is contraindicated. The providers are able to modify and add information to the discharge note template as they see appropriate.

The findings of this study suggest that modifying the existing discharge template to include additional provider prompts in a form of designated spaces asking for specific information may help improve HF care outcomes. If providers are prompted to answer whether an oral diuretic was continued for at least 24 hours after stopping IV diuretics for HF, adherence to the HF guideline-recommended duration of oral diuretic therapy may improve. Additionally, ß-blocker prescribing in appropriate systolic HF patients may increase if providers are prompted. To enhance continuity of care, the discharge note template may be modified to include a section in which the providers can document patients followed by outside providers. This can be done by incorporating a space in the discharge template to enter the patient’s non-VA provider information if applicable and may help further coordinate the care of such patients to ensure that they are not lost.

Furthermore, the discharge template may be modified to include a prompt to place a CHF clinic consult to increase provider awareness about the availability of CHF and CHF-PharmD clinics at JBVAMC. CHF and CHF-PharmD clinics collaborate to provide comprehensive care to HF patients. After an initial evaluation at the CHF clinic, patients are referred to the clinical pharmacist for further medication therapy management when necessary. Currently, the physicians are encouraged to refer HF patients to the CHF clinic after discharge, but not all providers know that such a service is available. The prompt within the discharge note template would provide CHF/CHF-PharmD clinic provider contact information, clinic times, and a link that would take the provider to an appropriate screen for placing the consult.

Limitations

There are several limitations to this study, including its retrospective design and small sample size. Another source of potential study limitation was the initial process for creating a study patient list. The study list was designed to use ICD-9 codes to capture readmissions only for HF and only at JBVAMC. This was achieved by specifying any of the HF ICD-9 codes as the principal discharge diagnosis. However, the providers may not have always used a HF specific ICD-9 code for the principal discharge diagnosis, even if a patient was admitted primarily for HF. The provider may have chosen another principal discharge diagnosis for which the patient received treatment during the hospitalization.

There are multiple ways to obtain HF patient lists, one includes using the diagnosis-related group codes instead of ICD-9 codes. Due to the way the patient list was obtained and an inherent possibility that some patients admitted for HF had a non-HF ICD-9 code recorded as their principal discharge diagnosis, some eligible patients may not have appeared on the generated list. Additionally, this study captured readmission rates for only HF whereas the national HF 30-day readmission rate represents all-cause readmissions for HF patients. This difference may be reflected in the low 30-day readmission rate observed.

Another possible limitation was the timing of the launch of the CHF-PharmD clinic and the initiative for telephone follow-up 48 hours postdischarge. The CHF-PharmD clinic was launched in April 2011, and the initiative for telephone follow-up 48 hours postdischarge began in January 2011. As the start dates fell within the study period, these services may not have been available to all patients. Therefore, the data describing patient enrollment in CHF-PharmD clinic and those who received postdischarge telephone follow-up may not accurately reflect current practice. Last, statistical tests were not used in the study data analysis leaving any differences found open to interpretation. To minimize these limitations, larger prospective studies with statistical analysis capturing all-cause readmissions are necessary to further evaluate patient characteristics that may be contributing to HF readmissions at JBVAMC.

Conclusions

In general, earlier and more frequent readmissions were more common in patients who were converted to oral diuretic therapy for < 24 hours before discharge and were not discharged on an ACE-I/ARB and a b-blocker when appropriate. Additionally, most of the readmitted patients had no follow-up scheduled at discharge, were nonadherent with medications and follow-up appointments, and were not enrolled in the CHF and/or CHF-PharmD clinic. The majority of patients with systolic HF were not at target doses of either the ACE-I/ARB or the ß-blocker when readmitted. Overall, JBVAMC had a low percentage of patients readmitted for HF within 30 days, but there is still room for improvement in reducing HF readmissions.

At the time of discharge, all JBVAMC patients receive printed instructions and recommendations for their care after hospitalization. The patient handout includes the most current medications, diet/weight/medication instructions, and actual or suggested dates for follow-up appointments and/or tests. It may enhance awareness regarding dietician services to patients if the current discharge instruction template can be modified to provide information regarding the outpatient dietitian class. This could include date, time, and location of classes as well as dietician contact information. (See Appendixes 1 and 2.)

When these recommendations have been implemented, further studies will be warranted to assess the impact of the interventions. 

Acknowledgments
The authors thank Ms. Yvette Bloodson for her assistance in generating the initial patient list.

Author disclosures
The authors report no actual or potential conflicts of interest with regard to this article.

Disclaimer
The opinions expressed herein are those of the authors and do not necessarily reflect those of Federal Practitioner, Frontline Medical Communications Inc., the U.S. Government, or any of its agencies. This article may discuss unlabeled or investigational use of certain drugs. Please review the complete prescribing information for specific drugs or drug combinations—including indications, contraindications, warnings, and adverse effects—before administering pharmacologic therapy to patients.

Heart failure (HF) continues to grow as a significant health problem in the U.S., accounting for 1.1 million hospitalizations annually.¹ About 5.8 million Americans have HF, and 670,000 new cases are diagnosed each year.¹ The prevalence of HF increases with age. Persons aged > 65 years comprise the largest group of patients hospitalized for the condition. Heart failure-related hospitalizations place a major financial burden on patients, caregivers, and the national health care system. In 2010, the estimated cost of health care for HF was $35 billion with hospitalizations accounting for 1% to 2% of the total annual health care costs.^1-3 Furthermore, > 50% of patients with HF are rehospitalized before their first outpatient follow-up.⁴

To ensure patients are ready for discharge, HF guidelines recommend specific interventions for all hospitalized patients. These recommendations include successful transition from IV to oral diuretic therapy as well as the initiation of a ß-blocker and an angiotensin-converting enzyme inhibitor (ACE-I) or an angiotensin receptor blocker (ARB) in stable patients with a left ventricular ejection fraction (LVEF) < 40% and without contraindications. Additionally, patients and their caregivers should receive comprehensive discharge instructions regarding medications, the importance of adherence and regular follow-up, sodium and fluid restriction, weight monitoring, physical activity, and a plan for worsening symptoms. When available, assistance with the hospital-to-home transition should also be provided.^5,6

Heart Failure Measures

Recognizing common factors essential to HF care, the Joint Commission has implemented HF core measures that all U.S. hospitals are required to meet to maintain accreditation status. These guideline-supported measures include receipt of diet, weight, and medication instructions; measured or scheduled assessment of LVEF; ACE-I or an ARB prescribed in patients with LVEF < 40%; and smoking cessation counseling before discharge for all patients with HF.

In addition to HF core measures, 30-day HF readmission rates have also become available to the general public as another hospital quality indicator. In 2009, the Centers for Medicare & Medicaid Services began publicly reporting 30-day HF readmission rates for Medicare patients. A CMS report indicated a 24.8% national 30-day HF readmission rate from July 1, 2007, through June 30, 2010.⁷ Unfortunately, even with the increased quality improvement effort, national HF rehospitalization rates have remained relatively steady in recent years.³

The VA health care system has a growing number of veterans with HF, and it is the leading discharge diagnosis in patients treated at VA hospitals. The number of HF-related hospitalizations at the VA health care system increased from just over 74,000 in fiscal year 2002 to 96,000 in 2009.⁸

To advance the care of veterans with HF and implement best practices, the VA launched the Chronic HF-Quality Enhancement Research Initiative (CHF-QUERI). The major goals of this initiative are to reduce hospitalization rates, increase use of life-prolonging care, empower patients and their caregivers in self-management, and improve appropriateness of HF therapies and tests. As part of its efforts, CHF-QUERI launched the HF Provider Network (HF Network), involving more than 712 VA health care providers (as of July 2014 there were more than 900 providers) committed to improving HF management throughout the entire VA health care system. The HF Network has already put into practice several quality improvement initiatives.

The National Hospital to Home initiative led by the American College of Cardiology and the Institute for Healthcare Improvement was launched throughout the VA system in January 2010.⁹ The main goal of this initiative is to reduce all-cause hospital readmission rates in patients with a discharge diagnosis of HF by improving medication management, early follow-up after discharge, and symptom management.

The Jesse Brown VAMC (JBVAMC) is an active participant of the Hospital to Home initiative, embracing the goals of reducing HF readmission rates and improving the transition of veterans from inpatient to outpatient care. The JBVAMC also has been successfully meeting or exceeding HF core measures except for providing discharge instructions. In May 2011, 91% of patients received discharge instructions, falling just slightly below the 93% target goal. Despite the implementation of HF care improvement initiatives and successful core measure performance, from July 1, 2007, to June 30, 2010, the average HF 30-day readmission rate at JBVAMC was reported to be 28.4%, compared with the national average of 24.8%. Additionally, the average readmission rate for fiscal year 2011 was 31% at JBVAMC, showing a further increase in readmission rates.

The cost of a hospital bed at JBVAMC ranges from about $2,000 to $5,000 per day. According to the American Heart Association’s Get With the Guidelines-HF registry, the mean hospital length of stay for HF in 2009 was 5.5 days.¹ Consequently, HF hospitalizations could potentially cost JBVAMC nearly $7 million annually. Therefore, HF readmissions not only affect patients and caregivers, but also represent a financial burden for JBVAMC. 

METHODS

The purpose of this study was to identify factors contributing to the high HF readmission rates in veterans enrolled at JBVAMC. This study was an Institutional Review Board and VA Research and Development Committee-approved retrospective, electronic chart review of patients with an ICD-9 principal discharge diagnosis code for HF and hospitalization for HF exacerbation anytime between October 1, 2010, and March 1, 2011. A patient chart was reviewed for 6 months after inclusion. A report was generated to identify patients discharged from JBVAMC with a principal discharge diagnosis of HF between October 1, 2010, and March 1, 2011, using the following ICD-9 HF codes: 402.01, 402.11, 402.91, 404.01, 404.03, 404.11, 404.13, 404.91, 404.93, 428.0, 428.1, 428.20, 428.21, 428.22, 428.23, 428.30, 428.31, 428.32, 428.33, 428.40, 428.41, 428.42, 428.43, and 428.9.

Patients were included if aged  ≥ 18 years with one of the ICD-9 HF codes as the principal discharge diagnosis within the study period. Patients were excluded from the study if transferred to or from an outside hospital, discharged without an ICD-9 principal diagnosis code for HF, electively admitted for HF, not treated for HF during hospitalization, left the hospital against medical advice, had chart documentation with comfort measures only, were discharged/transferred to hospice, had active HF medications listed under non-VA medications in the electronic medication profile, or did not receive follow-up at JBVAMC. Study participants were included in the study once, which was classified as their index HF hospitalization, and were followed for 6 months thereafter.

The primary endpoint was the difference in patient characteristics between 2 groups of patients: those readmitted for HF within 30 days of the index hospitalization and those readmitted after 30 days or not at all.

The study had multiple secondary endpoints. One was the difference in patient characteristics between 2 groups of patients: those readmitted for HF within 90 days of the index hospitalization and those readmitted after 90 days or not at all. Another secondary endpoint was the difference in patient characteristics between 2 groups of patients: those with ≥ 2 readmissions for HF within 6 months and those with < 2 HF readmissions within 6 months. Additional secondary endpoints included percentage of patients readmitted for HF within 30 days of the index HF hospitalization, time to readmission if applicable, time to death if applicable, and average number of readmissions per patient within 6 months.

Index data collected included age, gender, ethnicity, prior diagnosis of HF, date of diagnosis, hospitalization for HF within 30 days of the index HF admission, in-hospital cardiac arrest, comorbid conditions, systolic blood pressure (BP), heart rate, respiratory rate, weight, serum sodium, blood urea nitrogen, serum creatinine, hematocrit, and glucose. For this study, comorbid conditions gathered were diabetes mellitus, coronary artery disease, prior percutaneous coronary intervention, aortic stenosis, stroke, chronic obstructive pulmonary disease, and dementia.

Medication profiles were reviewed at the time of admission to determine whether the patient was prescribed an ACE-I/ARB, ß-blocker, diuretic, hydralazine and isosorbide dinitrate, aldosterone antagonist, digoxin, NSAIDs, nonvasoselective calcium channel blocker, and an antiarrhytmic other than amiodarone and dofetilide. Hospitalization data included the most recent LVEF, the number of days on oral diuretic therapy after stopping IV diuretics, the number of days admitted, and documentation of an in-person inpatient dietitian consultation.

Data collected at discharge included diet/weight/medication instructions, weight, BP,  American College of Cardiology/American Heart Association HF stage and New York Heart Association (NYHA) HF functional class, if documented. Discharge medication profiles were assessed for the number of medications (< 9 or ≥ 9), documentation of active prescriptions for an ACE-I/ARB and a ß-blocker (or contraindication documented), diuretic, hydralazine and isosorbide dinitrate, aldosterone antagonist, and digoxin. Other data collected were documentation of a scheduled follow-up appointment with primary care physician, urgent care, chronic HF (CHF) clinic, or cardiologist, and whether the patient was discharged to home, skilled nursing facility, shelter, or homeless. Additionally, if the patient was discharged on a diuretic, the dose was compared with the baseline diuretic. If the diuretic at discharge was different from the home diuretic, equivalent doses were used for comparison with that of the baseline diuretic.

Postdischarge data collection included telephone follow-up within 48 hours of discharge, medication compliance since the initial hospitalization, date of first outpatient follow-up after initial hospital discharge, enrollment in CHF clinic/CHF-PharmD/Care Coordination Home Telehealth (CCHT) program, outpatient dietitian consultations, and date of death if applicable. Medication adherence was defined as ≥ 80% of lowest percentage filled medication of all HF medications, determined by the refill history in the computerized patient record system (CPRS). First outpatient follow-up was defined as a visit in which HF was addressed in the assessment and plan.

If readmitted within the study period, data collection included the date of first nonelective hospital readmission for HF, BP, heart rate, weight, serum digoxin level, serum creatinine, serum potassium, and whether the patient was on a target dose of HF recommended medications (if LVEF < 40% and no contraindication). Heart failure recommended medications for which target doses are established include ACE-I/ARB and ß-blockers. For this study, target doses of ACE-Is were captopril 50 mg 3 times daily, enalapril 10 mg twice daily, fosinopril 40 mg daily, lisinopril 20 mg daily, ramipril 10 mg daily, and trandolapril 4 mg daily. Target doses for ARBs were candesartan 32 mg daily, losartan 50 mg daily, and valsartan 160 mg twice daily. ß-blocker target doses were bisoprolol 10 mg daily, carvedilol 25 mg twice daily (50 mg twice daily if patients’ weight was > 85 kg), and metoprolol succinate 200 mg daily.^5,6 A statistical analysis was not performed on the data.

RESULTS

A total of 137 patient charts were reviewed, and 109 patients were included in the study. Patients were excluded if they transferred to or from an outside hospital (n = 8), had no follow-up at JBVAMC (n = 8), left the hospital against medical advice (n = 4), were electively admitted (n = 4), were not treated for HF (n = 3), or only had comfort measures documented in the chart (n = 1). The patients included were predominantly male (99%) and African American (78%) and had a mean age of 70 years. The majority of the patients had a prior diagnosis of HF (87%) and a history of systolic HF (58%). Most patients were previously prescribed an ACE-I/ARB (83%) and a ß-blocker (76%) at the time of admission (Table 1).

Six patients were readmitted within 30 days of the index hospitalization, whereas 103 patients were readmitted after 30 days or not at all. With respect to secondary endpoints, there were 21 patients readmitted within 90 days of the index hospitalization, whereas 88 patients were readmitted after 90 days or not at all. Additionally, 6 patients were readmitted ≥ 2 times within 6 months of the index hospitalization, whereas 103 patients were readmitted < 2 times within 6 months.

Baseline characteristics seemed similar across the study groups, except a greater percentage of patients readmitted within 30 days of the index HF hospitalization had a prior history of systolic HF and were hospitalized for HF 30 days prior to the index hospitalization (Table 2). In addition, patients readmitted within 30 days tended to receive a shorter duration of oral diuretic therapy after discontinuation of IV diuretics (mean 0.2 days vs 1.1 days). Patients in this group with an LVEF < 40% were less likely to be discharged on an ACE-I/ARB (75% vs 95%) and a ß-blocker (50% vs 85%) than were the patients who were readmitted after 30 days or not at all. These trends continued for patients readmitted within 90 days of the index hospitalization and for those readmitted after 90 days or not at all. The mean length of stay for the index HF hospitalization was about 5 days and was comparable among all study groups.

From the evaluation of postdischarge characteristics, no patients readmitted within 30 days had a follow-up appointment scheduled with the CHF clinic. In comparison with patients readmitted after 30 days or not at all, more patients had follow-up at an urgent care clinic (33% vs 6%) or no follow-up appointment scheduled at the time of discharge (17% vs 2%). Half of all the patients with a scheduled follow-up missed their appointment. Additionally, medication adherence was lower (33% vs 80%), and none of the patients were enrolled in the CHF-PharmD clinic (0% vs 5%). A similar trend continued for the secondary endpoint groups (Table 3). Last, none of the study patients had an outpatient dietitian consultation.

On readmission, the majority of patients readmitted within 30 days were not on a target dose of an ACE-I/ARB (75%), and none were on a target dose of a ß-blocker. The same trend continued for the secondary endpoint groups. None of the study patients had a serum digoxin level > 0.9 ng/mL. However, serum digoxin level was not measured in all readmitted patients prescribed digoxin (Table 4).

In regard to other secondary endpoints, 6 patients (5.5%) were readmitted for HF within 30 days of the index HF hospitalization. The average number of readmissions per patient in 6 months was < 1, mean time to readmission was 85 days (n = 33), and mean time to death was 88 days (n = 5) when applicable.

DISCUSSION

Based on the trends observed in this study, multiple recommendations can be made to further improve the quality of care and reduce HF readmissions at JBVAMC. The medical center physicians currently use a discharge note template, which already includes sections such as HF discharge instructions and follow-up appointments. The template also prompts providers to prescribe an ACE-I in appropriate patients.

When JBVAMC providers are ready to enter discharge notes into the CPRS, they first select the discharge note template from available note template options. The electronic template contains spaces for the provider to enter a patient’s primary reason for hospitalization, date of admission, discharge medication list, specific or suggested dates for follow-up with outpatient provider(s), general diet/weight/medication instructions, a space to answer whether the patient has HF, a space to record NYHA HF class if applicable, and a space to record whether the patient is prescribed or will be prescribed an ACE-I if appropriate, or whether ACE-I is contraindicated. The providers are able to modify and add information to the discharge note template as they see appropriate.

The findings of this study suggest that modifying the existing discharge template to include additional provider prompts in a form of designated spaces asking for specific information may help improve HF care outcomes. If providers are prompted to answer whether an oral diuretic was continued for at least 24 hours after stopping IV diuretics for HF, adherence to the HF guideline-recommended duration of oral diuretic therapy may improve. Additionally, ß-blocker prescribing in appropriate systolic HF patients may increase if providers are prompted. To enhance continuity of care, the discharge note template may be modified to include a section in which the providers can document patients followed by outside providers. This can be done by incorporating a space in the discharge template to enter the patient’s non-VA provider information if applicable and may help further coordinate the care of such patients to ensure that they are not lost.

Furthermore, the discharge template may be modified to include a prompt to place a CHF clinic consult to increase provider awareness about the availability of CHF and CHF-PharmD clinics at JBVAMC. CHF and CHF-PharmD clinics collaborate to provide comprehensive care to HF patients. After an initial evaluation at the CHF clinic, patients are referred to the clinical pharmacist for further medication therapy management when necessary. Currently, the physicians are encouraged to refer HF patients to the CHF clinic after discharge, but not all providers know that such a service is available. The prompt within the discharge note template would provide CHF/CHF-PharmD clinic provider contact information, clinic times, and a link that would take the provider to an appropriate screen for placing the consult.

Limitations

There are several limitations to this study, including its retrospective design and small sample size. Another source of potential study limitation was the initial process for creating a study patient list. The study list was designed to use ICD-9 codes to capture readmissions only for HF and only at JBVAMC. This was achieved by specifying any of the HF ICD-9 codes as the principal discharge diagnosis. However, the providers may not have always used a HF specific ICD-9 code for the principal discharge diagnosis, even if a patient was admitted primarily for HF. The provider may have chosen another principal discharge diagnosis for which the patient received treatment during the hospitalization.

There are multiple ways to obtain HF patient lists, one includes using the diagnosis-related group codes instead of ICD-9 codes. Due to the way the patient list was obtained and an inherent possibility that some patients admitted for HF had a non-HF ICD-9 code recorded as their principal discharge diagnosis, some eligible patients may not have appeared on the generated list. Additionally, this study captured readmission rates for only HF whereas the national HF 30-day readmission rate represents all-cause readmissions for HF patients. This difference may be reflected in the low 30-day readmission rate observed.

Another possible limitation was the timing of the launch of the CHF-PharmD clinic and the initiative for telephone follow-up 48 hours postdischarge. The CHF-PharmD clinic was launched in April 2011, and the initiative for telephone follow-up 48 hours postdischarge began in January 2011. As the start dates fell within the study period, these services may not have been available to all patients. Therefore, the data describing patient enrollment in CHF-PharmD clinic and those who received postdischarge telephone follow-up may not accurately reflect current practice. Last, statistical tests were not used in the study data analysis leaving any differences found open to interpretation. To minimize these limitations, larger prospective studies with statistical analysis capturing all-cause readmissions are necessary to further evaluate patient characteristics that may be contributing to HF readmissions at JBVAMC.

Conclusions

In general, earlier and more frequent readmissions were more common in patients who were converted to oral diuretic therapy for < 24 hours before discharge and were not discharged on an ACE-I/ARB and a b-blocker when appropriate. Additionally, most of the readmitted patients had no follow-up scheduled at discharge, were nonadherent with medications and follow-up appointments, and were not enrolled in the CHF and/or CHF-PharmD clinic. The majority of patients with systolic HF were not at target doses of either the ACE-I/ARB or the ß-blocker when readmitted. Overall, JBVAMC had a low percentage of patients readmitted for HF within 30 days, but there is still room for improvement in reducing HF readmissions.

At the time of discharge, all JBVAMC patients receive printed instructions and recommendations for their care after hospitalization. The patient handout includes the most current medications, diet/weight/medication instructions, and actual or suggested dates for follow-up appointments and/or tests. It may enhance awareness regarding dietician services to patients if the current discharge instruction template can be modified to provide information regarding the outpatient dietitian class. This could include date, time, and location of classes as well as dietician contact information. (See Appendixes 1 and 2.)

When these recommendations have been implemented, further studies will be warranted to assess the impact of the interventions. 

Acknowledgments
The authors thank Ms. Yvette Bloodson for her assistance in generating the initial patient list.

Author disclosures
The authors report no actual or potential conflicts of interest with regard to this article.

Disclaimer
The opinions expressed herein are those of the authors and do not necessarily reflect those of Federal Practitioner, Frontline Medical Communications Inc., the U.S. Government, or any of its agencies. This article may discuss unlabeled or investigational use of certain drugs. Please review the complete prescribing information for specific drugs or drug combinations—including indications, contraindications, warnings, and adverse effects—before administering pharmacologic therapy to patients.

Although the elevated risks and rates of veterans’ substance abuse patterns are well documented, little has been written about veterans’ use of designer drugs.^1-6 In recent months throughout Europe and the U.S., there has been a flurry of media attention for 2 classes of designer drugs: synthetic cathinones and synthetic cannabinoids.^7,8 In the U.S., the popularity of these drugs has surged, and a disproportionate amount of use of these 2 drug classes is coming from locations near military instillations.^9,10

The purpose of this article is to raise awareness regarding these 2 burgeoning designer drug classes and their impact on veterans. Designer drugs affecting vulnerable populations are not a new phenomenon, yet many providers are unfamiliar with the effects of these unique drugs of abuse on their veteran populations.^11-13

Many designer drugs begin their existence as variations of other addictive or psychoactive drugs. Others begin in laboratories as investigative research compounds that end up on the street, often promising a novel mind-altering experience as a “legal high.”^14-18 The Designer Drug Enforcement Act of 1986 was an initial attempt in the U.S. to define and control the early rise of copycat drugs that appeared on the streets and mimicked the effects of other illicit substances. More recent legislation enacted in the U.S. has imposed Schedule I controls on the manufacture, distribution, possession, importation, and exportation of these types of drugs, including both synthetic cathinones and synthetic cannabinoids. State laws are perennially in flux trying to keep up with the latest drug trends.^19-21

Similar efforts have been made by the European Union to control mephedrone, a synthetic cathinone, citing multiple fatalities, seizures, related crime, lack of medical use, and risk of dependence.²² Although uniform levels of control do not exist in Europe for synthetic cannabinoids, many countries have independently acted to limit their use.²³

In its recent World Drug Report 2013, the United Nations Office on Drugs and Crime documents its growing concern about the “new psychoactive substances” category of illicit recreational substances (in which synthetic cannabinoids and cathinones are included) that has increased by 50% since 2009.²⁴ Alone, this category now outnumbers the total number of substances controlled by international drug conventions.

The novelty and variability of designer drugs causes difficulties with detection and regulation. Innovative chemists can legally manufacture new versions of known molecules intended for illicit use with a rapidity that outpaces bureaucratic control. Local law enforcement officials may be unaware of the latest designer drug trends, stifling efforts at public education or restriction. Designer drugs are often deceptively packaged and are available in convenience stores, tobacco outlets, gas stations, pawnshops, tattoo parlors, and truck stops.^25-28 The Internet may be the singular reason, however, that designer drugs continue to be widely available to veterans.^11,18

Innumerable websites discuss, promote, and sell designer drugs or deceitfully market them as safe, legitimate household products (“not for human consumption”), which can be ordered online and shipped by commercial carriers.¹² Little accurate information is known about their effects or about the specific compounds they contain. When the recreational nature of the drugs is actually acknowledged, information on how the buyer can evade prosecution is often provided in tandem. The suppliers’ inventory of the drugs has been shown to be variable and inconsistent, and the product ingredients can be similarly unpredictable despite comparatively more stable naming and labeling.^14,29

In the clinical setting, a reliable patient drug history may not be available. This ensures that the diagnosis of designer drug use will be an exclusionary process involving routine laboratory work, physical examination, and at times electroencephalogram and/or neuroimaging. Psychiatric consultation is often useful in this setting. Routine immunoassay tests do not detect either synthetic cathinones or synthetic cannabinoids.³⁰

Both cannabinoids and cathinones can be identified using gas chromatography-mass spectroscopy (GC-MS) or liquid chromatography-mass spectroscopy (LC-MS). However, this technology is limited to specialized laboratories.^31,32 The laboratory results often are not immediately available, potentially limiting the tests’ use in emergency or inpatient settings, as the patient may have left the hospital by the time the results are available. Additionally, these drugs’ prevalence of use, while increasing, often does not justify the cost of these tests.

The inability to routinely detect metabolites in urine may increase the enticement of these drugs given the likelihood that active-duty personnel could use them surreptitiously. Further, these compounds are evolving and seemingly limitless in their variability, and there is often a paucity of pure reference materials. As such, it is impossible to guarantee reliable test results.

The following profiles of each of these drug classes will be accompanied by clinical cases depicting the drugs’ effects and how an affected veteran might present clinically. The severe effects of these novel agents illustrate the value in maintaining a functional knowledge base about emerging drug trends. The accuracy of diagnosis as well as the outcome of a veteran’s treatment may depend on the provider’s ability to identify the presence of a drug and manage its effects.

Synthetic Cathinones

Mr. H is a 28-year-old Iraq War veteran with a history of posttraumatic stress disorder (PTSD), alcohol abuse, and opioid dependence who presented for inpatient psychiatric admission after making suicidal statements to his wife in the context of 2 weeks of “bath salts” use. A family member initially introduced him to the drug. His first drug purchase had been 1 gram ($30) at a local movie rental store.

After discharge from the hospital, Mr. H began purchasing increasing amounts online with a credit card. Although he initially had been insufflating and inhaling the substance, he later began injecting it (dissolving it in tap water and loading it through a cotton filter in a syringe). The patient admitted to finding the drug significantly more addictive than any others he had used, and his use resulted in leaving his job and abandoning his family.

Severe cravings and depression were present between episodes of use. He spent $40,000 over 6 months of use. Insomnia lasted for several days, his appearance changed dramatically (including persistent skin infections), and he became paranoid, believing that everyone around him was an undercover police officer. He remained on medications for persistent anxiety. His daily drug cravings continued,
although he remained uncertain about the actual ingredients of bath salts.

Cathinone is a naturally occurring stimulant from the khat plant (Catha edulis), which grows indigenously in Egypt and on the Arabian Peninsula. The recreational and religious use of this plant has occurred for thousands of years, though it is not without risk: The chewing of the leaves containing natural cathinones has been associated with esophagitis, gastritis, oral keratosis, myocardial infarction, dilated cardiomyopathy, hypertension, cerebral ischemia, thromboembolism, diabetes, sexual dysfunction, duodenal ulcer, and hepatitis.^33,34

The stimulants known colloquially as bath salts are synthetic cathinones, which have become more widely available within the past 10 years: first in the Middle East, then Europe, and now in the U.S.^{5,9,10,14,19,25,35-41,44} Although the current rise in use has occurred in the past few years, the first documented abuse of synthetic cathinones in the U.S. dates to the early 1990s in Michigan.⁴²

Bath salts is the most common of the many names used to denote synthetic cathinones. The compounds have no utility when used as such but often are marketed as research chemicals, plant fertilizer, or shoe polish. It is this deliberate counterfeit of household product names that allows many distributors to avoid classifying the compounds according to the true, intended use. More appealing brand names may also be used to entice the user (Table 1).²⁵

Synthetic cathinones owe their popularity to similarities with cocaine and methamphetamine. They are sympathomimetic with synaptic increases of monoamines after use: Surges in norepinephrine and dopamine account for the stimulant qualities, and serotonergic changes mediate distinct psychoactive effects (Table 2).⁴⁰ Users are interested in the drugs for many of the same reasons that other recreational stimulants have appeal: euphoria, energy, empathy, heightened sexuality, sociability, and an overall intensification of senses. Synthetic cathinones have become preferred to cocaine for some users.⁴³

The drugs can be used via oral and anal routes. Using methods known as “bombing” or “keystering,” users deliver boluses of the powder wrapped in cigarette paper, which they swallow or insert into the rectum. Insufflation and IV injection are also common methods of administration with a quicker onset of action expected.⁴⁰ The prices of the drugs range from $25 to $50 per 500-mg packet (though the cost is increasing with more regulation). Users typically use 500 mg to 2 g in one session.

The 2 most commonly abused synthetic cathinones are mephedrone and MDPV (methylenedioxypyrovalerone). There is some regional variability about which ingredient is present; mephedrone tends to be more prevalent in Europe, whereas MDPV is noted to be more common in the U.S.^10,44

When ordering a laboratory test to evaluate for the presence of these drugs, a specific request should be given to the technicians to look for signals of MDPV (most common metabolite is dimethylenyl-methyl-MDPV), mephedrone (4-methylmethcathinone), 3-bromomethcathinone (3-BMC), or 3-fluoromethcathinone (fluphedrone).^45-48 The study testing (both in VA and civilian settings) for Mr. H was done by a commercial laboratory several states away where patented techniques can screen for more than 30 compounds via LC-MS. The laboratory offered bath salts panels for urine, serum/plasma, and blood samples.

Synthetic cathinones are dangerous, and as the body of medical literature continues to expand, reports of significant morbidity and death related to their use are appearing. The harmful effects of recreational synthetic cathinone use has been documented across the globe in the form of serotonin syndrome, intoxication delirium, hyperthermia and multi-organ failure, myocarditis, hypo-osmotic hyponatremia with encephalopathy, agitation, psychosis, and death after cardiac arrest.^{5,12,38,39,49-53} Published treatment methods are largely supportive with the available literature, suggesting that benzodiazepines, antipsychotics (both typical and atypical), restraints to maintain safety, and IV fluids may be indicated.^5,9,50

Synthetic Cannabinoids

Mr. W is a 58-year-old veteran with a history of alcohol dependence and PTSD who reported use of the synthetic cannabinoid “Spice” during intake assessment for treatment of alcohol dependence. He reported using Spice about 4 times over a 2-month period. He purchased a small jar of the substance from a party store for $15 per gram and understood its contents to be synthetic marijuana, which he appreciated for its low cost and assumed legality. He denied having any understanding of the package’s contents beyond “synthetic marijuana.”

The patient ingested the drug by smoking and inhaling from a pipe. For the first 3 times that he used the substance, Mr. W reported feeling a pleasant sensation that started quickly and lasted about 30 minutes. The fourth time that he used synthetic cannabis he felt nauseated and vomited several times, had auditory hallucinations, and increased anxiety; he also reported a hangover effect after this use. He identified that the effects may have been different the fourth time “because the brands were changing.”

Mr. W also reported that his neighbor—a daily user of synthetic cannabinoids for several months—became paranoid, suspicious, and developed incomprehensible speech. His neighbor’s symptoms and his own unpleasant experiences prompted a discontinuation of use.

Synthetic cannabinoids are a diverse group of agents numbering in excess of 100 artificial compounds that act as agonists at cannabinoid receptors, mimicking the effects of tetrahydrocannabinol (THC), an active ingredient in marijuana.^28,54 The availability of these drugs online and in specialty shops has been documented since the mid-2000s.^27,28,32 Their packaging often describes the contents as incense or herbal blends, using various names. Spice is a common name, but these products are also known by a myriad of other designations (Table 1).²⁸ A single packet usually contains several grams of the drug and costs about $30.⁵⁵

To the user, who may already be familiar with marijuana, the contents intentionally appear similar to the dried buds of cannabis.^30,56 In reality, the drug has just been sprayed onto inert plant material.⁵⁷ The drug is smoked, and the psychoactive dose can be as little as 1 mg.³⁰ Users describe potent drug effects (Table 2). There is a rapid onset of action, and duration of effects last 1 to 2 hours.⁵⁸

The compounds’ mechanism of action and appeal are derived from their high affinity for the cannabinoid receptors. The CB1 receptor is located primarily in the central nervous system and is responsible for the psychoactive component of the drugs’ actions.^27,30,58,59 Two particular synthetic cannabinoids, cannibicyclohexanol and JWH-018, are potent cannabinoid CB1 agonists with affinity exceeding their natural counterparts.^{27,30, 32,56,58,59}

Chemically, these drugs are varied. The largest structural family of these compounds is the JWH group, which includes JWH-018.⁶⁰ Also common are CP 47,497 and other CP compounds.⁵⁸ HU compounds, such as HU-210, have also been identified and have been shown to be 100 to 800 times more potent at the CB1 and CB2 receptors than is THC.^60,61 A final group includes the benzoylindoles, such as AM-964 and RCS-4, which also bind strongly to CB1 and CB2.^60,62

Constitutional symptoms of synthetic cannabinoid intoxication include disorientation, anxiety, tremulousness, palpitations, tachycardia, agitation, injected conjunctivae, hyperreflexia, nausea, vomiting, lateral gaze nystagmus, and myoclonic jerks, which have been mistaken for seizure activity.^27,30,55 Pupils are often normal sized.⁵⁵ Withdrawal phenomena are similar to those of cannabis withdrawal: irritability, anxiety, tremor, palpitations, diaphoresis, insomnia, headache, diarrhea, nausea, and vomiting.⁵⁹

Given the established link between cannabis use and psychosis, synthetic cannabinoids may stand as a precipitant of psychotic symptoms, which may include visual hallucinations, auditory hallucinations, disorganized speech, paranoia, grandiose delusions, disorganization, or bizarre behavior.^58,63-66 These symptoms may represent a relapse of a primary thought disorder or, for some unfortunate individuals, a de novo psychotic illness.^58,65,66 Symptoms can linger for months after drug use.⁶⁵

A key risk in the use of synthetic cannabinoid moieties may involve the absence of cannabidiol. Cannabidiol naturally occurs in many strains of cannabis and is thought to have antipsychotic, neuroprotective properties.⁶⁷ The absence of this molecule in synthetic cannabinoids may at least partially explain their severe psychoactive effects. Treatment for synthetic cannabinoid intoxication and related psychosis is largely supportive and may include the use of antipsychotic medication.⁶⁶

Detection of synthetic cannabinoids in urine is difficult, yet many compounds can be detected via GC-MS or LC-MS. Molecules of significance include JWH-018, JWH-073, JWH-015, JWH-250, CP-47 497, HU-210, cannabicyclohexanol, and oleamide; however, these compounds are rarely excreted in urine in their pure form. The many hydroxylated or dealkylated metabolites of these compounds, mostly unnamed, are more consistently detected in urine.^68,69 One author has noted that the pentanoic acid metabolite of JWH-018 seems to appear most reliably in urine specimens.⁶⁸

Many synthetic cannabinoid herbal mixes also contain a detectable compound called tocopherol, seemingly added as an antioxidant.^69,70 Synthetic cannabinoids are an evolving drug class, and reliable detection will require that laboratories stay up-to-date in their detection methods. As stated earlier, a commercial laboratory in the region accepted the civilian and veteran patient samples for these case studies. The synthetic cannabinoid panels offered evaluation of the drug itself (GC-MS), an oral fluid screen (LC-MS), and isolation of metabolites in urine (enzyme-linked immunosorbent assay).

Conclusion

Designer drugs will remain a challenge for providers caring for veterans for several key reasons: (1) Veterans are a vulnerable population who abuse substances at higher rates than do their civilian counterparts; (2) Chemists are able to manufacture variations of known habit-forming substances; (3) Modern technology facilitates the purchase and wide distribution of addictive substances; (4) Many designer drugs are deceptively packaged and marketed; (5) The effects of the drugs are often severe; (6) No standardized treatment guidelines exist; and (7) Detection of the drugs is difficult, and new versions of the molecules may evade even cutting-edge techniques.

Due to the high cost of detecting synthetic cathinones and synthetic cannabinoids in body fluids, screening should be considered only in settings where severe symptoms are accompanied by reasonable clinical suspicion of use and an otherwise negative toxicologic workup. As more designer drugs inevitably emerge, research will be needed on their pharmacology, toxidromes, and detection. Military and civilian practitioners must remain abreast of the dynamic trends in designer drugs to ensure that their patients receive the highest level of medical care.

Author disclosures
The authors report no actual or potential conflicts of interest with regard to this article.  

Disclaimer
The opinions expressed herein are those of the authors and do not necessarily reflect those of Federal Practitioner, Frontline Medical Communications Inc., the U.S. Government, or any of its agencies. This article may discuss unlabeled or investigational use of certain drugs. Please review complete prescribing information for specific drugs or drug combinations—including indications, contraindications, warnings, and adverse effects—before administering pharmacologic therapy to patients.

Although the elevated risks and rates of veterans’ substance abuse patterns are well documented, little has been written about veterans’ use of designer drugs.^1-6 In recent months throughout Europe and the U.S., there has been a flurry of media attention for 2 classes of designer drugs: synthetic cathinones and synthetic cannabinoids.^7,8 In the U.S., the popularity of these drugs has surged, and a disproportionate amount of use of these 2 drug classes is coming from locations near military instillations.^9,10

The purpose of this article is to raise awareness regarding these 2 burgeoning designer drug classes and their impact on veterans. Designer drugs affecting vulnerable populations are not a new phenomenon, yet many providers are unfamiliar with the effects of these unique drugs of abuse on their veteran populations.^11-13

Many designer drugs begin their existence as variations of other addictive or psychoactive drugs. Others begin in laboratories as investigative research compounds that end up on the street, often promising a novel mind-altering experience as a “legal high.”^14-18 The Designer Drug Enforcement Act of 1986 was an initial attempt in the U.S. to define and control the early rise of copycat drugs that appeared on the streets and mimicked the effects of other illicit substances. More recent legislation enacted in the U.S. has imposed Schedule I controls on the manufacture, distribution, possession, importation, and exportation of these types of drugs, including both synthetic cathinones and synthetic cannabinoids. State laws are perennially in flux trying to keep up with the latest drug trends.^19-21

Similar efforts have been made by the European Union to control mephedrone, a synthetic cathinone, citing multiple fatalities, seizures, related crime, lack of medical use, and risk of dependence.²² Although uniform levels of control do not exist in Europe for synthetic cannabinoids, many countries have independently acted to limit their use.²³

In its recent World Drug Report 2013, the United Nations Office on Drugs and Crime documents its growing concern about the “new psychoactive substances” category of illicit recreational substances (in which synthetic cannabinoids and cathinones are included) that has increased by 50% since 2009.²⁴ Alone, this category now outnumbers the total number of substances controlled by international drug conventions.

The novelty and variability of designer drugs causes difficulties with detection and regulation. Innovative chemists can legally manufacture new versions of known molecules intended for illicit use with a rapidity that outpaces bureaucratic control. Local law enforcement officials may be unaware of the latest designer drug trends, stifling efforts at public education or restriction. Designer drugs are often deceptively packaged and are available in convenience stores, tobacco outlets, gas stations, pawnshops, tattoo parlors, and truck stops.^25-28 The Internet may be the singular reason, however, that designer drugs continue to be widely available to veterans.^11,18

Innumerable websites discuss, promote, and sell designer drugs or deceitfully market them as safe, legitimate household products (“not for human consumption”), which can be ordered online and shipped by commercial carriers.¹² Little accurate information is known about their effects or about the specific compounds they contain. When the recreational nature of the drugs is actually acknowledged, information on how the buyer can evade prosecution is often provided in tandem. The suppliers’ inventory of the drugs has been shown to be variable and inconsistent, and the product ingredients can be similarly unpredictable despite comparatively more stable naming and labeling.^14,29

In the clinical setting, a reliable patient drug history may not be available. This ensures that the diagnosis of designer drug use will be an exclusionary process involving routine laboratory work, physical examination, and at times electroencephalogram and/or neuroimaging. Psychiatric consultation is often useful in this setting. Routine immunoassay tests do not detect either synthetic cathinones or synthetic cannabinoids.³⁰

Both cannabinoids and cathinones can be identified using gas chromatography-mass spectroscopy (GC-MS) or liquid chromatography-mass spectroscopy (LC-MS). However, this technology is limited to specialized laboratories.^31,32 The laboratory results often are not immediately available, potentially limiting the tests’ use in emergency or inpatient settings, as the patient may have left the hospital by the time the results are available. Additionally, these drugs’ prevalence of use, while increasing, often does not justify the cost of these tests.

The inability to routinely detect metabolites in urine may increase the enticement of these drugs given the likelihood that active-duty personnel could use them surreptitiously. Further, these compounds are evolving and seemingly limitless in their variability, and there is often a paucity of pure reference materials. As such, it is impossible to guarantee reliable test results.

The following profiles of each of these drug classes will be accompanied by clinical cases depicting the drugs’ effects and how an affected veteran might present clinically. The severe effects of these novel agents illustrate the value in maintaining a functional knowledge base about emerging drug trends. The accuracy of diagnosis as well as the outcome of a veteran’s treatment may depend on the provider’s ability to identify the presence of a drug and manage its effects.

Synthetic Cathinones

Mr. H is a 28-year-old Iraq War veteran with a history of posttraumatic stress disorder (PTSD), alcohol abuse, and opioid dependence who presented for inpatient psychiatric admission after making suicidal statements to his wife in the context of 2 weeks of “bath salts” use. A family member initially introduced him to the drug. His first drug purchase had been 1 gram ($30) at a local movie rental store.

After discharge from the hospital, Mr. H began purchasing increasing amounts online with a credit card. Although he initially had been insufflating and inhaling the substance, he later began injecting it (dissolving it in tap water and loading it through a cotton filter in a syringe). The patient admitted to finding the drug significantly more addictive than any others he had used, and his use resulted in leaving his job and abandoning his family.

Severe cravings and depression were present between episodes of use. He spent $40,000 over 6 months of use. Insomnia lasted for several days, his appearance changed dramatically (including persistent skin infections), and he became paranoid, believing that everyone around him was an undercover police officer. He remained on medications for persistent anxiety. His daily drug cravings continued,
although he remained uncertain about the actual ingredients of bath salts.

Cathinone is a naturally occurring stimulant from the khat plant (Catha edulis), which grows indigenously in Egypt and on the Arabian Peninsula. The recreational and religious use of this plant has occurred for thousands of years, though it is not without risk: The chewing of the leaves containing natural cathinones has been associated with esophagitis, gastritis, oral keratosis, myocardial infarction, dilated cardiomyopathy, hypertension, cerebral ischemia, thromboembolism, diabetes, sexual dysfunction, duodenal ulcer, and hepatitis.^33,34

The stimulants known colloquially as bath salts are synthetic cathinones, which have become more widely available within the past 10 years: first in the Middle East, then Europe, and now in the U.S.^{5,9,10,14,19,25,35-41,44} Although the current rise in use has occurred in the past few years, the first documented abuse of synthetic cathinones in the U.S. dates to the early 1990s in Michigan.⁴²

Bath salts is the most common of the many names used to denote synthetic cathinones. The compounds have no utility when used as such but often are marketed as research chemicals, plant fertilizer, or shoe polish. It is this deliberate counterfeit of household product names that allows many distributors to avoid classifying the compounds according to the true, intended use. More appealing brand names may also be used to entice the user (Table 1).²⁵

Synthetic cathinones owe their popularity to similarities with cocaine and methamphetamine. They are sympathomimetic with synaptic increases of monoamines after use: Surges in norepinephrine and dopamine account for the stimulant qualities, and serotonergic changes mediate distinct psychoactive effects (Table 2).⁴⁰ Users are interested in the drugs for many of the same reasons that other recreational stimulants have appeal: euphoria, energy, empathy, heightened sexuality, sociability, and an overall intensification of senses. Synthetic cathinones have become preferred to cocaine for some users.⁴³

The drugs can be used via oral and anal routes. Using methods known as “bombing” or “keystering,” users deliver boluses of the powder wrapped in cigarette paper, which they swallow or insert into the rectum. Insufflation and IV injection are also common methods of administration with a quicker onset of action expected.⁴⁰ The prices of the drugs range from $25 to $50 per 500-mg packet (though the cost is increasing with more regulation). Users typically use 500 mg to 2 g in one session.

The 2 most commonly abused synthetic cathinones are mephedrone and MDPV (methylenedioxypyrovalerone). There is some regional variability about which ingredient is present; mephedrone tends to be more prevalent in Europe, whereas MDPV is noted to be more common in the U.S.^10,44

When ordering a laboratory test to evaluate for the presence of these drugs, a specific request should be given to the technicians to look for signals of MDPV (most common metabolite is dimethylenyl-methyl-MDPV), mephedrone (4-methylmethcathinone), 3-bromomethcathinone (3-BMC), or 3-fluoromethcathinone (fluphedrone).^45-48 The study testing (both in VA and civilian settings) for Mr. H was done by a commercial laboratory several states away where patented techniques can screen for more than 30 compounds via LC-MS. The laboratory offered bath salts panels for urine, serum/plasma, and blood samples.

Synthetic cathinones are dangerous, and as the body of medical literature continues to expand, reports of significant morbidity and death related to their use are appearing. The harmful effects of recreational synthetic cathinone use has been documented across the globe in the form of serotonin syndrome, intoxication delirium, hyperthermia and multi-organ failure, myocarditis, hypo-osmotic hyponatremia with encephalopathy, agitation, psychosis, and death after cardiac arrest.^{5,12,38,39,49-53} Published treatment methods are largely supportive with the available literature, suggesting that benzodiazepines, antipsychotics (both typical and atypical), restraints to maintain safety, and IV fluids may be indicated.^5,9,50

Synthetic Cannabinoids

Mr. W is a 58-year-old veteran with a history of alcohol dependence and PTSD who reported use of the synthetic cannabinoid “Spice” during intake assessment for treatment of alcohol dependence. He reported using Spice about 4 times over a 2-month period. He purchased a small jar of the substance from a party store for $15 per gram and understood its contents to be synthetic marijuana, which he appreciated for its low cost and assumed legality. He denied having any understanding of the package’s contents beyond “synthetic marijuana.”

The patient ingested the drug by smoking and inhaling from a pipe. For the first 3 times that he used the substance, Mr. W reported feeling a pleasant sensation that started quickly and lasted about 30 minutes. The fourth time that he used synthetic cannabis he felt nauseated and vomited several times, had auditory hallucinations, and increased anxiety; he also reported a hangover effect after this use. He identified that the effects may have been different the fourth time “because the brands were changing.”

Mr. W also reported that his neighbor—a daily user of synthetic cannabinoids for several months—became paranoid, suspicious, and developed incomprehensible speech. His neighbor’s symptoms and his own unpleasant experiences prompted a discontinuation of use.

Synthetic cannabinoids are a diverse group of agents numbering in excess of 100 artificial compounds that act as agonists at cannabinoid receptors, mimicking the effects of tetrahydrocannabinol (THC), an active ingredient in marijuana.^28,54 The availability of these drugs online and in specialty shops has been documented since the mid-2000s.^27,28,32 Their packaging often describes the contents as incense or herbal blends, using various names. Spice is a common name, but these products are also known by a myriad of other designations (Table 1).²⁸ A single packet usually contains several grams of the drug and costs about $30.⁵⁵

To the user, who may already be familiar with marijuana, the contents intentionally appear similar to the dried buds of cannabis.^30,56 In reality, the drug has just been sprayed onto inert plant material.⁵⁷ The drug is smoked, and the psychoactive dose can be as little as 1 mg.³⁰ Users describe potent drug effects (Table 2). There is a rapid onset of action, and duration of effects last 1 to 2 hours.⁵⁸

The compounds’ mechanism of action and appeal are derived from their high affinity for the cannabinoid receptors. The CB1 receptor is located primarily in the central nervous system and is responsible for the psychoactive component of the drugs’ actions.^27,30,58,59 Two particular synthetic cannabinoids, cannibicyclohexanol and JWH-018, are potent cannabinoid CB1 agonists with affinity exceeding their natural counterparts.^{27,30, 32,56,58,59}

Chemically, these drugs are varied. The largest structural family of these compounds is the JWH group, which includes JWH-018.⁶⁰ Also common are CP 47,497 and other CP compounds.⁵⁸ HU compounds, such as HU-210, have also been identified and have been shown to be 100 to 800 times more potent at the CB1 and CB2 receptors than is THC.^60,61 A final group includes the benzoylindoles, such as AM-964 and RCS-4, which also bind strongly to CB1 and CB2.^60,62

Constitutional symptoms of synthetic cannabinoid intoxication include disorientation, anxiety, tremulousness, palpitations, tachycardia, agitation, injected conjunctivae, hyperreflexia, nausea, vomiting, lateral gaze nystagmus, and myoclonic jerks, which have been mistaken for seizure activity.^27,30,55 Pupils are often normal sized.⁵⁵ Withdrawal phenomena are similar to those of cannabis withdrawal: irritability, anxiety, tremor, palpitations, diaphoresis, insomnia, headache, diarrhea, nausea, and vomiting.⁵⁹

Given the established link between cannabis use and psychosis, synthetic cannabinoids may stand as a precipitant of psychotic symptoms, which may include visual hallucinations, auditory hallucinations, disorganized speech, paranoia, grandiose delusions, disorganization, or bizarre behavior.^58,63-66 These symptoms may represent a relapse of a primary thought disorder or, for some unfortunate individuals, a de novo psychotic illness.^58,65,66 Symptoms can linger for months after drug use.⁶⁵

A key risk in the use of synthetic cannabinoid moieties may involve the absence of cannabidiol. Cannabidiol naturally occurs in many strains of cannabis and is thought to have antipsychotic, neuroprotective properties.⁶⁷ The absence of this molecule in synthetic cannabinoids may at least partially explain their severe psychoactive effects. Treatment for synthetic cannabinoid intoxication and related psychosis is largely supportive and may include the use of antipsychotic medication.⁶⁶

Detection of synthetic cannabinoids in urine is difficult, yet many compounds can be detected via GC-MS or LC-MS. Molecules of significance include JWH-018, JWH-073, JWH-015, JWH-250, CP-47 497, HU-210, cannabicyclohexanol, and oleamide; however, these compounds are rarely excreted in urine in their pure form. The many hydroxylated or dealkylated metabolites of these compounds, mostly unnamed, are more consistently detected in urine.^68,69 One author has noted that the pentanoic acid metabolite of JWH-018 seems to appear most reliably in urine specimens.⁶⁸

Many synthetic cannabinoid herbal mixes also contain a detectable compound called tocopherol, seemingly added as an antioxidant.^69,70 Synthetic cannabinoids are an evolving drug class, and reliable detection will require that laboratories stay up-to-date in their detection methods. As stated earlier, a commercial laboratory in the region accepted the civilian and veteran patient samples for these case studies. The synthetic cannabinoid panels offered evaluation of the drug itself (GC-MS), an oral fluid screen (LC-MS), and isolation of metabolites in urine (enzyme-linked immunosorbent assay).

Conclusion

Designer drugs will remain a challenge for providers caring for veterans for several key reasons: (1) Veterans are a vulnerable population who abuse substances at higher rates than do their civilian counterparts; (2) Chemists are able to manufacture variations of known habit-forming substances; (3) Modern technology facilitates the purchase and wide distribution of addictive substances; (4) Many designer drugs are deceptively packaged and marketed; (5) The effects of the drugs are often severe; (6) No standardized treatment guidelines exist; and (7) Detection of the drugs is difficult, and new versions of the molecules may evade even cutting-edge techniques.

Due to the high cost of detecting synthetic cathinones and synthetic cannabinoids in body fluids, screening should be considered only in settings where severe symptoms are accompanied by reasonable clinical suspicion of use and an otherwise negative toxicologic workup. As more designer drugs inevitably emerge, research will be needed on their pharmacology, toxidromes, and detection. Military and civilian practitioners must remain abreast of the dynamic trends in designer drugs to ensure that their patients receive the highest level of medical care.

Author disclosures
The authors report no actual or potential conflicts of interest with regard to this article.  

Disclaimer
The opinions expressed herein are those of the authors and do not necessarily reflect those of Federal Practitioner, Frontline Medical Communications Inc., the U.S. Government, or any of its agencies. This article may discuss unlabeled or investigational use of certain drugs. Please review complete prescribing information for specific drugs or drug combinations—including indications, contraindications, warnings, and adverse effects—before administering pharmacologic therapy to patients.

Although the elevated risks and rates of veterans’ substance abuse patterns are well documented, little has been written about veterans’ use of designer drugs.^1-6 In recent months throughout Europe and the U.S., there has been a flurry of media attention for 2 classes of designer drugs: synthetic cathinones and synthetic cannabinoids.^7,8 In the U.S., the popularity of these drugs has surged, and a disproportionate amount of use of these 2 drug classes is coming from locations near military instillations.^9,10

The purpose of this article is to raise awareness regarding these 2 burgeoning designer drug classes and their impact on veterans. Designer drugs affecting vulnerable populations are not a new phenomenon, yet many providers are unfamiliar with the effects of these unique drugs of abuse on their veteran populations.^11-13

Many designer drugs begin their existence as variations of other addictive or psychoactive drugs. Others begin in laboratories as investigative research compounds that end up on the street, often promising a novel mind-altering experience as a “legal high.”^14-18 The Designer Drug Enforcement Act of 1986 was an initial attempt in the U.S. to define and control the early rise of copycat drugs that appeared on the streets and mimicked the effects of other illicit substances. More recent legislation enacted in the U.S. has imposed Schedule I controls on the manufacture, distribution, possession, importation, and exportation of these types of drugs, including both synthetic cathinones and synthetic cannabinoids. State laws are perennially in flux trying to keep up with the latest drug trends.^19-21

Similar efforts have been made by the European Union to control mephedrone, a synthetic cathinone, citing multiple fatalities, seizures, related crime, lack of medical use, and risk of dependence.²² Although uniform levels of control do not exist in Europe for synthetic cannabinoids, many countries have independently acted to limit their use.²³

In its recent World Drug Report 2013, the United Nations Office on Drugs and Crime documents its growing concern about the “new psychoactive substances” category of illicit recreational substances (in which synthetic cannabinoids and cathinones are included) that has increased by 50% since 2009.²⁴ Alone, this category now outnumbers the total number of substances controlled by international drug conventions.

The novelty and variability of designer drugs causes difficulties with detection and regulation. Innovative chemists can legally manufacture new versions of known molecules intended for illicit use with a rapidity that outpaces bureaucratic control. Local law enforcement officials may be unaware of the latest designer drug trends, stifling efforts at public education or restriction. Designer drugs are often deceptively packaged and are available in convenience stores, tobacco outlets, gas stations, pawnshops, tattoo parlors, and truck stops.^25-28 The Internet may be the singular reason, however, that designer drugs continue to be widely available to veterans.^11,18

Innumerable websites discuss, promote, and sell designer drugs or deceitfully market them as safe, legitimate household products (“not for human consumption”), which can be ordered online and shipped by commercial carriers.¹² Little accurate information is known about their effects or about the specific compounds they contain. When the recreational nature of the drugs is actually acknowledged, information on how the buyer can evade prosecution is often provided in tandem. The suppliers’ inventory of the drugs has been shown to be variable and inconsistent, and the product ingredients can be similarly unpredictable despite comparatively more stable naming and labeling.^14,29

In the clinical setting, a reliable patient drug history may not be available. This ensures that the diagnosis of designer drug use will be an exclusionary process involving routine laboratory work, physical examination, and at times electroencephalogram and/or neuroimaging. Psychiatric consultation is often useful in this setting. Routine immunoassay tests do not detect either synthetic cathinones or synthetic cannabinoids.³⁰

Both cannabinoids and cathinones can be identified using gas chromatography-mass spectroscopy (GC-MS) or liquid chromatography-mass spectroscopy (LC-MS). However, this technology is limited to specialized laboratories.^31,32 The laboratory results often are not immediately available, potentially limiting the tests’ use in emergency or inpatient settings, as the patient may have left the hospital by the time the results are available. Additionally, these drugs’ prevalence of use, while increasing, often does not justify the cost of these tests.

The inability to routinely detect metabolites in urine may increase the enticement of these drugs given the likelihood that active-duty personnel could use them surreptitiously. Further, these compounds are evolving and seemingly limitless in their variability, and there is often a paucity of pure reference materials. As such, it is impossible to guarantee reliable test results.

The following profiles of each of these drug classes will be accompanied by clinical cases depicting the drugs’ effects and how an affected veteran might present clinically. The severe effects of these novel agents illustrate the value in maintaining a functional knowledge base about emerging drug trends. The accuracy of diagnosis as well as the outcome of a veteran’s treatment may depend on the provider’s ability to identify the presence of a drug and manage its effects.

Synthetic Cathinones

Mr. H is a 28-year-old Iraq War veteran with a history of posttraumatic stress disorder (PTSD), alcohol abuse, and opioid dependence who presented for inpatient psychiatric admission after making suicidal statements to his wife in the context of 2 weeks of “bath salts” use. A family member initially introduced him to the drug. His first drug purchase had been 1 gram ($30) at a local movie rental store.

After discharge from the hospital, Mr. H began purchasing increasing amounts online with a credit card. Although he initially had been insufflating and inhaling the substance, he later began injecting it (dissolving it in tap water and loading it through a cotton filter in a syringe). The patient admitted to finding the drug significantly more addictive than any others he had used, and his use resulted in leaving his job and abandoning his family.

Severe cravings and depression were present between episodes of use. He spent $40,000 over 6 months of use. Insomnia lasted for several days, his appearance changed dramatically (including persistent skin infections), and he became paranoid, believing that everyone around him was an undercover police officer. He remained on medications for persistent anxiety. His daily drug cravings continued,
although he remained uncertain about the actual ingredients of bath salts.

Cathinone is a naturally occurring stimulant from the khat plant (Catha edulis), which grows indigenously in Egypt and on the Arabian Peninsula. The recreational and religious use of this plant has occurred for thousands of years, though it is not without risk: The chewing of the leaves containing natural cathinones has been associated with esophagitis, gastritis, oral keratosis, myocardial infarction, dilated cardiomyopathy, hypertension, cerebral ischemia, thromboembolism, diabetes, sexual dysfunction, duodenal ulcer, and hepatitis.^33,34

The stimulants known colloquially as bath salts are synthetic cathinones, which have become more widely available within the past 10 years: first in the Middle East, then Europe, and now in the U.S.^{5,9,10,14,19,25,35-41,44} Although the current rise in use has occurred in the past few years, the first documented abuse of synthetic cathinones in the U.S. dates to the early 1990s in Michigan.⁴²

Bath salts is the most common of the many names used to denote synthetic cathinones. The compounds have no utility when used as such but often are marketed as research chemicals, plant fertilizer, or shoe polish. It is this deliberate counterfeit of household product names that allows many distributors to avoid classifying the compounds according to the true, intended use. More appealing brand names may also be used to entice the user (Table 1).²⁵

Synthetic cathinones owe their popularity to similarities with cocaine and methamphetamine. They are sympathomimetic with synaptic increases of monoamines after use: Surges in norepinephrine and dopamine account for the stimulant qualities, and serotonergic changes mediate distinct psychoactive effects (Table 2).⁴⁰ Users are interested in the drugs for many of the same reasons that other recreational stimulants have appeal: euphoria, energy, empathy, heightened sexuality, sociability, and an overall intensification of senses. Synthetic cathinones have become preferred to cocaine for some users.⁴³

The drugs can be used via oral and anal routes. Using methods known as “bombing” or “keystering,” users deliver boluses of the powder wrapped in cigarette paper, which they swallow or insert into the rectum. Insufflation and IV injection are also common methods of administration with a quicker onset of action expected.⁴⁰ The prices of the drugs range from $25 to $50 per 500-mg packet (though the cost is increasing with more regulation). Users typically use 500 mg to 2 g in one session.

The 2 most commonly abused synthetic cathinones are mephedrone and MDPV (methylenedioxypyrovalerone). There is some regional variability about which ingredient is present; mephedrone tends to be more prevalent in Europe, whereas MDPV is noted to be more common in the U.S.^10,44

When ordering a laboratory test to evaluate for the presence of these drugs, a specific request should be given to the technicians to look for signals of MDPV (most common metabolite is dimethylenyl-methyl-MDPV), mephedrone (4-methylmethcathinone), 3-bromomethcathinone (3-BMC), or 3-fluoromethcathinone (fluphedrone).^45-48 The study testing (both in VA and civilian settings) for Mr. H was done by a commercial laboratory several states away where patented techniques can screen for more than 30 compounds via LC-MS. The laboratory offered bath salts panels for urine, serum/plasma, and blood samples.

Synthetic cathinones are dangerous, and as the body of medical literature continues to expand, reports of significant morbidity and death related to their use are appearing. The harmful effects of recreational synthetic cathinone use has been documented across the globe in the form of serotonin syndrome, intoxication delirium, hyperthermia and multi-organ failure, myocarditis, hypo-osmotic hyponatremia with encephalopathy, agitation, psychosis, and death after cardiac arrest.^{5,12,38,39,49-53} Published treatment methods are largely supportive with the available literature, suggesting that benzodiazepines, antipsychotics (both typical and atypical), restraints to maintain safety, and IV fluids may be indicated.^5,9,50

Synthetic Cannabinoids

Mr. W is a 58-year-old veteran with a history of alcohol dependence and PTSD who reported use of the synthetic cannabinoid “Spice” during intake assessment for treatment of alcohol dependence. He reported using Spice about 4 times over a 2-month period. He purchased a small jar of the substance from a party store for $15 per gram and understood its contents to be synthetic marijuana, which he appreciated for its low cost and assumed legality. He denied having any understanding of the package’s contents beyond “synthetic marijuana.”

The patient ingested the drug by smoking and inhaling from a pipe. For the first 3 times that he used the substance, Mr. W reported feeling a pleasant sensation that started quickly and lasted about 30 minutes. The fourth time that he used synthetic cannabis he felt nauseated and vomited several times, had auditory hallucinations, and increased anxiety; he also reported a hangover effect after this use. He identified that the effects may have been different the fourth time “because the brands were changing.”

Mr. W also reported that his neighbor—a daily user of synthetic cannabinoids for several months—became paranoid, suspicious, and developed incomprehensible speech. His neighbor’s symptoms and his own unpleasant experiences prompted a discontinuation of use.

Synthetic cannabinoids are a diverse group of agents numbering in excess of 100 artificial compounds that act as agonists at cannabinoid receptors, mimicking the effects of tetrahydrocannabinol (THC), an active ingredient in marijuana.^28,54 The availability of these drugs online and in specialty shops has been documented since the mid-2000s.^27,28,32 Their packaging often describes the contents as incense or herbal blends, using various names. Spice is a common name, but these products are also known by a myriad of other designations (Table 1).²⁸ A single packet usually contains several grams of the drug and costs about $30.⁵⁵

To the user, who may already be familiar with marijuana, the contents intentionally appear similar to the dried buds of cannabis.^30,56 In reality, the drug has just been sprayed onto inert plant material.⁵⁷ The drug is smoked, and the psychoactive dose can be as little as 1 mg.³⁰ Users describe potent drug effects (Table 2). There is a rapid onset of action, and duration of effects last 1 to 2 hours.⁵⁸

The compounds’ mechanism of action and appeal are derived from their high affinity for the cannabinoid receptors. The CB1 receptor is located primarily in the central nervous system and is responsible for the psychoactive component of the drugs’ actions.^27,30,58,59 Two particular synthetic cannabinoids, cannibicyclohexanol and JWH-018, are potent cannabinoid CB1 agonists with affinity exceeding their natural counterparts.^{27,30, 32,56,58,59}

Chemically, these drugs are varied. The largest structural family of these compounds is the JWH group, which includes JWH-018.⁶⁰ Also common are CP 47,497 and other CP compounds.⁵⁸ HU compounds, such as HU-210, have also been identified and have been shown to be 100 to 800 times more potent at the CB1 and CB2 receptors than is THC.^60,61 A final group includes the benzoylindoles, such as AM-964 and RCS-4, which also bind strongly to CB1 and CB2.^60,62

Constitutional symptoms of synthetic cannabinoid intoxication include disorientation, anxiety, tremulousness, palpitations, tachycardia, agitation, injected conjunctivae, hyperreflexia, nausea, vomiting, lateral gaze nystagmus, and myoclonic jerks, which have been mistaken for seizure activity.^27,30,55 Pupils are often normal sized.⁵⁵ Withdrawal phenomena are similar to those of cannabis withdrawal: irritability, anxiety, tremor, palpitations, diaphoresis, insomnia, headache, diarrhea, nausea, and vomiting.⁵⁹

Given the established link between cannabis use and psychosis, synthetic cannabinoids may stand as a precipitant of psychotic symptoms, which may include visual hallucinations, auditory hallucinations, disorganized speech, paranoia, grandiose delusions, disorganization, or bizarre behavior.^58,63-66 These symptoms may represent a relapse of a primary thought disorder or, for some unfortunate individuals, a de novo psychotic illness.^58,65,66 Symptoms can linger for months after drug use.⁶⁵

A key risk in the use of synthetic cannabinoid moieties may involve the absence of cannabidiol. Cannabidiol naturally occurs in many strains of cannabis and is thought to have antipsychotic, neuroprotective properties.⁶⁷ The absence of this molecule in synthetic cannabinoids may at least partially explain their severe psychoactive effects. Treatment for synthetic cannabinoid intoxication and related psychosis is largely supportive and may include the use of antipsychotic medication.⁶⁶

Detection of synthetic cannabinoids in urine is difficult, yet many compounds can be detected via GC-MS or LC-MS. Molecules of significance include JWH-018, JWH-073, JWH-015, JWH-250, CP-47 497, HU-210, cannabicyclohexanol, and oleamide; however, these compounds are rarely excreted in urine in their pure form. The many hydroxylated or dealkylated metabolites of these compounds, mostly unnamed, are more consistently detected in urine.^68,69 One author has noted that the pentanoic acid metabolite of JWH-018 seems to appear most reliably in urine specimens.⁶⁸

Many synthetic cannabinoid herbal mixes also contain a detectable compound called tocopherol, seemingly added as an antioxidant.^69,70 Synthetic cannabinoids are an evolving drug class, and reliable detection will require that laboratories stay up-to-date in their detection methods. As stated earlier, a commercial laboratory in the region accepted the civilian and veteran patient samples for these case studies. The synthetic cannabinoid panels offered evaluation of the drug itself (GC-MS), an oral fluid screen (LC-MS), and isolation of metabolites in urine (enzyme-linked immunosorbent assay).

Conclusion

Designer drugs will remain a challenge for providers caring for veterans for several key reasons: (1) Veterans are a vulnerable population who abuse substances at higher rates than do their civilian counterparts; (2) Chemists are able to manufacture variations of known habit-forming substances; (3) Modern technology facilitates the purchase and wide distribution of addictive substances; (4) Many designer drugs are deceptively packaged and marketed; (5) The effects of the drugs are often severe; (6) No standardized treatment guidelines exist; and (7) Detection of the drugs is difficult, and new versions of the molecules may evade even cutting-edge techniques.

Due to the high cost of detecting synthetic cathinones and synthetic cannabinoids in body fluids, screening should be considered only in settings where severe symptoms are accompanied by reasonable clinical suspicion of use and an otherwise negative toxicologic workup. As more designer drugs inevitably emerge, research will be needed on their pharmacology, toxidromes, and detection. Military and civilian practitioners must remain abreast of the dynamic trends in designer drugs to ensure that their patients receive the highest level of medical care.

Author disclosures
The authors report no actual or potential conflicts of interest with regard to this article.  

Disclaimer
The opinions expressed herein are those of the authors and do not necessarily reflect those of Federal Practitioner, Frontline Medical Communications Inc., the U.S. Government, or any of its agencies. This article may discuss unlabeled or investigational use of certain drugs. Please review complete prescribing information for specific drugs or drug combinations—including indications, contraindications, warnings, and adverse effects—before administering pharmacologic therapy to patients.

The older adult population in the VA is growing. Adults aged > 85 years are the fastest growing segment of the older veteran population and many are afflicted with multiple medical problems and functional impairments.^1,2 The majority of older veterans (94.6%, or about 1.9 million veterans) who seek care at the VA obtain care through primary care providers (PCPs) who are often not formally trained in geriatrics.^1,3 With the increasing number of older patients, new models of care are needed to provide coordinated, comprehensive, efficient, and patient-centered care.^4,5

Common themes found in successful models of care for older patients include a team approach, care management (comprehensive and coordinated), and patients who are active partners.⁴ These themes are reflected in the VA Patient Aligned Care Team (PACT) primary care program. PACT, a model of care that was initiated in 2010 and is built on a foundation of patient-centered care, encompasses a team approach to provide comprehensive, coordinated, and personalized care.^6-8 The challenge for the VA is to integrate gerontologic principles and tools into the daily practice of all PACTs in order to improve care provided for older veterans.⁹

This article discusses current challenges in caring for older veterans in the VA system and recommends tools that can be used to infuse geriatric care principles into VA primary care by the PACT, to improve the quality of care provided to older veterans. In addition, the article also describes VA geriatric programs that PACT clinicians can access to supplement older veterans’ care.

Challenges of Caring for Older Veterans

One concern when caring for older veterans arises when the veteran accesses both VA and non-VA health care services to offset medication costs and obtain services not covered by Medicare or other insurance companies.^2,3 This “dual care” can exacerbate polypharmacy issues and increase confusion regarding plans of care. Problems may arise when multiple providers from different systems of care prescribe medications available only within their own formulary and/or order diagnostic and laboratory tests with results available only within their own health care system.

The VA is also challenged by health care delivery for rural veterans. Thirty-six percent of all veterans live in rural areas, and they often depend on non-VA services to meet their health care needs due to difficulty traveling to the nearest VA facility.¹⁰ Seasonal residency also presents challenges. An increasing number of older veterans are seen at different VA facilities when they “winter” in a different section of the country.

Fortunately, a VA provider in one facility can access a patient’s electronic medical records in another facility, using the VA Computerized Patient Record System (CPRS). However, it is unclear to what extent busy VA PCPs use this function when seeing patients. Although individual pilot programs have shown promise, integrated electronic health records between VA and non-VA health care have not advanced to the point of sharing data or reconciling care plans (R. Rupper, personal communication, March 1, 2013).

Many PCPs and other PACT staff are not formally trained in geriatrics and may have had limited exposure to geriatric principles.³ Clinic time pressures, multiple clinical reminders (eg, vaccinations), and panel management of specific diseases make it challenging to find time to focus on complex geriatric syndromes. Current PACT performance measures also do not routinely include geriatric-specific quality of care criteria or focus on patient function (K. Shay, personal communication, February 12, 2013), a hallmark of geriatric care.⁸ Furthermore, with increasing complexity of the health care system and limited availability of resources, it is often time consuming to identify and collaborate with non-VA resources to ensure patients’ needs are met in their communities.

Opportunities for Improvement in Care

The VA transformation to PACTs has led to process changes in clinic workflow that may aid in addressing the aforementioned challenges in caring for older veterans. Each patient is assigned to a PCP-led team that includes a registered nurse care manager, a clinical associate, and an administrative associate. The PACT model of care has increased access to care by redesigning face-to-face visits, increasingly moving toward open access, and through the increased use of virtual access via secured e-mail, telephone visits, and telehealth.⁸

In addition to process changes, the VA has created new tools to assist teams in patient management. One of these is the Care Assessment Need (CAN) score, a risk stratification tool available for use by PACTs to identify patients at highest risk for hospital admission and/or death for focused care management.¹¹ It is based on statistical prediction models of veterans enrolled in primary care, using patient characteristics and health care use information.¹¹ Although the CAN score looks promising, more research is needed to evaluate its effectiveness  in improving care for older veterans and its association with better patient functioning—an important focus in quality geriatric care.

A tool that takes into account daily function is the Vulnerable Elders Survey-13 (VES-13). As measured by the VES-13, functional ability has been shown to be a strong predictor of decline and death in older adults independent of gender or comorbidities.¹² Integration of the VES-13 into the evaluation of older veterans could assist PACTs in considering patients’ current function and life expectancy in their care plans along with patient and family goals.

Another potentially useful tool for the PACT team is the SPICES mnemonic (Sleeping, Problems with feeding/eating, Incontinence or urinary problems, Confusion, Evidence of falls, and Skin breakdown).¹³ Although SPICES is not comprehensive, this mnemonic highlights potential problems facing older patients that may not be brought up routinely. It provides a concise, formalized format that can be used by clerks or patient support assistants  as part of the check-in process.

This tool has been used successfully by the Geriatric Evaluation and Management Clinic of the South Texas Veterans Health Care System (STVHCS) to improve communication between the PCP and nurse so that pertinent patient information is relayed concisely. SPICES was helpful in identifying patients needing interventions for fall risk. In a retrospective chart review of 100 randomly selected patients aged 75 to 90 years enrolled in the clinic, a 75% reduction in falls was noted during the first year of implementation (STVHCS unpublished data, 2012).

Additional tools that focus on identifying specific geriatric syndromes are available online from the Hartford Institute for Geriatric Nursing, which provides evidence-based information and training on how to assess, evaluate, and manage common geriatric syndromes such as depression, dementia, and delirium.¹⁴ The site also includes videos on how to use common brief geriatric assessment tools that can be performed by nurses and health care associates while the patient is in the waiting room. Though promising, further research is needed to study the effects of these tools on patient, provider, and system outcomes.

Infusing quality of care indicators (QI) can play an integral role in achieving PACT goals while improving the older veterans’ quality of life. For example, polypharmacy and medication-related injuries in older adults continue to pose both a safety and economic challenge to patients and the health care system.^15-17 The 2012 Beers criteria for Potentially Inappropriate Medications in Older Adults lists 53 medication classes that have been identified as potentially inappropriate medications for use in older adults.¹⁷ Use of this tool by PACTs in the development of patient care plans has the potential to reduce medication-related adverse reactions and improper prescribing.^18,19

Assessing Care of Vulnerable Elders (ACOVE ) also provides QIs that are specific to vulnerable older persons.^20-24 The most recent version, ACOVE-3, includes 392 QIs for 26 conditions and 14 types of care processes and covers all domains of care.²⁰ Findings from a study applying QIs involving vulnerable elderly patients in 2 managed care programs revealed that recipients of better-quality care had a 10% higher survival rate over 3 years.²⁵

The VA currently monitors 6 frail elderly QIs based on ACOVE criteria via reviews of medical records in veterans aged > 75 years. These QIs cover falls, incontinence, functional assessment, and the presence of a surrogate decision maker. PACT staff, unfortunately, do not receive feedback on these, because they are still QIs and not part of the performance measures (K. Shay, personal communication, February 12, 2013). Though some VA sites have adopted these QIs to some extent, until these frail elderly QIs become performance measures throughout VA, other competing priorities may be more at the forefront of quality improvement projects done by PACT teams.⁵

The American Geriatrics Society recently published recommendations on the care of older adults with multiple chronic conditions, to aid PCPs in practicing a more individualized, patient-centered care in complex cases.²⁶ In addition to focusing on a patient’s primary concern during a clinic visit and eliciting preferences, considering prognosis in deciding on treatment options allows patients to better weigh the potential benefits and burdens in their daily living.²⁶ A discussion on how aggressive potential treatments are and what the patient is willing to undertake is an important component of patient-centered care and should be incorporated during routine PACT clinic visits.

VA Geriatric Programs

It is important for PACT clinicians to be familiar with the geriatric programs and resources available within the VA medical home “neighborhood,” which can supplement care. One such resource is the Geriatric Research Education and Clinical Centers (GRECCs). There are currently 19 GRECCs throughout the nation that serve as Centers of Excellence in the care of older veterans.²⁷ The GRECCs provide training for clinicians, test innovative ways to care for older veterans, and collaborate with other staff to improve the care provided. Some have also developed Geriatric Primary Care Clinics (or Geri PACTs) to provide team care to very frail and high-risk older veterans. Since not all VA facilities have access to Geri PACTs, the GRECCs play an important role in making geriatric expertise and training available to the PACTs.³

To address this limitation in access, VA programs have begun using telehealth technology to increase competencies of PCPs in caring for older veterans. For example, the VA Geriatric Scholars Program is a national educational program with different avenues to “geriatricize” VA primary care services and improve knowledge and care provided to older veterans.²⁸ It consists of several subprograms: Geriatric Scholars Program for Rural Community Based Outpatient Clinics; Geriatric Scholars Program for Primary Care Providers; Rural Interdisciplinary Team Training; and the Geriatric Assessment Pocket Guide.²⁹ These components may include didactics both face-to-face and online, clinical experience with performing common geriatric screening tools, and a quality improvement project.

Some local VAMCs have also developed programs to address this need to improve care provided to older veterans in PACT. The VA Greater Los Angeles Healthcare System (GLA) GRECC, for example, has started several programs to infuse geriatrics into PACTs, including the Geri Specialty Care Access Network-Extension for Community Healthcare Outcomes (SCAN-ECHO). VA SCAN-ECHO was developed to increase access to specialty care in rural/underserved areas. The PCP presents a case and a specialty provider gives guidance in the assessment and/or management of a specific clinical problem.³⁰ Unlike many other SCAN-ECHO programs, the GLA Geri SCAN-ECHO program encourages not only PCPs, but also nurses and social workers to submit consults for discussion and encourages team management (a hallmark of quality geriatric care). Another important GLA GRECC project is the Veterans Cognitive Assessment and Management Program (V-CAMP), which uses videoconferencing to assess and manage veterans with cognitive impairment/dementia who reside in underserved areas in the GLA region. The program provides dementia care management and access to neuropsychological examinations—services that are often not available in rural areas.³¹

Various VA program offices have also published useful resources to help PACT clinicians infuse gerontologic principles into their practice. The VA Office of Nursing Services has a Geriatrics and Extended Care Field Advisory Committee, which recently produced on-demand lectures in the virtual VA eHealth University (also known as myVeHU campus) on improving the PACT’s management of progressive chronic diseases and dementia recognition and initial evaluation. They also produced a resource guide for VA clinicians (nursing and non-nursing), based on a team consensus of what the workgroup thinks a clinician would find helpful in clinical practice to improve care of older veterans. The VA Office of Geriatrics and Extended Care Service also identified a list of clinical and educational resources to help PACT clinicians. These include the Geriatrics Evaluation and management (GEM) Tools Booklet (http://geriatricscareonline.org) and a SharePoint site to improve dementia care in all settings.

The VA Office of Geriatrics and Extended Care provides additional geriatric-specific programs (http://va.gov/geriatrics). These programs may be useful for consultation and collaboration for patients whom the PACT teams have found to be more challenging and require more assistance to meet performance measures and patient needs. A recent evidence synthesis notes that direct involvement of geriatricians (as opposed to indirect care with limited contact) is more likely to result in positive patient outcomes and should be considered for those patients who are the most frail and/or high utilizers of services.³²

Conclusion

The PACT initiative in the VA health care system may prove to be an important vehicle for improving and standardizing the care provided to older veterans. Use of reliable and valid tools in the identification and assessment of geriatric syndromes, provision of quality standards, and use of innovative telehealth practices are promising enhancements for the primary care of older veterans.

Acknowledgements
We would like to thank the following contributors for their thoughtful review of the initial drafts of this article: Dr. Balmatee Bidassie; Dr. Kathryn Corrigan; Dr. Gail McNut; Dr. Linda Kinsinger, chief consultant for preventive medicine in the Office of Patient Care Services; Dr. Theodore Hahn, GRECC deputy director from VA Greater Los Angeles Healthcare System; Dr. James Hallenbeck, associate chief of staff, Extended Care at VA Palo Alto Health Care System; Ms. Storm Morgan, VA Office of Nursing Services PACT program manager; and Dr. Kenneth Shay, director of Geriatric Programs for the VA Office of Geriatrics and Extended Care.

The authors also would like to express their gratitude to the VA Office of Nursing Services, Clinical Practice Program, Geriatrics and Extended Care Field Advisory Committee for the opportunity to work on this manuscript.

Author disclosures
The authors report no actual or potential conflicts of interest with regard to this article.

Disclaimer
The opinions expressed herein are those of the authors and do not necessarily reflect those of Federal Practitioner, Frontline Medical Communications Inc., the U.S. Government, or any of its agencies. This article may discuss unlabeled or investigational use of certain drugs. Please review the complete prescribing information for specific drugs or drug combinations—including indications, contraindications, warnings, and adverse effects—before administering pharmacologic therapy to patients.

The older adult population in the VA is growing. Adults aged > 85 years are the fastest growing segment of the older veteran population and many are afflicted with multiple medical problems and functional impairments.^1,2 The majority of older veterans (94.6%, or about 1.9 million veterans) who seek care at the VA obtain care through primary care providers (PCPs) who are often not formally trained in geriatrics.^1,3 With the increasing number of older patients, new models of care are needed to provide coordinated, comprehensive, efficient, and patient-centered care.^4,5

Common themes found in successful models of care for older patients include a team approach, care management (comprehensive and coordinated), and patients who are active partners.⁴ These themes are reflected in the VA Patient Aligned Care Team (PACT) primary care program. PACT, a model of care that was initiated in 2010 and is built on a foundation of patient-centered care, encompasses a team approach to provide comprehensive, coordinated, and personalized care.^6-8 The challenge for the VA is to integrate gerontologic principles and tools into the daily practice of all PACTs in order to improve care provided for older veterans.⁹

This article discusses current challenges in caring for older veterans in the VA system and recommends tools that can be used to infuse geriatric care principles into VA primary care by the PACT, to improve the quality of care provided to older veterans. In addition, the article also describes VA geriatric programs that PACT clinicians can access to supplement older veterans’ care.

Challenges of Caring for Older Veterans

One concern when caring for older veterans arises when the veteran accesses both VA and non-VA health care services to offset medication costs and obtain services not covered by Medicare or other insurance companies.^2,3 This “dual care” can exacerbate polypharmacy issues and increase confusion regarding plans of care. Problems may arise when multiple providers from different systems of care prescribe medications available only within their own formulary and/or order diagnostic and laboratory tests with results available only within their own health care system.

The VA is also challenged by health care delivery for rural veterans. Thirty-six percent of all veterans live in rural areas, and they often depend on non-VA services to meet their health care needs due to difficulty traveling to the nearest VA facility.¹⁰ Seasonal residency also presents challenges. An increasing number of older veterans are seen at different VA facilities when they “winter” in a different section of the country.

Fortunately, a VA provider in one facility can access a patient’s electronic medical records in another facility, using the VA Computerized Patient Record System (CPRS). However, it is unclear to what extent busy VA PCPs use this function when seeing patients. Although individual pilot programs have shown promise, integrated electronic health records between VA and non-VA health care have not advanced to the point of sharing data or reconciling care plans (R. Rupper, personal communication, March 1, 2013).

Many PCPs and other PACT staff are not formally trained in geriatrics and may have had limited exposure to geriatric principles.³ Clinic time pressures, multiple clinical reminders (eg, vaccinations), and panel management of specific diseases make it challenging to find time to focus on complex geriatric syndromes. Current PACT performance measures also do not routinely include geriatric-specific quality of care criteria or focus on patient function (K. Shay, personal communication, February 12, 2013), a hallmark of geriatric care.⁸ Furthermore, with increasing complexity of the health care system and limited availability of resources, it is often time consuming to identify and collaborate with non-VA resources to ensure patients’ needs are met in their communities.

Opportunities for Improvement in Care

The VA transformation to PACTs has led to process changes in clinic workflow that may aid in addressing the aforementioned challenges in caring for older veterans. Each patient is assigned to a PCP-led team that includes a registered nurse care manager, a clinical associate, and an administrative associate. The PACT model of care has increased access to care by redesigning face-to-face visits, increasingly moving toward open access, and through the increased use of virtual access via secured e-mail, telephone visits, and telehealth.⁸

In addition to process changes, the VA has created new tools to assist teams in patient management. One of these is the Care Assessment Need (CAN) score, a risk stratification tool available for use by PACTs to identify patients at highest risk for hospital admission and/or death for focused care management.¹¹ It is based on statistical prediction models of veterans enrolled in primary care, using patient characteristics and health care use information.¹¹ Although the CAN score looks promising, more research is needed to evaluate its effectiveness  in improving care for older veterans and its association with better patient functioning—an important focus in quality geriatric care.

A tool that takes into account daily function is the Vulnerable Elders Survey-13 (VES-13). As measured by the VES-13, functional ability has been shown to be a strong predictor of decline and death in older adults independent of gender or comorbidities.¹² Integration of the VES-13 into the evaluation of older veterans could assist PACTs in considering patients’ current function and life expectancy in their care plans along with patient and family goals.

Another potentially useful tool for the PACT team is the SPICES mnemonic (Sleeping, Problems with feeding/eating, Incontinence or urinary problems, Confusion, Evidence of falls, and Skin breakdown).¹³ Although SPICES is not comprehensive, this mnemonic highlights potential problems facing older patients that may not be brought up routinely. It provides a concise, formalized format that can be used by clerks or patient support assistants  as part of the check-in process.

This tool has been used successfully by the Geriatric Evaluation and Management Clinic of the South Texas Veterans Health Care System (STVHCS) to improve communication between the PCP and nurse so that pertinent patient information is relayed concisely. SPICES was helpful in identifying patients needing interventions for fall risk. In a retrospective chart review of 100 randomly selected patients aged 75 to 90 years enrolled in the clinic, a 75% reduction in falls was noted during the first year of implementation (STVHCS unpublished data, 2012).

Additional tools that focus on identifying specific geriatric syndromes are available online from the Hartford Institute for Geriatric Nursing, which provides evidence-based information and training on how to assess, evaluate, and manage common geriatric syndromes such as depression, dementia, and delirium.¹⁴ The site also includes videos on how to use common brief geriatric assessment tools that can be performed by nurses and health care associates while the patient is in the waiting room. Though promising, further research is needed to study the effects of these tools on patient, provider, and system outcomes.

Infusing quality of care indicators (QI) can play an integral role in achieving PACT goals while improving the older veterans’ quality of life. For example, polypharmacy and medication-related injuries in older adults continue to pose both a safety and economic challenge to patients and the health care system.^15-17 The 2012 Beers criteria for Potentially Inappropriate Medications in Older Adults lists 53 medication classes that have been identified as potentially inappropriate medications for use in older adults.¹⁷ Use of this tool by PACTs in the development of patient care plans has the potential to reduce medication-related adverse reactions and improper prescribing.^18,19

Assessing Care of Vulnerable Elders (ACOVE ) also provides QIs that are specific to vulnerable older persons.^20-24 The most recent version, ACOVE-3, includes 392 QIs for 26 conditions and 14 types of care processes and covers all domains of care.²⁰ Findings from a study applying QIs involving vulnerable elderly patients in 2 managed care programs revealed that recipients of better-quality care had a 10% higher survival rate over 3 years.²⁵

The VA currently monitors 6 frail elderly QIs based on ACOVE criteria via reviews of medical records in veterans aged > 75 years. These QIs cover falls, incontinence, functional assessment, and the presence of a surrogate decision maker. PACT staff, unfortunately, do not receive feedback on these, because they are still QIs and not part of the performance measures (K. Shay, personal communication, February 12, 2013). Though some VA sites have adopted these QIs to some extent, until these frail elderly QIs become performance measures throughout VA, other competing priorities may be more at the forefront of quality improvement projects done by PACT teams.⁵

The American Geriatrics Society recently published recommendations on the care of older adults with multiple chronic conditions, to aid PCPs in practicing a more individualized, patient-centered care in complex cases.²⁶ In addition to focusing on a patient’s primary concern during a clinic visit and eliciting preferences, considering prognosis in deciding on treatment options allows patients to better weigh the potential benefits and burdens in their daily living.²⁶ A discussion on how aggressive potential treatments are and what the patient is willing to undertake is an important component of patient-centered care and should be incorporated during routine PACT clinic visits.

VA Geriatric Programs

It is important for PACT clinicians to be familiar with the geriatric programs and resources available within the VA medical home “neighborhood,” which can supplement care. One such resource is the Geriatric Research Education and Clinical Centers (GRECCs). There are currently 19 GRECCs throughout the nation that serve as Centers of Excellence in the care of older veterans.²⁷ The GRECCs provide training for clinicians, test innovative ways to care for older veterans, and collaborate with other staff to improve the care provided. Some have also developed Geriatric Primary Care Clinics (or Geri PACTs) to provide team care to very frail and high-risk older veterans. Since not all VA facilities have access to Geri PACTs, the GRECCs play an important role in making geriatric expertise and training available to the PACTs.³

To address this limitation in access, VA programs have begun using telehealth technology to increase competencies of PCPs in caring for older veterans. For example, the VA Geriatric Scholars Program is a national educational program with different avenues to “geriatricize” VA primary care services and improve knowledge and care provided to older veterans.²⁸ It consists of several subprograms: Geriatric Scholars Program for Rural Community Based Outpatient Clinics; Geriatric Scholars Program for Primary Care Providers; Rural Interdisciplinary Team Training; and the Geriatric Assessment Pocket Guide.²⁹ These components may include didactics both face-to-face and online, clinical experience with performing common geriatric screening tools, and a quality improvement project.

Some local VAMCs have also developed programs to address this need to improve care provided to older veterans in PACT. The VA Greater Los Angeles Healthcare System (GLA) GRECC, for example, has started several programs to infuse geriatrics into PACTs, including the Geri Specialty Care Access Network-Extension for Community Healthcare Outcomes (SCAN-ECHO). VA SCAN-ECHO was developed to increase access to specialty care in rural/underserved areas. The PCP presents a case and a specialty provider gives guidance in the assessment and/or management of a specific clinical problem.³⁰ Unlike many other SCAN-ECHO programs, the GLA Geri SCAN-ECHO program encourages not only PCPs, but also nurses and social workers to submit consults for discussion and encourages team management (a hallmark of quality geriatric care). Another important GLA GRECC project is the Veterans Cognitive Assessment and Management Program (V-CAMP), which uses videoconferencing to assess and manage veterans with cognitive impairment/dementia who reside in underserved areas in the GLA region. The program provides dementia care management and access to neuropsychological examinations—services that are often not available in rural areas.³¹

Various VA program offices have also published useful resources to help PACT clinicians infuse gerontologic principles into their practice. The VA Office of Nursing Services has a Geriatrics and Extended Care Field Advisory Committee, which recently produced on-demand lectures in the virtual VA eHealth University (also known as myVeHU campus) on improving the PACT’s management of progressive chronic diseases and dementia recognition and initial evaluation. They also produced a resource guide for VA clinicians (nursing and non-nursing), based on a team consensus of what the workgroup thinks a clinician would find helpful in clinical practice to improve care of older veterans. The VA Office of Geriatrics and Extended Care Service also identified a list of clinical and educational resources to help PACT clinicians. These include the Geriatrics Evaluation and management (GEM) Tools Booklet (http://geriatricscareonline.org) and a SharePoint site to improve dementia care in all settings.

The VA Office of Geriatrics and Extended Care provides additional geriatric-specific programs (http://va.gov/geriatrics). These programs may be useful for consultation and collaboration for patients whom the PACT teams have found to be more challenging and require more assistance to meet performance measures and patient needs. A recent evidence synthesis notes that direct involvement of geriatricians (as opposed to indirect care with limited contact) is more likely to result in positive patient outcomes and should be considered for those patients who are the most frail and/or high utilizers of services.³²

Conclusion

The PACT initiative in the VA health care system may prove to be an important vehicle for improving and standardizing the care provided to older veterans. Use of reliable and valid tools in the identification and assessment of geriatric syndromes, provision of quality standards, and use of innovative telehealth practices are promising enhancements for the primary care of older veterans.

Acknowledgements
We would like to thank the following contributors for their thoughtful review of the initial drafts of this article: Dr. Balmatee Bidassie; Dr. Kathryn Corrigan; Dr. Gail McNut; Dr. Linda Kinsinger, chief consultant for preventive medicine in the Office of Patient Care Services; Dr. Theodore Hahn, GRECC deputy director from VA Greater Los Angeles Healthcare System; Dr. James Hallenbeck, associate chief of staff, Extended Care at VA Palo Alto Health Care System; Ms. Storm Morgan, VA Office of Nursing Services PACT program manager; and Dr. Kenneth Shay, director of Geriatric Programs for the VA Office of Geriatrics and Extended Care.

The authors also would like to express their gratitude to the VA Office of Nursing Services, Clinical Practice Program, Geriatrics and Extended Care Field Advisory Committee for the opportunity to work on this manuscript.

Author disclosures
The authors report no actual or potential conflicts of interest with regard to this article.

Disclaimer
The opinions expressed herein are those of the authors and do not necessarily reflect those of Federal Practitioner, Frontline Medical Communications Inc., the U.S. Government, or any of its agencies. This article may discuss unlabeled or investigational use of certain drugs. Please review the complete prescribing information for specific drugs or drug combinations—including indications, contraindications, warnings, and adverse effects—before administering pharmacologic therapy to patients.

The older adult population in the VA is growing. Adults aged > 85 years are the fastest growing segment of the older veteran population and many are afflicted with multiple medical problems and functional impairments.^1,2 The majority of older veterans (94.6%, or about 1.9 million veterans) who seek care at the VA obtain care through primary care providers (PCPs) who are often not formally trained in geriatrics.^1,3 With the increasing number of older patients, new models of care are needed to provide coordinated, comprehensive, efficient, and patient-centered care.^4,5

Common themes found in successful models of care for older patients include a team approach, care management (comprehensive and coordinated), and patients who are active partners.⁴ These themes are reflected in the VA Patient Aligned Care Team (PACT) primary care program. PACT, a model of care that was initiated in 2010 and is built on a foundation of patient-centered care, encompasses a team approach to provide comprehensive, coordinated, and personalized care.^6-8 The challenge for the VA is to integrate gerontologic principles and tools into the daily practice of all PACTs in order to improve care provided for older veterans.⁹

This article discusses current challenges in caring for older veterans in the VA system and recommends tools that can be used to infuse geriatric care principles into VA primary care by the PACT, to improve the quality of care provided to older veterans. In addition, the article also describes VA geriatric programs that PACT clinicians can access to supplement older veterans’ care.

Challenges of Caring for Older Veterans

One concern when caring for older veterans arises when the veteran accesses both VA and non-VA health care services to offset medication costs and obtain services not covered by Medicare or other insurance companies.^2,3 This “dual care” can exacerbate polypharmacy issues and increase confusion regarding plans of care. Problems may arise when multiple providers from different systems of care prescribe medications available only within their own formulary and/or order diagnostic and laboratory tests with results available only within their own health care system.

The VA is also challenged by health care delivery for rural veterans. Thirty-six percent of all veterans live in rural areas, and they often depend on non-VA services to meet their health care needs due to difficulty traveling to the nearest VA facility.¹⁰ Seasonal residency also presents challenges. An increasing number of older veterans are seen at different VA facilities when they “winter” in a different section of the country.

Fortunately, a VA provider in one facility can access a patient’s electronic medical records in another facility, using the VA Computerized Patient Record System (CPRS). However, it is unclear to what extent busy VA PCPs use this function when seeing patients. Although individual pilot programs have shown promise, integrated electronic health records between VA and non-VA health care have not advanced to the point of sharing data or reconciling care plans (R. Rupper, personal communication, March 1, 2013).

Many PCPs and other PACT staff are not formally trained in geriatrics and may have had limited exposure to geriatric principles.³ Clinic time pressures, multiple clinical reminders (eg, vaccinations), and panel management of specific diseases make it challenging to find time to focus on complex geriatric syndromes. Current PACT performance measures also do not routinely include geriatric-specific quality of care criteria or focus on patient function (K. Shay, personal communication, February 12, 2013), a hallmark of geriatric care.⁸ Furthermore, with increasing complexity of the health care system and limited availability of resources, it is often time consuming to identify and collaborate with non-VA resources to ensure patients’ needs are met in their communities.

Opportunities for Improvement in Care

The VA transformation to PACTs has led to process changes in clinic workflow that may aid in addressing the aforementioned challenges in caring for older veterans. Each patient is assigned to a PCP-led team that includes a registered nurse care manager, a clinical associate, and an administrative associate. The PACT model of care has increased access to care by redesigning face-to-face visits, increasingly moving toward open access, and through the increased use of virtual access via secured e-mail, telephone visits, and telehealth.⁸

In addition to process changes, the VA has created new tools to assist teams in patient management. One of these is the Care Assessment Need (CAN) score, a risk stratification tool available for use by PACTs to identify patients at highest risk for hospital admission and/or death for focused care management.¹¹ It is based on statistical prediction models of veterans enrolled in primary care, using patient characteristics and health care use information.¹¹ Although the CAN score looks promising, more research is needed to evaluate its effectiveness  in improving care for older veterans and its association with better patient functioning—an important focus in quality geriatric care.

A tool that takes into account daily function is the Vulnerable Elders Survey-13 (VES-13). As measured by the VES-13, functional ability has been shown to be a strong predictor of decline and death in older adults independent of gender or comorbidities.¹² Integration of the VES-13 into the evaluation of older veterans could assist PACTs in considering patients’ current function and life expectancy in their care plans along with patient and family goals.

Another potentially useful tool for the PACT team is the SPICES mnemonic (Sleeping, Problems with feeding/eating, Incontinence or urinary problems, Confusion, Evidence of falls, and Skin breakdown).¹³ Although SPICES is not comprehensive, this mnemonic highlights potential problems facing older patients that may not be brought up routinely. It provides a concise, formalized format that can be used by clerks or patient support assistants  as part of the check-in process.

This tool has been used successfully by the Geriatric Evaluation and Management Clinic of the South Texas Veterans Health Care System (STVHCS) to improve communication between the PCP and nurse so that pertinent patient information is relayed concisely. SPICES was helpful in identifying patients needing interventions for fall risk. In a retrospective chart review of 100 randomly selected patients aged 75 to 90 years enrolled in the clinic, a 75% reduction in falls was noted during the first year of implementation (STVHCS unpublished data, 2012).

Additional tools that focus on identifying specific geriatric syndromes are available online from the Hartford Institute for Geriatric Nursing, which provides evidence-based information and training on how to assess, evaluate, and manage common geriatric syndromes such as depression, dementia, and delirium.¹⁴ The site also includes videos on how to use common brief geriatric assessment tools that can be performed by nurses and health care associates while the patient is in the waiting room. Though promising, further research is needed to study the effects of these tools on patient, provider, and system outcomes.

Infusing quality of care indicators (QI) can play an integral role in achieving PACT goals while improving the older veterans’ quality of life. For example, polypharmacy and medication-related injuries in older adults continue to pose both a safety and economic challenge to patients and the health care system.^15-17 The 2012 Beers criteria for Potentially Inappropriate Medications in Older Adults lists 53 medication classes that have been identified as potentially inappropriate medications for use in older adults.¹⁷ Use of this tool by PACTs in the development of patient care plans has the potential to reduce medication-related adverse reactions and improper prescribing.^18,19

Assessing Care of Vulnerable Elders (ACOVE ) also provides QIs that are specific to vulnerable older persons.^20-24 The most recent version, ACOVE-3, includes 392 QIs for 26 conditions and 14 types of care processes and covers all domains of care.²⁰ Findings from a study applying QIs involving vulnerable elderly patients in 2 managed care programs revealed that recipients of better-quality care had a 10% higher survival rate over 3 years.²⁵

The VA currently monitors 6 frail elderly QIs based on ACOVE criteria via reviews of medical records in veterans aged > 75 years. These QIs cover falls, incontinence, functional assessment, and the presence of a surrogate decision maker. PACT staff, unfortunately, do not receive feedback on these, because they are still QIs and not part of the performance measures (K. Shay, personal communication, February 12, 2013). Though some VA sites have adopted these QIs to some extent, until these frail elderly QIs become performance measures throughout VA, other competing priorities may be more at the forefront of quality improvement projects done by PACT teams.⁵

The American Geriatrics Society recently published recommendations on the care of older adults with multiple chronic conditions, to aid PCPs in practicing a more individualized, patient-centered care in complex cases.²⁶ In addition to focusing on a patient’s primary concern during a clinic visit and eliciting preferences, considering prognosis in deciding on treatment options allows patients to better weigh the potential benefits and burdens in their daily living.²⁶ A discussion on how aggressive potential treatments are and what the patient is willing to undertake is an important component of patient-centered care and should be incorporated during routine PACT clinic visits.

VA Geriatric Programs

It is important for PACT clinicians to be familiar with the geriatric programs and resources available within the VA medical home “neighborhood,” which can supplement care. One such resource is the Geriatric Research Education and Clinical Centers (GRECCs). There are currently 19 GRECCs throughout the nation that serve as Centers of Excellence in the care of older veterans.²⁷ The GRECCs provide training for clinicians, test innovative ways to care for older veterans, and collaborate with other staff to improve the care provided. Some have also developed Geriatric Primary Care Clinics (or Geri PACTs) to provide team care to very frail and high-risk older veterans. Since not all VA facilities have access to Geri PACTs, the GRECCs play an important role in making geriatric expertise and training available to the PACTs.³

To address this limitation in access, VA programs have begun using telehealth technology to increase competencies of PCPs in caring for older veterans. For example, the VA Geriatric Scholars Program is a national educational program with different avenues to “geriatricize” VA primary care services and improve knowledge and care provided to older veterans.²⁸ It consists of several subprograms: Geriatric Scholars Program for Rural Community Based Outpatient Clinics; Geriatric Scholars Program for Primary Care Providers; Rural Interdisciplinary Team Training; and the Geriatric Assessment Pocket Guide.²⁹ These components may include didactics both face-to-face and online, clinical experience with performing common geriatric screening tools, and a quality improvement project.

Some local VAMCs have also developed programs to address this need to improve care provided to older veterans in PACT. The VA Greater Los Angeles Healthcare System (GLA) GRECC, for example, has started several programs to infuse geriatrics into PACTs, including the Geri Specialty Care Access Network-Extension for Community Healthcare Outcomes (SCAN-ECHO). VA SCAN-ECHO was developed to increase access to specialty care in rural/underserved areas. The PCP presents a case and a specialty provider gives guidance in the assessment and/or management of a specific clinical problem.³⁰ Unlike many other SCAN-ECHO programs, the GLA Geri SCAN-ECHO program encourages not only PCPs, but also nurses and social workers to submit consults for discussion and encourages team management (a hallmark of quality geriatric care). Another important GLA GRECC project is the Veterans Cognitive Assessment and Management Program (V-CAMP), which uses videoconferencing to assess and manage veterans with cognitive impairment/dementia who reside in underserved areas in the GLA region. The program provides dementia care management and access to neuropsychological examinations—services that are often not available in rural areas.³¹

Various VA program offices have also published useful resources to help PACT clinicians infuse gerontologic principles into their practice. The VA Office of Nursing Services has a Geriatrics and Extended Care Field Advisory Committee, which recently produced on-demand lectures in the virtual VA eHealth University (also known as myVeHU campus) on improving the PACT’s management of progressive chronic diseases and dementia recognition and initial evaluation. They also produced a resource guide for VA clinicians (nursing and non-nursing), based on a team consensus of what the workgroup thinks a clinician would find helpful in clinical practice to improve care of older veterans. The VA Office of Geriatrics and Extended Care Service also identified a list of clinical and educational resources to help PACT clinicians. These include the Geriatrics Evaluation and management (GEM) Tools Booklet (http://geriatricscareonline.org) and a SharePoint site to improve dementia care in all settings.

The VA Office of Geriatrics and Extended Care provides additional geriatric-specific programs (http://va.gov/geriatrics). These programs may be useful for consultation and collaboration for patients whom the PACT teams have found to be more challenging and require more assistance to meet performance measures and patient needs. A recent evidence synthesis notes that direct involvement of geriatricians (as opposed to indirect care with limited contact) is more likely to result in positive patient outcomes and should be considered for those patients who are the most frail and/or high utilizers of services.³²

Conclusion

The PACT initiative in the VA health care system may prove to be an important vehicle for improving and standardizing the care provided to older veterans. Use of reliable and valid tools in the identification and assessment of geriatric syndromes, provision of quality standards, and use of innovative telehealth practices are promising enhancements for the primary care of older veterans.

Acknowledgements
We would like to thank the following contributors for their thoughtful review of the initial drafts of this article: Dr. Balmatee Bidassie; Dr. Kathryn Corrigan; Dr. Gail McNut; Dr. Linda Kinsinger, chief consultant for preventive medicine in the Office of Patient Care Services; Dr. Theodore Hahn, GRECC deputy director from VA Greater Los Angeles Healthcare System; Dr. James Hallenbeck, associate chief of staff, Extended Care at VA Palo Alto Health Care System; Ms. Storm Morgan, VA Office of Nursing Services PACT program manager; and Dr. Kenneth Shay, director of Geriatric Programs for the VA Office of Geriatrics and Extended Care.

The authors also would like to express their gratitude to the VA Office of Nursing Services, Clinical Practice Program, Geriatrics and Extended Care Field Advisory Committee for the opportunity to work on this manuscript.

Author disclosures
The authors report no actual or potential conflicts of interest with regard to this article.

Disclaimer
The opinions expressed herein are those of the authors and do not necessarily reflect those of Federal Practitioner, Frontline Medical Communications Inc., the U.S. Government, or any of its agencies. This article may discuss unlabeled or investigational use of certain drugs. Please review the complete prescribing information for specific drugs or drug combinations—including indications, contraindications, warnings, and adverse effects—before administering pharmacologic therapy to patients.

Although the VA has the largest health care system in the U.S., not every VA facility offers medical subspecialty care. As a result, patients are often separated by long distances from services they need.

At the VA Pittsburgh Healthcare System (VAPHS) in Pennsylvania, about 15,700 veterans received care in 2011. The Gastroenterology Department (GD) served many of these patients: 5,800 patients were seen in clinic appointments, 2,500 underwent colonoscopy, and 1,700 underwent esophagogastroduodenoscopy (EGD). Patients traveled up to 150 miles from 3 states for appointments and procedures. Prior to each procedure, a face-to-face appointment was standard practice for most patients to plan procedures and ensure medical stability. Patients expressed dissatisfaction with transportation, cost, time, and inconvenience, particularly when they were required to attend both the preprocedure and procedure appointments.

Patient satisfaction, timely care, and appropriate use of resources are important VA goals of care, so the VAPHS developed an electronic consultation (e-consult) program as a component of its long-term strategic plan. The goal was to increase access to care through the use of an e-consult in lieu of a face-to-face appointment for select patients. The e-consult program established guidelines and benchmark goals (Table). The program also established a database to track the benchmarks.

The purpose of this quality improvement (QI) project was to evaluate e-consults in the GD over a 6-month period from January 2012, when e-consults were implemented in the GD, to July 2012. Based on the outcomes, recommendations for program continuation, change, and sustainability were made.

Background

Telemedicine using information technologies has been reported as a viable solution to support health care delivery when distance limits patients’ access to care.^1,2 Telemedicine has also been shown to improve efficacy in clinical decision making and reduce costs. It also can increase patient satisfaction by reducing travel and time, minimize duplication of diagnostic testing, and integrate services effectively across multiple sites when an electronic medical record (EMR) is in place.^1-5

A randomized controlled trial in 2004 compared a standard outpatient referral appointment with a joint teleconsultation between provider, specialist, and patient.³ In the teleconsultation arm, there was higher patient satisfaction, fewer diagnostic tests (particularly in gastroenterology), and lower patient costs.

A study published in 2009 examined the impact of cardiac, dermatologic, and diabetes teleconsultations on organization and patient outcomes in 950 patients in 30 rural communities.² Rapid access to care was provided for 85% of the patients. Organizational benefits included resource savings and efficacy improvement measured by a provider opinion Likert scale. Patient benefits included reductions in wait times, transportation savings, avoidance of unnecessary office visits, and ease of use.

A large systematic review of telemedicine services across all medical specialties in 2006 included 106 published studies.⁴ Clinical outcomes (decision making, diagnosis, and management) were similar between conventional care and telemedicine in specialty care, particularly in neurology and psychiatry. Virtual consults provided equal care to traditional specialty visits.⁶

Communication and coordination of care via an e-consult instead of a face-to-face clinic visit was evaluated by Horner and colleagues.⁵ The authors identified e-consult benefits for patients and specialists and that e-consults can reduce unnecessary referrals and appointments by 30%. They concluded that reserved time to complete e-consults must be built into workflow systems and that an advanced EMR was necessary for successful use of e-consults.

Two studies have evaluated satisfaction with e-consults. A preliminary analysis of satisfaction with e-consults was conducted in 2009 by K. L. Rodriguez, PhD, and colleagues (unpublished data). Patients, primary care providers (PCPs), and medical specialists reported overall satisfaction in 8 satisfaction domains. A pilot study of 34 patients in 2011 with inflammatory bowel disease compared a standard patient-GD physician encounter with a video encounter.⁷ The authors reported patient satisfaction, appointment time, wait times, and quality of care were similar for the 2 groups.

Methods

The GD where this QI project was completed consisted of a clinic staffed by nurse practitioners (NPs) and a procedure lab staffed by gastroenterology physicians. Before e-consult implementation, a NP reviewed and triaged new referrals daily. During the project period, an average of 25 to 35 new referrals were received daily via the EMR. Referrals came mostly from PCPs requesting an evaluation of their patients’ gastroenterology problem. Patients were triaged either to an appointment for evaluation or directly to the GD procedure lab for EGD or colonoscopy.

When e-consults were implemented, several changes occurred. Two providers were assigned to new referral triage, and they were expected to process 20% as e-consults. In the EMR, e-consult note titles, templates, and an e-consult encounter form were created, and staff was given access to the e-consult tracking database. The EMR referral template was amended so the entering provider could say whether a face-to-face appointment was desired or whether an e-consult was acceptable. The patient was to be included in decision making about this choice. Department staff had permission to triage according to judgment and expertise; thus, appointment requests could be triaged to e-consults, and e-consult requests could be triaged to appointments.

To complete an e-consult, the EMR was reviewed for medical diagnoses, medications, diagnostics, and recent physical exam. A summary note outlining an impression, treatment recommendations, and follow-up was entered in the EMR and communicated to the PCP. In most cases, no discussion with the patient occurred. The encounter form was completed, and information was entered into the tracking database. The database was installed on each provider’s computer who processed e-consults. If EGD or colonoscopy was indicated, the scheduler was notified to call the patient. Once a procedure date was established, procedure orders were completed in the EMR, and instructions were mailed to the patient.

Project Description and Evaluation 

The project was reviewed by the Institutional Review Board and determined to be a QI project. VA organizational policies were followed for data collection and security. Benchmarks were identified from the e-consult program and from the GD, where available. Process variables were determined to measure benchmark outcomes. (Table)

To identify participants, a retrospective chart review was performed. A total of 203 potential patients were identified from the e-consult program database for the 6-month period between January and July 2012. For comparison, 50 patients who attended an appointment during the same time frame were systematically identified in the EMR. Although this comparison group was eligible for e-consults, they were triaged to in-person appointments and subsequently had colonoscopies completed.

Outcome data were extracted from the e-consult program database and the EMR. The data analysis was descriptive. Summary aggregate data were compared with the benchmarks and comparison patient outcomes. The Table summarizes the process variables, how they were measured, where they came from, and what the comparisons were.

Discussion 

Figure 1 illustrates the volume of completed e-consults from January to July 2012. A gastroenterology procedure was not indicated in 43 patients. A procedure was indicated for 160 patients and completed in 116 patients (72%). One hundred procedures were colonoscopies, and 16 were EGDs. Figure 2 provides reasons why procedures were not completed in 44 patients (27%). Group comparisons of colon prep quality and preprocedure reminder calls are displayed in Figures 3 and 4, respectively.

This project sought to evaluate VAPHS GD e-consults beginning in January 2012. Process variables were established to measure benchmarks in the e-consult program and in the GD. Some benchmarks were met with outcomes that were comparable between the groups, while others were not. To our knowledge, this project is the first to evaluate e-consults in the subspecialty of gastroenterology.

Volume of Completed E-Consults

The benchmark for 20% e-consults was not met (Figure 1). For weeks 1 through 8, the volume was between 10% and 20%. Lower volume in weeks 9 through 15 (late March and April) may have been due to staff vacations. Not only do the outcomes show a downward trend in e-consult volume, they also show a precipitous fall in volume at week 15 to almost no e-consults for the remainder of the project. To explore reasons for this outcome, the workflow process of new referral triage and e-consults was reviewed.

Two providers (1 NP, 1 physician) were assigned to new referral triage and e-consults from weeks 1 through 14. At week 15, the physician was reassigned to perform procedures. From this point, only 1 NP worked on e-consults and referral triage. Competing time demands included an e-consult encounter form, tracking database entry, procedure orders, patient instructions, appointment changes, phone calls, and resolution of medications issues for procedures. The triage NP was also required to see patients in the clinic. Each day, only a half-day was allotted to complete e-consults, new referral triage (25-35/day), and the aforementioned tasks.

Therefore, it became clear that a half-day was not sufficient to meet the 20% benchmark for e-consults. Horner and colleagues also found that dedicated time in workflow processes was necessary to allow for e-consult completion.⁵

E-Consults vs Appointment Groups 

All patients in both groups were offered the choice of e-consult or appointment; this benchmark was met. Of the 203 e-consult patients, 70% requested an appointment, but their evaluation was completed as an e-consult. By design, the appointment group patients chose e-consult but were triaged to appointment due to time constraints and the high volume of new referrals.

Evaluations via e-consult were completed within 2 to 3 days, whereas the mean for appointments was 19 days, with the longest time frame of 44 days. Thus, e-consult evaluations were completed sooner. Rapid access to care was also found by Zanaboni and colleagues.²

When appointments are delayed, patient complaints or status may change, which in turn may affect treatment plans. In addition, the reason for referral may have already resolved by the actual appointment, rendering the appointment unnecessary. This can be viewed as a missed opportunity for another patient to be seen. Ideally, it is best for a patient to be evaluated soon after a new referral is made.

The VA encounter form for an e-consult had only one 5-digit code, which allotted only 15 minutes of work credit. Encounter form codes were established by the Center for Medicare and Medicaid Services (CMS) for billing purposes in the private sector, with coding levels based on information documented in a chart note: review of systems, physical exam, and diagnoses decision making. Because all criteria could not be met in an e-consult, only 1 code was assigned for VAPHS e-consults. The CMS has specific telemedicine codes; it was unknown why they were not used for e-consults.

E-consults took an average of 19 minutes to complete, with 91 completed in ≤ 15 minutes and over half (112) having taken > 15 minutes. Therefore, the actual workload was not captured, and more work was done than was credited. To speculate, e-consults were in their infancy; a learning curve may have existed as staff became accustomed to this new process.

The EMRs were reviewed for the 7 e-consults that took > 30 minutes to complete. Two were in the early stage of e-consult implementation, but the remainder were scattered throughout. Patients in these e-consults had complicated medical histories and perhaps should not have been triaged to e-consult. Theoretically, only uncomplicated patients with simple reasons for gastroenterology referral should be triaged to e-consult, allowing for a shorter time frame and higher volume.

The wait times to procedure were 58 days for the e-consult group and 39 days in the appointment group. Although wait time was originally identified as an outcome, its relevance is questionable after looking at the outcome data. The procedure appointment date was a subjective decision by the patient; many factors affected what date the patient established, including weekday preference, time off from work, caretaker availability, season, and staffing. Many patients rescheduled their initial procedure dates, often several times. These factors are reflected in the variable ranges of wait times.

Colon Prep Quality

Colon prep refers to patient instructions on the day before the procedure and includes a clear liquid diet, drinking a liquid solution to empty bowel contents, and no food or liquid after midnight. Prep quality is stated in the colonoscopy report. During the procedure, the physician makes a visual decision based on presence or absence of stool inside the colon. Prep quality is important, because retained stool can preclude thorough visualization of the colon wall for polyps or abnormalities. In the event of fair and poor preps, the colonoscopy might be aborted and rescheduled or completed, but with the recommendation for another colonoscopy in a short time frame, such as 1 to 3 years.

Forty-four percent of the e-consult group and 62% of the appointment group had good or adequate preps. Thus, more patients in the appointment group achieved good and adequate preps, and far fewer achieved fair or poor preps. One important point was that almost half (47%) of the e-consult group had only a fair prep (Figure 3).

A number of reasons have been identified in the literature, which might help us understand these findings. First, patients may not fully understand or adhere to prep instructions.^8-10 Furthermore, certain medical diagnoses are known to affect prep quality (ie, diabetes, thyroid disease, constipation).^11,12 Another potential factor is the manner in which prep quality is determined.^13,14 However, due to the focus of this QI study, the influential drivers of prep quality can only be inferred from the literature; thus, a future research or QI study is warranted to ascertain the underlying mechanism of colon prep quality in our specific veteran population.

Preprocedure Reminder Calls

Outcomes were essentially reversed between the 2 groups (Figure 4). Between 50% and 60% of the e-consult group received a call, while the same percentage of appointment patients did not. All patients did attend their procedure appointment. A GD goal was to call every patient before their procedure, but the ability to make the calls was staffing-dependent, which may explain these findings.

Most Relevant Findings

Although this project provides a thorough analysis of various benchmarks within this recently implemented e-consult gastroenterology program, 3 findings emerged that were identified as most relevant. First, the benchmark of 20% volume of completed e-consults was not met. A review of the workflow processes revealed that a daily allotment of only a half-day was not sufficient to complete e-consults, referral triage, and related tasks.

Second, outcomes for colon prep quality and preprocedure reminder calls were also relevant. Although beyond the scope of this project, the question arose of a relationship between prep quality and the reminder call: Does the reminder call affect prep quality? The goal of colon prep is to achieve a good or adequate prep. The purpose of the reminder call was to confirm the appointment and to review the colon prep. Among patients in both groups who achieved only a fair prep, 62% in the e-consult group did receive the reminder call; thus, the call seemed to have failed in helping these patients achieve an adequate or good prep. The actual content of colon prep review during the call was unknown, but certainly bears improvement.

Another speculation concerned the prep instruction sheet. Although all patients received the same sheet, e-consult patients received it in the mail and read it themselves, while it was directly reviewed face-to-face with the appointment patients. Questions remain whether a face-to-face review increases the likelihood for a better prep and how to help e-consult patients achieve optimal prep, since they are not seen face-to-face. 

Practice Implications and Sustaining Measures

Theoretically, e-consults are a viable alternative to face-to-face appointments. Potential advantages include efficient use of an EMR, avoidance of unnecessary appointments, and improved access to care for patients who require an appointment. Although patient satisfaction was not measured in this project, the literature review revealed that satisfaction was increased through use of various virtual health care modalities, including a preliminary analysis in this facility by the aforementioned 2009 study by K. L. Rodriguez (unpublished data). Based on findings in this project, the following 4 recommendations were made to improve benchmark outcomes and quality of care.

1. Provide dedicated time to complete e-consults and related tasks. In this setting, a full day was recommended. An alternative was to hire a NP whose sole responsibility was e-consults.
2. Develop a selection process to determine which new referrals are best suited for e-consult. This process will increase e-consult efficiency and decrease the time to complete an e-consult. Recommended selection criteria were (A) gastroenterology referrals only for simple symptoms or issues; (B) referrals only for a procedure; and (C) stable patients with uncomplicated medical histories.
3. Sustain the preprocedure reminder phone call. The reminder call helps patients keep appointments and thus reduces a missed opportunity for care.
4. Plan a future QI project or research study on patient colon prep quality for colonoscopy. Such a project might evaluate types of colon prep, how prep quality is measured, patient instructions, and the timing/content of pre-procedure reminder phone calls, particularly for e-consult patients.

Conclusion

This QI project provided outcomes for e-consults in the subspecialty of gastroenterology at the VAPHS. Although some benchmark outcomes were met and favorable, others were less favorable. By conducting this benchmark analysis, the areas of needed improvement are now clear. This analysis provides information so recommendations for process improvements can be made.

Quality of care improvement is an ongoing process at VAPHS. Since completion of this project, several processes have been adjusted so that outcomes will be improved. For example, corrective actions were taken for patients who did not complete their gastroenterology procedure. The process for scheduling gastroenterology procedures was adjusted for appointments and cancellations. Ongoing efforts to sustain the reminder phone call were put in place. Changes in NP staffing and time assigned for both clinical and nonclinical work were proposed and are currently under review.

It is the mission of the VA to provide access to care, patient satisfaction, timely care, and appropriate use of resources. Having the ability to highlight our strengths, as well as the willingness to recognize weaknesses, allows us to create new improved processes to provide the best care possible for our veterans.

Acknowledgments
This project was Elena Swann’s capstone for the Doctor of Nursing Practice Program, University of Pittsburgh School of Nursing in Pennsylvania. Rich Laufer, Larry Priscella, and Janie Fleming assisted with the VA Gastroenterology Clinic and procedure metrics. Dr. Melissa Taylor, VA associate chief nurse for research provided project guidance and manuscript revisions. Dr. Sandra Engberg, University of Pittsburgh School of Nursing faculty, assisted with project and manuscript development.

Author disclosures
The author reports no actual or potential conflicts of interest with regard to this article.

Disclaimer
The opinions expressed herein are those of the author and do not necessarily reflect those of Federal Practitioner, Frontline Medical Communications Inc., the U.S. Government, or any of its agencies. This article may discuss unlabeled or investigational use of certain drugs. Please review complete prescribing information for specific drugs or drug combinations—including indications, contraindications, warnings, and adverse effects—before administering pharmacologic therapy to patients.

Although the VA has the largest health care system in the U.S., not every VA facility offers medical subspecialty care. As a result, patients are often separated by long distances from services they need.

At the VA Pittsburgh Healthcare System (VAPHS) in Pennsylvania, about 15,700 veterans received care in 2011. The Gastroenterology Department (GD) served many of these patients: 5,800 patients were seen in clinic appointments, 2,500 underwent colonoscopy, and 1,700 underwent esophagogastroduodenoscopy (EGD). Patients traveled up to 150 miles from 3 states for appointments and procedures. Prior to each procedure, a face-to-face appointment was standard practice for most patients to plan procedures and ensure medical stability. Patients expressed dissatisfaction with transportation, cost, time, and inconvenience, particularly when they were required to attend both the preprocedure and procedure appointments.

Patient satisfaction, timely care, and appropriate use of resources are important VA goals of care, so the VAPHS developed an electronic consultation (e-consult) program as a component of its long-term strategic plan. The goal was to increase access to care through the use of an e-consult in lieu of a face-to-face appointment for select patients. The e-consult program established guidelines and benchmark goals (Table). The program also established a database to track the benchmarks.

The purpose of this quality improvement (QI) project was to evaluate e-consults in the GD over a 6-month period from January 2012, when e-consults were implemented in the GD, to July 2012. Based on the outcomes, recommendations for program continuation, change, and sustainability were made.

Background

Telemedicine using information technologies has been reported as a viable solution to support health care delivery when distance limits patients’ access to care.^1,2 Telemedicine has also been shown to improve efficacy in clinical decision making and reduce costs. It also can increase patient satisfaction by reducing travel and time, minimize duplication of diagnostic testing, and integrate services effectively across multiple sites when an electronic medical record (EMR) is in place.^1-5

A randomized controlled trial in 2004 compared a standard outpatient referral appointment with a joint teleconsultation between provider, specialist, and patient.³ In the teleconsultation arm, there was higher patient satisfaction, fewer diagnostic tests (particularly in gastroenterology), and lower patient costs.

A study published in 2009 examined the impact of cardiac, dermatologic, and diabetes teleconsultations on organization and patient outcomes in 950 patients in 30 rural communities.² Rapid access to care was provided for 85% of the patients. Organizational benefits included resource savings and efficacy improvement measured by a provider opinion Likert scale. Patient benefits included reductions in wait times, transportation savings, avoidance of unnecessary office visits, and ease of use.

A large systematic review of telemedicine services across all medical specialties in 2006 included 106 published studies.⁴ Clinical outcomes (decision making, diagnosis, and management) were similar between conventional care and telemedicine in specialty care, particularly in neurology and psychiatry. Virtual consults provided equal care to traditional specialty visits.⁶

Communication and coordination of care via an e-consult instead of a face-to-face clinic visit was evaluated by Horner and colleagues.⁵ The authors identified e-consult benefits for patients and specialists and that e-consults can reduce unnecessary referrals and appointments by 30%. They concluded that reserved time to complete e-consults must be built into workflow systems and that an advanced EMR was necessary for successful use of e-consults.

Two studies have evaluated satisfaction with e-consults. A preliminary analysis of satisfaction with e-consults was conducted in 2009 by K. L. Rodriguez, PhD, and colleagues (unpublished data). Patients, primary care providers (PCPs), and medical specialists reported overall satisfaction in 8 satisfaction domains. A pilot study of 34 patients in 2011 with inflammatory bowel disease compared a standard patient-GD physician encounter with a video encounter.⁷ The authors reported patient satisfaction, appointment time, wait times, and quality of care were similar for the 2 groups.

Methods

The GD where this QI project was completed consisted of a clinic staffed by nurse practitioners (NPs) and a procedure lab staffed by gastroenterology physicians. Before e-consult implementation, a NP reviewed and triaged new referrals daily. During the project period, an average of 25 to 35 new referrals were received daily via the EMR. Referrals came mostly from PCPs requesting an evaluation of their patients’ gastroenterology problem. Patients were triaged either to an appointment for evaluation or directly to the GD procedure lab for EGD or colonoscopy.

When e-consults were implemented, several changes occurred. Two providers were assigned to new referral triage, and they were expected to process 20% as e-consults. In the EMR, e-consult note titles, templates, and an e-consult encounter form were created, and staff was given access to the e-consult tracking database. The EMR referral template was amended so the entering provider could say whether a face-to-face appointment was desired or whether an e-consult was acceptable. The patient was to be included in decision making about this choice. Department staff had permission to triage according to judgment and expertise; thus, appointment requests could be triaged to e-consults, and e-consult requests could be triaged to appointments.

To complete an e-consult, the EMR was reviewed for medical diagnoses, medications, diagnostics, and recent physical exam. A summary note outlining an impression, treatment recommendations, and follow-up was entered in the EMR and communicated to the PCP. In most cases, no discussion with the patient occurred. The encounter form was completed, and information was entered into the tracking database. The database was installed on each provider’s computer who processed e-consults. If EGD or colonoscopy was indicated, the scheduler was notified to call the patient. Once a procedure date was established, procedure orders were completed in the EMR, and instructions were mailed to the patient.

Project Description and Evaluation 

The project was reviewed by the Institutional Review Board and determined to be a QI project. VA organizational policies were followed for data collection and security. Benchmarks were identified from the e-consult program and from the GD, where available. Process variables were determined to measure benchmark outcomes. (Table)

To identify participants, a retrospective chart review was performed. A total of 203 potential patients were identified from the e-consult program database for the 6-month period between January and July 2012. For comparison, 50 patients who attended an appointment during the same time frame were systematically identified in the EMR. Although this comparison group was eligible for e-consults, they were triaged to in-person appointments and subsequently had colonoscopies completed.

Outcome data were extracted from the e-consult program database and the EMR. The data analysis was descriptive. Summary aggregate data were compared with the benchmarks and comparison patient outcomes. The Table summarizes the process variables, how they were measured, where they came from, and what the comparisons were.

Discussion 

Figure 1 illustrates the volume of completed e-consults from January to July 2012. A gastroenterology procedure was not indicated in 43 patients. A procedure was indicated for 160 patients and completed in 116 patients (72%). One hundred procedures were colonoscopies, and 16 were EGDs. Figure 2 provides reasons why procedures were not completed in 44 patients (27%). Group comparisons of colon prep quality and preprocedure reminder calls are displayed in Figures 3 and 4, respectively.

This project sought to evaluate VAPHS GD e-consults beginning in January 2012. Process variables were established to measure benchmarks in the e-consult program and in the GD. Some benchmarks were met with outcomes that were comparable between the groups, while others were not. To our knowledge, this project is the first to evaluate e-consults in the subspecialty of gastroenterology.

Volume of Completed E-Consults

The benchmark for 20% e-consults was not met (Figure 1). For weeks 1 through 8, the volume was between 10% and 20%. Lower volume in weeks 9 through 15 (late March and April) may have been due to staff vacations. Not only do the outcomes show a downward trend in e-consult volume, they also show a precipitous fall in volume at week 15 to almost no e-consults for the remainder of the project. To explore reasons for this outcome, the workflow process of new referral triage and e-consults was reviewed.

Two providers (1 NP, 1 physician) were assigned to new referral triage and e-consults from weeks 1 through 14. At week 15, the physician was reassigned to perform procedures. From this point, only 1 NP worked on e-consults and referral triage. Competing time demands included an e-consult encounter form, tracking database entry, procedure orders, patient instructions, appointment changes, phone calls, and resolution of medications issues for procedures. The triage NP was also required to see patients in the clinic. Each day, only a half-day was allotted to complete e-consults, new referral triage (25-35/day), and the aforementioned tasks.

Therefore, it became clear that a half-day was not sufficient to meet the 20% benchmark for e-consults. Horner and colleagues also found that dedicated time in workflow processes was necessary to allow for e-consult completion.⁵

E-Consults vs Appointment Groups 

All patients in both groups were offered the choice of e-consult or appointment; this benchmark was met. Of the 203 e-consult patients, 70% requested an appointment, but their evaluation was completed as an e-consult. By design, the appointment group patients chose e-consult but were triaged to appointment due to time constraints and the high volume of new referrals.

Evaluations via e-consult were completed within 2 to 3 days, whereas the mean for appointments was 19 days, with the longest time frame of 44 days. Thus, e-consult evaluations were completed sooner. Rapid access to care was also found by Zanaboni and colleagues.²

When appointments are delayed, patient complaints or status may change, which in turn may affect treatment plans. In addition, the reason for referral may have already resolved by the actual appointment, rendering the appointment unnecessary. This can be viewed as a missed opportunity for another patient to be seen. Ideally, it is best for a patient to be evaluated soon after a new referral is made.

The VA encounter form for an e-consult had only one 5-digit code, which allotted only 15 minutes of work credit. Encounter form codes were established by the Center for Medicare and Medicaid Services (CMS) for billing purposes in the private sector, with coding levels based on information documented in a chart note: review of systems, physical exam, and diagnoses decision making. Because all criteria could not be met in an e-consult, only 1 code was assigned for VAPHS e-consults. The CMS has specific telemedicine codes; it was unknown why they were not used for e-consults.

E-consults took an average of 19 minutes to complete, with 91 completed in ≤ 15 minutes and over half (112) having taken > 15 minutes. Therefore, the actual workload was not captured, and more work was done than was credited. To speculate, e-consults were in their infancy; a learning curve may have existed as staff became accustomed to this new process.

The EMRs were reviewed for the 7 e-consults that took > 30 minutes to complete. Two were in the early stage of e-consult implementation, but the remainder were scattered throughout. Patients in these e-consults had complicated medical histories and perhaps should not have been triaged to e-consult. Theoretically, only uncomplicated patients with simple reasons for gastroenterology referral should be triaged to e-consult, allowing for a shorter time frame and higher volume.

The wait times to procedure were 58 days for the e-consult group and 39 days in the appointment group. Although wait time was originally identified as an outcome, its relevance is questionable after looking at the outcome data. The procedure appointment date was a subjective decision by the patient; many factors affected what date the patient established, including weekday preference, time off from work, caretaker availability, season, and staffing. Many patients rescheduled their initial procedure dates, often several times. These factors are reflected in the variable ranges of wait times.

Colon Prep Quality

Colon prep refers to patient instructions on the day before the procedure and includes a clear liquid diet, drinking a liquid solution to empty bowel contents, and no food or liquid after midnight. Prep quality is stated in the colonoscopy report. During the procedure, the physician makes a visual decision based on presence or absence of stool inside the colon. Prep quality is important, because retained stool can preclude thorough visualization of the colon wall for polyps or abnormalities. In the event of fair and poor preps, the colonoscopy might be aborted and rescheduled or completed, but with the recommendation for another colonoscopy in a short time frame, such as 1 to 3 years.

Forty-four percent of the e-consult group and 62% of the appointment group had good or adequate preps. Thus, more patients in the appointment group achieved good and adequate preps, and far fewer achieved fair or poor preps. One important point was that almost half (47%) of the e-consult group had only a fair prep (Figure 3).

A number of reasons have been identified in the literature, which might help us understand these findings. First, patients may not fully understand or adhere to prep instructions.^8-10 Furthermore, certain medical diagnoses are known to affect prep quality (ie, diabetes, thyroid disease, constipation).^11,12 Another potential factor is the manner in which prep quality is determined.^13,14 However, due to the focus of this QI study, the influential drivers of prep quality can only be inferred from the literature; thus, a future research or QI study is warranted to ascertain the underlying mechanism of colon prep quality in our specific veteran population.

Preprocedure Reminder Calls

Outcomes were essentially reversed between the 2 groups (Figure 4). Between 50% and 60% of the e-consult group received a call, while the same percentage of appointment patients did not. All patients did attend their procedure appointment. A GD goal was to call every patient before their procedure, but the ability to make the calls was staffing-dependent, which may explain these findings.

Most Relevant Findings

Although this project provides a thorough analysis of various benchmarks within this recently implemented e-consult gastroenterology program, 3 findings emerged that were identified as most relevant. First, the benchmark of 20% volume of completed e-consults was not met. A review of the workflow processes revealed that a daily allotment of only a half-day was not sufficient to complete e-consults, referral triage, and related tasks.

Second, outcomes for colon prep quality and preprocedure reminder calls were also relevant. Although beyond the scope of this project, the question arose of a relationship between prep quality and the reminder call: Does the reminder call affect prep quality? The goal of colon prep is to achieve a good or adequate prep. The purpose of the reminder call was to confirm the appointment and to review the colon prep. Among patients in both groups who achieved only a fair prep, 62% in the e-consult group did receive the reminder call; thus, the call seemed to have failed in helping these patients achieve an adequate or good prep. The actual content of colon prep review during the call was unknown, but certainly bears improvement.

Another speculation concerned the prep instruction sheet. Although all patients received the same sheet, e-consult patients received it in the mail and read it themselves, while it was directly reviewed face-to-face with the appointment patients. Questions remain whether a face-to-face review increases the likelihood for a better prep and how to help e-consult patients achieve optimal prep, since they are not seen face-to-face. 

Practice Implications and Sustaining Measures

Theoretically, e-consults are a viable alternative to face-to-face appointments. Potential advantages include efficient use of an EMR, avoidance of unnecessary appointments, and improved access to care for patients who require an appointment. Although patient satisfaction was not measured in this project, the literature review revealed that satisfaction was increased through use of various virtual health care modalities, including a preliminary analysis in this facility by the aforementioned 2009 study by K. L. Rodriguez (unpublished data). Based on findings in this project, the following 4 recommendations were made to improve benchmark outcomes and quality of care.

1. Provide dedicated time to complete e-consults and related tasks. In this setting, a full day was recommended. An alternative was to hire a NP whose sole responsibility was e-consults.
2. Develop a selection process to determine which new referrals are best suited for e-consult. This process will increase e-consult efficiency and decrease the time to complete an e-consult. Recommended selection criteria were (A) gastroenterology referrals only for simple symptoms or issues; (B) referrals only for a procedure; and (C) stable patients with uncomplicated medical histories.
3. Sustain the preprocedure reminder phone call. The reminder call helps patients keep appointments and thus reduces a missed opportunity for care.
4. Plan a future QI project or research study on patient colon prep quality for colonoscopy. Such a project might evaluate types of colon prep, how prep quality is measured, patient instructions, and the timing/content of pre-procedure reminder phone calls, particularly for e-consult patients.

Conclusion

This QI project provided outcomes for e-consults in the subspecialty of gastroenterology at the VAPHS. Although some benchmark outcomes were met and favorable, others were less favorable. By conducting this benchmark analysis, the areas of needed improvement are now clear. This analysis provides information so recommendations for process improvements can be made.

Quality of care improvement is an ongoing process at VAPHS. Since completion of this project, several processes have been adjusted so that outcomes will be improved. For example, corrective actions were taken for patients who did not complete their gastroenterology procedure. The process for scheduling gastroenterology procedures was adjusted for appointments and cancellations. Ongoing efforts to sustain the reminder phone call were put in place. Changes in NP staffing and time assigned for both clinical and nonclinical work were proposed and are currently under review.

It is the mission of the VA to provide access to care, patient satisfaction, timely care, and appropriate use of resources. Having the ability to highlight our strengths, as well as the willingness to recognize weaknesses, allows us to create new improved processes to provide the best care possible for our veterans.

Acknowledgments
This project was Elena Swann’s capstone for the Doctor of Nursing Practice Program, University of Pittsburgh School of Nursing in Pennsylvania. Rich Laufer, Larry Priscella, and Janie Fleming assisted with the VA Gastroenterology Clinic and procedure metrics. Dr. Melissa Taylor, VA associate chief nurse for research provided project guidance and manuscript revisions. Dr. Sandra Engberg, University of Pittsburgh School of Nursing faculty, assisted with project and manuscript development.

Author disclosures
The author reports no actual or potential conflicts of interest with regard to this article.

Disclaimer
The opinions expressed herein are those of the author and do not necessarily reflect those of Federal Practitioner, Frontline Medical Communications Inc., the U.S. Government, or any of its agencies. This article may discuss unlabeled or investigational use of certain drugs. Please review complete prescribing information for specific drugs or drug combinations—including indications, contraindications, warnings, and adverse effects—before administering pharmacologic therapy to patients.

Although the VA has the largest health care system in the U.S., not every VA facility offers medical subspecialty care. As a result, patients are often separated by long distances from services they need.

At the VA Pittsburgh Healthcare System (VAPHS) in Pennsylvania, about 15,700 veterans received care in 2011. The Gastroenterology Department (GD) served many of these patients: 5,800 patients were seen in clinic appointments, 2,500 underwent colonoscopy, and 1,700 underwent esophagogastroduodenoscopy (EGD). Patients traveled up to 150 miles from 3 states for appointments and procedures. Prior to each procedure, a face-to-face appointment was standard practice for most patients to plan procedures and ensure medical stability. Patients expressed dissatisfaction with transportation, cost, time, and inconvenience, particularly when they were required to attend both the preprocedure and procedure appointments.

Patient satisfaction, timely care, and appropriate use of resources are important VA goals of care, so the VAPHS developed an electronic consultation (e-consult) program as a component of its long-term strategic plan. The goal was to increase access to care through the use of an e-consult in lieu of a face-to-face appointment for select patients. The e-consult program established guidelines and benchmark goals (Table). The program also established a database to track the benchmarks.

The purpose of this quality improvement (QI) project was to evaluate e-consults in the GD over a 6-month period from January 2012, when e-consults were implemented in the GD, to July 2012. Based on the outcomes, recommendations for program continuation, change, and sustainability were made.

Background

Telemedicine using information technologies has been reported as a viable solution to support health care delivery when distance limits patients’ access to care.^1,2 Telemedicine has also been shown to improve efficacy in clinical decision making and reduce costs. It also can increase patient satisfaction by reducing travel and time, minimize duplication of diagnostic testing, and integrate services effectively across multiple sites when an electronic medical record (EMR) is in place.^1-5

A randomized controlled trial in 2004 compared a standard outpatient referral appointment with a joint teleconsultation between provider, specialist, and patient.³ In the teleconsultation arm, there was higher patient satisfaction, fewer diagnostic tests (particularly in gastroenterology), and lower patient costs.

A study published in 2009 examined the impact of cardiac, dermatologic, and diabetes teleconsultations on organization and patient outcomes in 950 patients in 30 rural communities.² Rapid access to care was provided for 85% of the patients. Organizational benefits included resource savings and efficacy improvement measured by a provider opinion Likert scale. Patient benefits included reductions in wait times, transportation savings, avoidance of unnecessary office visits, and ease of use.

A large systematic review of telemedicine services across all medical specialties in 2006 included 106 published studies.⁴ Clinical outcomes (decision making, diagnosis, and management) were similar between conventional care and telemedicine in specialty care, particularly in neurology and psychiatry. Virtual consults provided equal care to traditional specialty visits.⁶

Communication and coordination of care via an e-consult instead of a face-to-face clinic visit was evaluated by Horner and colleagues.⁵ The authors identified e-consult benefits for patients and specialists and that e-consults can reduce unnecessary referrals and appointments by 30%. They concluded that reserved time to complete e-consults must be built into workflow systems and that an advanced EMR was necessary for successful use of e-consults.

Two studies have evaluated satisfaction with e-consults. A preliminary analysis of satisfaction with e-consults was conducted in 2009 by K. L. Rodriguez, PhD, and colleagues (unpublished data). Patients, primary care providers (PCPs), and medical specialists reported overall satisfaction in 8 satisfaction domains. A pilot study of 34 patients in 2011 with inflammatory bowel disease compared a standard patient-GD physician encounter with a video encounter.⁷ The authors reported patient satisfaction, appointment time, wait times, and quality of care were similar for the 2 groups.

Methods

The GD where this QI project was completed consisted of a clinic staffed by nurse practitioners (NPs) and a procedure lab staffed by gastroenterology physicians. Before e-consult implementation, a NP reviewed and triaged new referrals daily. During the project period, an average of 25 to 35 new referrals were received daily via the EMR. Referrals came mostly from PCPs requesting an evaluation of their patients’ gastroenterology problem. Patients were triaged either to an appointment for evaluation or directly to the GD procedure lab for EGD or colonoscopy.

When e-consults were implemented, several changes occurred. Two providers were assigned to new referral triage, and they were expected to process 20% as e-consults. In the EMR, e-consult note titles, templates, and an e-consult encounter form were created, and staff was given access to the e-consult tracking database. The EMR referral template was amended so the entering provider could say whether a face-to-face appointment was desired or whether an e-consult was acceptable. The patient was to be included in decision making about this choice. Department staff had permission to triage according to judgment and expertise; thus, appointment requests could be triaged to e-consults, and e-consult requests could be triaged to appointments.

To complete an e-consult, the EMR was reviewed for medical diagnoses, medications, diagnostics, and recent physical exam. A summary note outlining an impression, treatment recommendations, and follow-up was entered in the EMR and communicated to the PCP. In most cases, no discussion with the patient occurred. The encounter form was completed, and information was entered into the tracking database. The database was installed on each provider’s computer who processed e-consults. If EGD or colonoscopy was indicated, the scheduler was notified to call the patient. Once a procedure date was established, procedure orders were completed in the EMR, and instructions were mailed to the patient.

Project Description and Evaluation 

The project was reviewed by the Institutional Review Board and determined to be a QI project. VA organizational policies were followed for data collection and security. Benchmarks were identified from the e-consult program and from the GD, where available. Process variables were determined to measure benchmark outcomes. (Table)

To identify participants, a retrospective chart review was performed. A total of 203 potential patients were identified from the e-consult program database for the 6-month period between January and July 2012. For comparison, 50 patients who attended an appointment during the same time frame were systematically identified in the EMR. Although this comparison group was eligible for e-consults, they were triaged to in-person appointments and subsequently had colonoscopies completed.

Outcome data were extracted from the e-consult program database and the EMR. The data analysis was descriptive. Summary aggregate data were compared with the benchmarks and comparison patient outcomes. The Table summarizes the process variables, how they were measured, where they came from, and what the comparisons were.

Discussion 

Figure 1 illustrates the volume of completed e-consults from January to July 2012. A gastroenterology procedure was not indicated in 43 patients. A procedure was indicated for 160 patients and completed in 116 patients (72%). One hundred procedures were colonoscopies, and 16 were EGDs. Figure 2 provides reasons why procedures were not completed in 44 patients (27%). Group comparisons of colon prep quality and preprocedure reminder calls are displayed in Figures 3 and 4, respectively.

This project sought to evaluate VAPHS GD e-consults beginning in January 2012. Process variables were established to measure benchmarks in the e-consult program and in the GD. Some benchmarks were met with outcomes that were comparable between the groups, while others were not. To our knowledge, this project is the first to evaluate e-consults in the subspecialty of gastroenterology.

Volume of Completed E-Consults

The benchmark for 20% e-consults was not met (Figure 1). For weeks 1 through 8, the volume was between 10% and 20%. Lower volume in weeks 9 through 15 (late March and April) may have been due to staff vacations. Not only do the outcomes show a downward trend in e-consult volume, they also show a precipitous fall in volume at week 15 to almost no e-consults for the remainder of the project. To explore reasons for this outcome, the workflow process of new referral triage and e-consults was reviewed.

Two providers (1 NP, 1 physician) were assigned to new referral triage and e-consults from weeks 1 through 14. At week 15, the physician was reassigned to perform procedures. From this point, only 1 NP worked on e-consults and referral triage. Competing time demands included an e-consult encounter form, tracking database entry, procedure orders, patient instructions, appointment changes, phone calls, and resolution of medications issues for procedures. The triage NP was also required to see patients in the clinic. Each day, only a half-day was allotted to complete e-consults, new referral triage (25-35/day), and the aforementioned tasks.

Therefore, it became clear that a half-day was not sufficient to meet the 20% benchmark for e-consults. Horner and colleagues also found that dedicated time in workflow processes was necessary to allow for e-consult completion.⁵

E-Consults vs Appointment Groups 

All patients in both groups were offered the choice of e-consult or appointment; this benchmark was met. Of the 203 e-consult patients, 70% requested an appointment, but their evaluation was completed as an e-consult. By design, the appointment group patients chose e-consult but were triaged to appointment due to time constraints and the high volume of new referrals.

Evaluations via e-consult were completed within 2 to 3 days, whereas the mean for appointments was 19 days, with the longest time frame of 44 days. Thus, e-consult evaluations were completed sooner. Rapid access to care was also found by Zanaboni and colleagues.²

When appointments are delayed, patient complaints or status may change, which in turn may affect treatment plans. In addition, the reason for referral may have already resolved by the actual appointment, rendering the appointment unnecessary. This can be viewed as a missed opportunity for another patient to be seen. Ideally, it is best for a patient to be evaluated soon after a new referral is made.

The VA encounter form for an e-consult had only one 5-digit code, which allotted only 15 minutes of work credit. Encounter form codes were established by the Center for Medicare and Medicaid Services (CMS) for billing purposes in the private sector, with coding levels based on information documented in a chart note: review of systems, physical exam, and diagnoses decision making. Because all criteria could not be met in an e-consult, only 1 code was assigned for VAPHS e-consults. The CMS has specific telemedicine codes; it was unknown why they were not used for e-consults.

E-consults took an average of 19 minutes to complete, with 91 completed in ≤ 15 minutes and over half (112) having taken > 15 minutes. Therefore, the actual workload was not captured, and more work was done than was credited. To speculate, e-consults were in their infancy; a learning curve may have existed as staff became accustomed to this new process.

The EMRs were reviewed for the 7 e-consults that took > 30 minutes to complete. Two were in the early stage of e-consult implementation, but the remainder were scattered throughout. Patients in these e-consults had complicated medical histories and perhaps should not have been triaged to e-consult. Theoretically, only uncomplicated patients with simple reasons for gastroenterology referral should be triaged to e-consult, allowing for a shorter time frame and higher volume.

The wait times to procedure were 58 days for the e-consult group and 39 days in the appointment group. Although wait time was originally identified as an outcome, its relevance is questionable after looking at the outcome data. The procedure appointment date was a subjective decision by the patient; many factors affected what date the patient established, including weekday preference, time off from work, caretaker availability, season, and staffing. Many patients rescheduled their initial procedure dates, often several times. These factors are reflected in the variable ranges of wait times.

Colon Prep Quality

Colon prep refers to patient instructions on the day before the procedure and includes a clear liquid diet, drinking a liquid solution to empty bowel contents, and no food or liquid after midnight. Prep quality is stated in the colonoscopy report. During the procedure, the physician makes a visual decision based on presence or absence of stool inside the colon. Prep quality is important, because retained stool can preclude thorough visualization of the colon wall for polyps or abnormalities. In the event of fair and poor preps, the colonoscopy might be aborted and rescheduled or completed, but with the recommendation for another colonoscopy in a short time frame, such as 1 to 3 years.

Forty-four percent of the e-consult group and 62% of the appointment group had good or adequate preps. Thus, more patients in the appointment group achieved good and adequate preps, and far fewer achieved fair or poor preps. One important point was that almost half (47%) of the e-consult group had only a fair prep (Figure 3).

A number of reasons have been identified in the literature, which might help us understand these findings. First, patients may not fully understand or adhere to prep instructions.^8-10 Furthermore, certain medical diagnoses are known to affect prep quality (ie, diabetes, thyroid disease, constipation).^11,12 Another potential factor is the manner in which prep quality is determined.^13,14 However, due to the focus of this QI study, the influential drivers of prep quality can only be inferred from the literature; thus, a future research or QI study is warranted to ascertain the underlying mechanism of colon prep quality in our specific veteran population.

Preprocedure Reminder Calls

Outcomes were essentially reversed between the 2 groups (Figure 4). Between 50% and 60% of the e-consult group received a call, while the same percentage of appointment patients did not. All patients did attend their procedure appointment. A GD goal was to call every patient before their procedure, but the ability to make the calls was staffing-dependent, which may explain these findings.

Most Relevant Findings

Although this project provides a thorough analysis of various benchmarks within this recently implemented e-consult gastroenterology program, 3 findings emerged that were identified as most relevant. First, the benchmark of 20% volume of completed e-consults was not met. A review of the workflow processes revealed that a daily allotment of only a half-day was not sufficient to complete e-consults, referral triage, and related tasks.

Second, outcomes for colon prep quality and preprocedure reminder calls were also relevant. Although beyond the scope of this project, the question arose of a relationship between prep quality and the reminder call: Does the reminder call affect prep quality? The goal of colon prep is to achieve a good or adequate prep. The purpose of the reminder call was to confirm the appointment and to review the colon prep. Among patients in both groups who achieved only a fair prep, 62% in the e-consult group did receive the reminder call; thus, the call seemed to have failed in helping these patients achieve an adequate or good prep. The actual content of colon prep review during the call was unknown, but certainly bears improvement.

Another speculation concerned the prep instruction sheet. Although all patients received the same sheet, e-consult patients received it in the mail and read it themselves, while it was directly reviewed face-to-face with the appointment patients. Questions remain whether a face-to-face review increases the likelihood for a better prep and how to help e-consult patients achieve optimal prep, since they are not seen face-to-face. 

Practice Implications and Sustaining Measures

Theoretically, e-consults are a viable alternative to face-to-face appointments. Potential advantages include efficient use of an EMR, avoidance of unnecessary appointments, and improved access to care for patients who require an appointment. Although patient satisfaction was not measured in this project, the literature review revealed that satisfaction was increased through use of various virtual health care modalities, including a preliminary analysis in this facility by the aforementioned 2009 study by K. L. Rodriguez (unpublished data). Based on findings in this project, the following 4 recommendations were made to improve benchmark outcomes and quality of care.

1. Provide dedicated time to complete e-consults and related tasks. In this setting, a full day was recommended. An alternative was to hire a NP whose sole responsibility was e-consults.
2. Develop a selection process to determine which new referrals are best suited for e-consult. This process will increase e-consult efficiency and decrease the time to complete an e-consult. Recommended selection criteria were (A) gastroenterology referrals only for simple symptoms or issues; (B) referrals only for a procedure; and (C) stable patients with uncomplicated medical histories.
3. Sustain the preprocedure reminder phone call. The reminder call helps patients keep appointments and thus reduces a missed opportunity for care.
4. Plan a future QI project or research study on patient colon prep quality for colonoscopy. Such a project might evaluate types of colon prep, how prep quality is measured, patient instructions, and the timing/content of pre-procedure reminder phone calls, particularly for e-consult patients.

Conclusion

This QI project provided outcomes for e-consults in the subspecialty of gastroenterology at the VAPHS. Although some benchmark outcomes were met and favorable, others were less favorable. By conducting this benchmark analysis, the areas of needed improvement are now clear. This analysis provides information so recommendations for process improvements can be made.

Quality of care improvement is an ongoing process at VAPHS. Since completion of this project, several processes have been adjusted so that outcomes will be improved. For example, corrective actions were taken for patients who did not complete their gastroenterology procedure. The process for scheduling gastroenterology procedures was adjusted for appointments and cancellations. Ongoing efforts to sustain the reminder phone call were put in place. Changes in NP staffing and time assigned for both clinical and nonclinical work were proposed and are currently under review.

It is the mission of the VA to provide access to care, patient satisfaction, timely care, and appropriate use of resources. Having the ability to highlight our strengths, as well as the willingness to recognize weaknesses, allows us to create new improved processes to provide the best care possible for our veterans.

Acknowledgments
This project was Elena Swann’s capstone for the Doctor of Nursing Practice Program, University of Pittsburgh School of Nursing in Pennsylvania. Rich Laufer, Larry Priscella, and Janie Fleming assisted with the VA Gastroenterology Clinic and procedure metrics. Dr. Melissa Taylor, VA associate chief nurse for research provided project guidance and manuscript revisions. Dr. Sandra Engberg, University of Pittsburgh School of Nursing faculty, assisted with project and manuscript development.

Author disclosures
The author reports no actual or potential conflicts of interest with regard to this article.

Disclaimer
The opinions expressed herein are those of the author and do not necessarily reflect those of Federal Practitioner, Frontline Medical Communications Inc., the U.S. Government, or any of its agencies. This article may discuss unlabeled or investigational use of certain drugs. Please review complete prescribing information for specific drugs or drug combinations—including indications, contraindications, warnings, and adverse effects—before administering pharmacologic therapy to patients.

Actinic keratosis (AK) is regarded as a lesion on a continuum of progression to squamous cell carcinoma (SCC).¹ Studies have estimated that 44% to 97% of SCCs were associated with AK lesions either in contiguous skin or within the same histologic section and that AK lesions progress to SCCs at a rate of 0.6% at 1 year.² In 1993-1994 there were 3.7 million reported office visits for AK lesions, while in 2002 alone there were 8.2 million office visits.^3,4 As the burden of disease from AKs has increased, so has the associated costs from office-based visits, treatments, and subsequent surveillance.

There are a number of highly effective approaches to AK treatment that are based on several factors such as the number of and extent of the lesions, history of skin cancer, provider practice characteristics (eg, location, appointment availability), patient preferences, cost, and tolerability. Cryosurgery is the most commonly used lesion-directed modality in the treatment of individual AKs based on its effectiveness and relative ease of use. Cryosurgery alone has been shown to have a success rate of 67% on AK lesions.⁵ Patients often experience erythema, edema, pain, and crusting at treated sites; there also is potential for ulceration, scarring, hypopigmentation, hyperpigmentation, and secondary infection, but these effects are less common. Recurrence may be an indicator of treatment-resistant lesions or new lesions appearing in the field.

A field-directed approach with topical 5-fluorouracil (5-FU) may be preferred in patients with a history of substantial photodamage, AKs that are resistant to cryosurgery, or multiple AKs. Field-directed treatments address multiple AKs simultaneously and treat subclinical lesions. Fluorouracil is a common therapy for AKs that often is implemented by dermatologists due to its efficacy and well-understood mechanism of action. Fluorouracil inhibits thymidylate synthase during DNA synthesis, thereby halting cellular proliferation. 5-Fluorouracil cream 0.5% has been approved for 1-, 2-, and 4-week treatment periods. In one study, resolution of AK lesions was greatest in the 4-week treatment group; however, side effects also were greatest in this group.⁶ Patients commonly may experience a range of local reactions including erythema, pruritus, erosions, ulcerations, scabbing, crusting, and facial irritation. For patients with substantial photodamage and AKs, a robust response can lead to perceived adverse events (AEs) and considerable downtime, possibly affecting patient satisfaction and treatment compliance.⁷

Many alternative and combination approaches have been studied to decrease AEs and improve compliance and efficacy in the treatment of AKs. In this study, we examined the efficacy and perceived side effects of cryosurgery and 5-FU cream 0.5% combination therapy in the treatment of AKs.

Methods

Study Design and Participants

This single-blind, single-center, comparator cream–controlled pilot study was parallel designed with a balanced randomization (1:1 frequency). The study protocol and consent form were approved by the Wake Forest University Health Sciences institutional review board (Winston-Salem, North Carolina). Participants were 18 years or older with 8 clinically typical, visible, and discrete AK lesions on the face (forehead and temples) or balding scalp. Typical inclusion and exclusion criteria were observed. No other topical agents or therapies were permitted to be applied to the affected areas at least 4 weeks prior to treatment, depending on the treatment modality.

Assessment

During the screening (baseline) visit, eligible participants provided informed consent, baseline lesion counts and investigator global assessments (IGAs) were performed, and cryosurgery was administered to all visible AK lesions in the study areas. Participants returned at weeks 3, 4, 8, and 26. Three weeks following cryosurgery, participants were randomized according to standard randomization tables into 1 of 2 treatment groups to receive once-daily treatment with either 5-FU cream 0.5% or a moisturizing comparator cream. The cream was applied at bedtime to the affected sites for 1 week. Randomization was investigator blinded, but participants and the study administrators were not blinded. Participants were instructed to record their treatment compliance in daily diary entries, which were reviewed at week 4 using the medication tolerability assessment rating for burning, stinging, and ulceration. Investigator global assessment, IGA of improvement, lesion counts, and quality of life (QOL) survey responses were gathered at weeks 3, 8, and 26. The IGA measured the overall severity of AK disease involvement on a 6-point scale (clear; very severe). The IGA of improvement measured the overall improvement from baseline on a 6-point scale (clear; worse). Adverse events were measured at each visit.

Efficacy End Points

The primary end point was 100% clearance of all AK lesions at the end of the study (week 26) relative to the baseline AK lesion count. Secondary end points included comparisons between the groups for the number of participants with greater than 75% reduction of baseline lesion counts at the end of the study as well as differences at each visit in medication tolerability assessments, QOL measures, IGA improvement scores, and medication adherence based on diary entries at week 4.

Statistical Analysis

An intention-to-treat analysis was performed. The number of participants with 100% or greater than 75% clearance of AK lesions by specified time points were compared using relative risks and risk differences with Poisson regression analysis log and identity link functions, respectively, to obtain robust error variance 95% confidence intervals. Medication tolerability assessment, QOL, and IGA improvement scores were compared between the 2 groups using the Mann-Whitney U test. The significance level was set at α=.05. All analyses were performed using SAS data analysis software.

Results

Sixty age-eligible participants were enrolled in the study with 30 participants in each treatment group. All of the participants completed the 26-week study period and were included in the intention-to-treat analysis. All of the participants were white with a median age of 67 years; the median number of baseline AK lesions was 12. Participant baseline demographics and clinical characteristics are provided in Table 1. Treatment compliance in both groups was good with only a few participants reporting missed doses.

In our evaluation of the rate of change in the number of AK lesions at week 8 compared to baseline, the 5-FU cream 0.5% group showed an 84% reduction in the number of AK lesions versus a 69% reduction in the control group. At week 26, the 5-FU cream 0.5% group showed a 72% reduction in the number of lesions versus 73% in the control group. There was no significant difference between 5-FU cream 0.5% and the comparator cream for either 100% or 75% clearance of AK lesions by the end of the study; however, comparing the AK lesion count from baseline to 8 weeks following the initiation of the study, participants in the 5-FU cream 0.5% group were more likely than the control group to achieve 75% or 100% clearance on the relative risk and risk difference scales (Table 2). 

There were no significant differences between the 2 groups for the IGA of improvement at any time point (Table 3). On average, participants in the 5-FU cream 0.5% group experienced more dryness, erosion, fissuring, and redness than the control group but not more ulcerations by the end of week 4 (Table 4). All other QOL measures were statistically comparable between the 2 treatment groups for all time points.

A total of 25 AEs were reported throughout the study but none were considered to be serious. One AE (redness, burning, and itching over the eyebrow) was considered to be related to the study drug. No participants withdrew from the study due to AEs. A total of 12 participants in the 5-FU cream 0.5% group and 10 in the control group reported AEs.

Comment

After a 1-week course of 5-FU cream 0.5% following cryosurgery, a greater reduction in the number of AK lesions for a period of 2 months was noted in the treatment group compared to the control group. These findings are consistent with a similar study from 2006 that used 5-FU cream 0.5% or a vehicle 1 week prior to cryosurgery and then counted the number of AK lesions that remained.⁸ In the 2006 study, remarkable improvement out to week 26 was noted,⁸ unlike our study; however, there was insufficient power in our study to demonstrate a continued effect out to week 26.

Both the 2006 study and our current study support the benefit of using a combination treatment to clear AK lesions versus either treatment alone. Of note, these studies also show that combination treatments are equally effective, regardless of the order of treatments, in lowering AK lesion counts compared to cryosurgery alone.

Although participants in the 5-FU cream 0.5% group reported slightly more AEs on average at week 4, the rate of side effects was lower than those reported in a study documenting the side effects of a 4-week course of 5-FU.⁶ This rate of side effects must be considered in light of the added benefits this combination treatment has demonstrated.

Results of this pilot study suggest that a larger sample size would yield a difference in the study arms for all time periods (weeks 8 and 26). In an effort to maintain exchangeability of the study arms, patients were randomized at baseline treatment, but behaviors of patients in the 6 months following treatment, such as variation in sun exposure or other habits that promote AK lesion development, may have attenuated the results.

Key strengths of this study include no loss to follow-up and high medication adherence rates. The key limitation was the small sample size, which did not demonstrate a statistical advantage of the 5-FU cream 0.5% at 26 weeks; however, our study does show promise for larger future studies in illustrating this difference. A study by Krawtchenko et al⁹ noted that long-term efficacy of field therapy with 5-FU may ultimately be less than imiquimod cream 5%, suggesting that a possible alteration of the study protocol to compare the efficacy of different forms of field therapy may ultimately achieve better outcomes.

Conclusion

Overall, individuals with AK may benefit from a combination of treatment with cryosurgery and topical 5-FU to resolve lesions for longer periods than with cryosurgery alone. Although prior studies have found statistically significant differences in short-term and long-term treatment efficacy when cryosurgery is combined with an active field therapy versus a placebo vehicle,^8,9 the current study aimed to find the best combination of efficacy with the fewest side effects. Therefore, the results of prior literature studies only further the feelings of the authors that with a protocol that looks at a slightly different treatment regimen within the treatment arm, the results can be extremely beneficial to patients. Further studies should be implemented to confirm the longer-term benefits of this combination therapy.

Actinic keratosis (AK) is regarded as a lesion on a continuum of progression to squamous cell carcinoma (SCC).¹ Studies have estimated that 44% to 97% of SCCs were associated with AK lesions either in contiguous skin or within the same histologic section and that AK lesions progress to SCCs at a rate of 0.6% at 1 year.² In 1993-1994 there were 3.7 million reported office visits for AK lesions, while in 2002 alone there were 8.2 million office visits.^3,4 As the burden of disease from AKs has increased, so has the associated costs from office-based visits, treatments, and subsequent surveillance.

There are a number of highly effective approaches to AK treatment that are based on several factors such as the number of and extent of the lesions, history of skin cancer, provider practice characteristics (eg, location, appointment availability), patient preferences, cost, and tolerability. Cryosurgery is the most commonly used lesion-directed modality in the treatment of individual AKs based on its effectiveness and relative ease of use. Cryosurgery alone has been shown to have a success rate of 67% on AK lesions.⁵ Patients often experience erythema, edema, pain, and crusting at treated sites; there also is potential for ulceration, scarring, hypopigmentation, hyperpigmentation, and secondary infection, but these effects are less common. Recurrence may be an indicator of treatment-resistant lesions or new lesions appearing in the field.

A field-directed approach with topical 5-fluorouracil (5-FU) may be preferred in patients with a history of substantial photodamage, AKs that are resistant to cryosurgery, or multiple AKs. Field-directed treatments address multiple AKs simultaneously and treat subclinical lesions. Fluorouracil is a common therapy for AKs that often is implemented by dermatologists due to its efficacy and well-understood mechanism of action. Fluorouracil inhibits thymidylate synthase during DNA synthesis, thereby halting cellular proliferation. 5-Fluorouracil cream 0.5% has been approved for 1-, 2-, and 4-week treatment periods. In one study, resolution of AK lesions was greatest in the 4-week treatment group; however, side effects also were greatest in this group.⁶ Patients commonly may experience a range of local reactions including erythema, pruritus, erosions, ulcerations, scabbing, crusting, and facial irritation. For patients with substantial photodamage and AKs, a robust response can lead to perceived adverse events (AEs) and considerable downtime, possibly affecting patient satisfaction and treatment compliance.⁷

Many alternative and combination approaches have been studied to decrease AEs and improve compliance and efficacy in the treatment of AKs. In this study, we examined the efficacy and perceived side effects of cryosurgery and 5-FU cream 0.5% combination therapy in the treatment of AKs.

Methods

Study Design and Participants

This single-blind, single-center, comparator cream–controlled pilot study was parallel designed with a balanced randomization (1:1 frequency). The study protocol and consent form were approved by the Wake Forest University Health Sciences institutional review board (Winston-Salem, North Carolina). Participants were 18 years or older with 8 clinically typical, visible, and discrete AK lesions on the face (forehead and temples) or balding scalp. Typical inclusion and exclusion criteria were observed. No other topical agents or therapies were permitted to be applied to the affected areas at least 4 weeks prior to treatment, depending on the treatment modality.

Assessment

During the screening (baseline) visit, eligible participants provided informed consent, baseline lesion counts and investigator global assessments (IGAs) were performed, and cryosurgery was administered to all visible AK lesions in the study areas. Participants returned at weeks 3, 4, 8, and 26. Three weeks following cryosurgery, participants were randomized according to standard randomization tables into 1 of 2 treatment groups to receive once-daily treatment with either 5-FU cream 0.5% or a moisturizing comparator cream. The cream was applied at bedtime to the affected sites for 1 week. Randomization was investigator blinded, but participants and the study administrators were not blinded. Participants were instructed to record their treatment compliance in daily diary entries, which were reviewed at week 4 using the medication tolerability assessment rating for burning, stinging, and ulceration. Investigator global assessment, IGA of improvement, lesion counts, and quality of life (QOL) survey responses were gathered at weeks 3, 8, and 26. The IGA measured the overall severity of AK disease involvement on a 6-point scale (clear; very severe). The IGA of improvement measured the overall improvement from baseline on a 6-point scale (clear; worse). Adverse events were measured at each visit.

Efficacy End Points

The primary end point was 100% clearance of all AK lesions at the end of the study (week 26) relative to the baseline AK lesion count. Secondary end points included comparisons between the groups for the number of participants with greater than 75% reduction of baseline lesion counts at the end of the study as well as differences at each visit in medication tolerability assessments, QOL measures, IGA improvement scores, and medication adherence based on diary entries at week 4.

Statistical Analysis

An intention-to-treat analysis was performed. The number of participants with 100% or greater than 75% clearance of AK lesions by specified time points were compared using relative risks and risk differences with Poisson regression analysis log and identity link functions, respectively, to obtain robust error variance 95% confidence intervals. Medication tolerability assessment, QOL, and IGA improvement scores were compared between the 2 groups using the Mann-Whitney U test. The significance level was set at α=.05. All analyses were performed using SAS data analysis software.

Results

Sixty age-eligible participants were enrolled in the study with 30 participants in each treatment group. All of the participants completed the 26-week study period and were included in the intention-to-treat analysis. All of the participants were white with a median age of 67 years; the median number of baseline AK lesions was 12. Participant baseline demographics and clinical characteristics are provided in Table 1. Treatment compliance in both groups was good with only a few participants reporting missed doses.

In our evaluation of the rate of change in the number of AK lesions at week 8 compared to baseline, the 5-FU cream 0.5% group showed an 84% reduction in the number of AK lesions versus a 69% reduction in the control group. At week 26, the 5-FU cream 0.5% group showed a 72% reduction in the number of lesions versus 73% in the control group. There was no significant difference between 5-FU cream 0.5% and the comparator cream for either 100% or 75% clearance of AK lesions by the end of the study; however, comparing the AK lesion count from baseline to 8 weeks following the initiation of the study, participants in the 5-FU cream 0.5% group were more likely than the control group to achieve 75% or 100% clearance on the relative risk and risk difference scales (Table 2). 

There were no significant differences between the 2 groups for the IGA of improvement at any time point (Table 3). On average, participants in the 5-FU cream 0.5% group experienced more dryness, erosion, fissuring, and redness than the control group but not more ulcerations by the end of week 4 (Table 4). All other QOL measures were statistically comparable between the 2 treatment groups for all time points.

A total of 25 AEs were reported throughout the study but none were considered to be serious. One AE (redness, burning, and itching over the eyebrow) was considered to be related to the study drug. No participants withdrew from the study due to AEs. A total of 12 participants in the 5-FU cream 0.5% group and 10 in the control group reported AEs.

Comment

After a 1-week course of 5-FU cream 0.5% following cryosurgery, a greater reduction in the number of AK lesions for a period of 2 months was noted in the treatment group compared to the control group. These findings are consistent with a similar study from 2006 that used 5-FU cream 0.5% or a vehicle 1 week prior to cryosurgery and then counted the number of AK lesions that remained.⁸ In the 2006 study, remarkable improvement out to week 26 was noted,⁸ unlike our study; however, there was insufficient power in our study to demonstrate a continued effect out to week 26.

Both the 2006 study and our current study support the benefit of using a combination treatment to clear AK lesions versus either treatment alone. Of note, these studies also show that combination treatments are equally effective, regardless of the order of treatments, in lowering AK lesion counts compared to cryosurgery alone.

Although participants in the 5-FU cream 0.5% group reported slightly more AEs on average at week 4, the rate of side effects was lower than those reported in a study documenting the side effects of a 4-week course of 5-FU.⁶ This rate of side effects must be considered in light of the added benefits this combination treatment has demonstrated.

Results of this pilot study suggest that a larger sample size would yield a difference in the study arms for all time periods (weeks 8 and 26). In an effort to maintain exchangeability of the study arms, patients were randomized at baseline treatment, but behaviors of patients in the 6 months following treatment, such as variation in sun exposure or other habits that promote AK lesion development, may have attenuated the results.

Key strengths of this study include no loss to follow-up and high medication adherence rates. The key limitation was the small sample size, which did not demonstrate a statistical advantage of the 5-FU cream 0.5% at 26 weeks; however, our study does show promise for larger future studies in illustrating this difference. A study by Krawtchenko et al⁹ noted that long-term efficacy of field therapy with 5-FU may ultimately be less than imiquimod cream 5%, suggesting that a possible alteration of the study protocol to compare the efficacy of different forms of field therapy may ultimately achieve better outcomes.

Conclusion

Overall, individuals with AK may benefit from a combination of treatment with cryosurgery and topical 5-FU to resolve lesions for longer periods than with cryosurgery alone. Although prior studies have found statistically significant differences in short-term and long-term treatment efficacy when cryosurgery is combined with an active field therapy versus a placebo vehicle,^8,9 the current study aimed to find the best combination of efficacy with the fewest side effects. Therefore, the results of prior literature studies only further the feelings of the authors that with a protocol that looks at a slightly different treatment regimen within the treatment arm, the results can be extremely beneficial to patients. Further studies should be implemented to confirm the longer-term benefits of this combination therapy.

Actinic keratosis (AK) is regarded as a lesion on a continuum of progression to squamous cell carcinoma (SCC).¹ Studies have estimated that 44% to 97% of SCCs were associated with AK lesions either in contiguous skin or within the same histologic section and that AK lesions progress to SCCs at a rate of 0.6% at 1 year.² In 1993-1994 there were 3.7 million reported office visits for AK lesions, while in 2002 alone there were 8.2 million office visits.^3,4 As the burden of disease from AKs has increased, so has the associated costs from office-based visits, treatments, and subsequent surveillance.

There are a number of highly effective approaches to AK treatment that are based on several factors such as the number of and extent of the lesions, history of skin cancer, provider practice characteristics (eg, location, appointment availability), patient preferences, cost, and tolerability. Cryosurgery is the most commonly used lesion-directed modality in the treatment of individual AKs based on its effectiveness and relative ease of use. Cryosurgery alone has been shown to have a success rate of 67% on AK lesions.⁵ Patients often experience erythema, edema, pain, and crusting at treated sites; there also is potential for ulceration, scarring, hypopigmentation, hyperpigmentation, and secondary infection, but these effects are less common. Recurrence may be an indicator of treatment-resistant lesions or new lesions appearing in the field.

A field-directed approach with topical 5-fluorouracil (5-FU) may be preferred in patients with a history of substantial photodamage, AKs that are resistant to cryosurgery, or multiple AKs. Field-directed treatments address multiple AKs simultaneously and treat subclinical lesions. Fluorouracil is a common therapy for AKs that often is implemented by dermatologists due to its efficacy and well-understood mechanism of action. Fluorouracil inhibits thymidylate synthase during DNA synthesis, thereby halting cellular proliferation. 5-Fluorouracil cream 0.5% has been approved for 1-, 2-, and 4-week treatment periods. In one study, resolution of AK lesions was greatest in the 4-week treatment group; however, side effects also were greatest in this group.⁶ Patients commonly may experience a range of local reactions including erythema, pruritus, erosions, ulcerations, scabbing, crusting, and facial irritation. For patients with substantial photodamage and AKs, a robust response can lead to perceived adverse events (AEs) and considerable downtime, possibly affecting patient satisfaction and treatment compliance.⁷

Many alternative and combination approaches have been studied to decrease AEs and improve compliance and efficacy in the treatment of AKs. In this study, we examined the efficacy and perceived side effects of cryosurgery and 5-FU cream 0.5% combination therapy in the treatment of AKs.

Methods

Study Design and Participants

This single-blind, single-center, comparator cream–controlled pilot study was parallel designed with a balanced randomization (1:1 frequency). The study protocol and consent form were approved by the Wake Forest University Health Sciences institutional review board (Winston-Salem, North Carolina). Participants were 18 years or older with 8 clinically typical, visible, and discrete AK lesions on the face (forehead and temples) or balding scalp. Typical inclusion and exclusion criteria were observed. No other topical agents or therapies were permitted to be applied to the affected areas at least 4 weeks prior to treatment, depending on the treatment modality.

Assessment

During the screening (baseline) visit, eligible participants provided informed consent, baseline lesion counts and investigator global assessments (IGAs) were performed, and cryosurgery was administered to all visible AK lesions in the study areas. Participants returned at weeks 3, 4, 8, and 26. Three weeks following cryosurgery, participants were randomized according to standard randomization tables into 1 of 2 treatment groups to receive once-daily treatment with either 5-FU cream 0.5% or a moisturizing comparator cream. The cream was applied at bedtime to the affected sites for 1 week. Randomization was investigator blinded, but participants and the study administrators were not blinded. Participants were instructed to record their treatment compliance in daily diary entries, which were reviewed at week 4 using the medication tolerability assessment rating for burning, stinging, and ulceration. Investigator global assessment, IGA of improvement, lesion counts, and quality of life (QOL) survey responses were gathered at weeks 3, 8, and 26. The IGA measured the overall severity of AK disease involvement on a 6-point scale (clear; very severe). The IGA of improvement measured the overall improvement from baseline on a 6-point scale (clear; worse). Adverse events were measured at each visit.

Efficacy End Points

The primary end point was 100% clearance of all AK lesions at the end of the study (week 26) relative to the baseline AK lesion count. Secondary end points included comparisons between the groups for the number of participants with greater than 75% reduction of baseline lesion counts at the end of the study as well as differences at each visit in medication tolerability assessments, QOL measures, IGA improvement scores, and medication adherence based on diary entries at week 4.

Statistical Analysis

An intention-to-treat analysis was performed. The number of participants with 100% or greater than 75% clearance of AK lesions by specified time points were compared using relative risks and risk differences with Poisson regression analysis log and identity link functions, respectively, to obtain robust error variance 95% confidence intervals. Medication tolerability assessment, QOL, and IGA improvement scores were compared between the 2 groups using the Mann-Whitney U test. The significance level was set at α=.05. All analyses were performed using SAS data analysis software.

Results

Sixty age-eligible participants were enrolled in the study with 30 participants in each treatment group. All of the participants completed the 26-week study period and were included in the intention-to-treat analysis. All of the participants were white with a median age of 67 years; the median number of baseline AK lesions was 12. Participant baseline demographics and clinical characteristics are provided in Table 1. Treatment compliance in both groups was good with only a few participants reporting missed doses.

In our evaluation of the rate of change in the number of AK lesions at week 8 compared to baseline, the 5-FU cream 0.5% group showed an 84% reduction in the number of AK lesions versus a 69% reduction in the control group. At week 26, the 5-FU cream 0.5% group showed a 72% reduction in the number of lesions versus 73% in the control group. There was no significant difference between 5-FU cream 0.5% and the comparator cream for either 100% or 75% clearance of AK lesions by the end of the study; however, comparing the AK lesion count from baseline to 8 weeks following the initiation of the study, participants in the 5-FU cream 0.5% group were more likely than the control group to achieve 75% or 100% clearance on the relative risk and risk difference scales (Table 2). 

There were no significant differences between the 2 groups for the IGA of improvement at any time point (Table 3). On average, participants in the 5-FU cream 0.5% group experienced more dryness, erosion, fissuring, and redness than the control group but not more ulcerations by the end of week 4 (Table 4). All other QOL measures were statistically comparable between the 2 treatment groups for all time points.

A total of 25 AEs were reported throughout the study but none were considered to be serious. One AE (redness, burning, and itching over the eyebrow) was considered to be related to the study drug. No participants withdrew from the study due to AEs. A total of 12 participants in the 5-FU cream 0.5% group and 10 in the control group reported AEs.

Comment

After a 1-week course of 5-FU cream 0.5% following cryosurgery, a greater reduction in the number of AK lesions for a period of 2 months was noted in the treatment group compared to the control group. These findings are consistent with a similar study from 2006 that used 5-FU cream 0.5% or a vehicle 1 week prior to cryosurgery and then counted the number of AK lesions that remained.⁸ In the 2006 study, remarkable improvement out to week 26 was noted,⁸ unlike our study; however, there was insufficient power in our study to demonstrate a continued effect out to week 26.

Both the 2006 study and our current study support the benefit of using a combination treatment to clear AK lesions versus either treatment alone. Of note, these studies also show that combination treatments are equally effective, regardless of the order of treatments, in lowering AK lesion counts compared to cryosurgery alone.

Although participants in the 5-FU cream 0.5% group reported slightly more AEs on average at week 4, the rate of side effects was lower than those reported in a study documenting the side effects of a 4-week course of 5-FU.⁶ This rate of side effects must be considered in light of the added benefits this combination treatment has demonstrated.

Results of this pilot study suggest that a larger sample size would yield a difference in the study arms for all time periods (weeks 8 and 26). In an effort to maintain exchangeability of the study arms, patients were randomized at baseline treatment, but behaviors of patients in the 6 months following treatment, such as variation in sun exposure or other habits that promote AK lesion development, may have attenuated the results.

Key strengths of this study include no loss to follow-up and high medication adherence rates. The key limitation was the small sample size, which did not demonstrate a statistical advantage of the 5-FU cream 0.5% at 26 weeks; however, our study does show promise for larger future studies in illustrating this difference. A study by Krawtchenko et al⁹ noted that long-term efficacy of field therapy with 5-FU may ultimately be less than imiquimod cream 5%, suggesting that a possible alteration of the study protocol to compare the efficacy of different forms of field therapy may ultimately achieve better outcomes.

Conclusion

Overall, individuals with AK may benefit from a combination of treatment with cryosurgery and topical 5-FU to resolve lesions for longer periods than with cryosurgery alone. Although prior studies have found statistically significant differences in short-term and long-term treatment efficacy when cryosurgery is combined with an active field therapy versus a placebo vehicle,^8,9 the current study aimed to find the best combination of efficacy with the fewest side effects. Therefore, the results of prior literature studies only further the feelings of the authors that with a protocol that looks at a slightly different treatment regimen within the treatment arm, the results can be extremely beneficial to patients. Further studies should be implemented to confirm the longer-term benefits of this combination therapy.