Rotavirus gastroenteritis (RGE) accounts for approximately 70,000 pediatric hospitalizations annually in the United States.1 Costly microbiological assays are frequently performed in these patients to exclude concurrent serious bacterial infection (SBI), though the actual incidence of SBI is quite low.28 Our objectives were to describe the incidence of SBI in children evaluated at a community hospital and subsequently diagnosed with laboratory‐confirmed RGE and to determine whether ancillary testing was associated with prolonged length of stay (LOS) in hospitalized patients.

Materials and Methods

Results

One hundred cases of RGE were initially identified; 6 patients were excluded4 with negative rotavirus stool antigen tests and 2 because the infection was nosocomially‐acquired. The remaining 94 cases were included in the analysis. Fifty‐eight (61.7%) of the patients were male, and 80 (85.1%) were African‐American. The median age was 8 months (IQR, 1 month to 16 years) and 83 patients (88.3%) were admitted to hospital. Fifty patients (53.2%) were febrile at presentation. The median length of stay was 2 days (IQR, 1‐3 days).

There were no patients with SBI (95% confidence interval [CI], 0%‐3.8%). Ten patients (12%) had received antibiotics in the 72 hours prior to evaluation; 6 of these 10 patients had blood cultures obtained. Peripheral blood cultures were drawn from 47 patients (50%). Of these, 43 (91.5%) were negative. Three cultures yielded viridans group streptococci, and 1 culture yielded vancomycin‐resistant Enterococcus species (VRE). The cultures yielding viridans group streptococci were drawn from 3 infants aged 42 days, 4 months, and 12 months. All 3 infants were febrile at presentation. In 2 of the 3 infants, 2 sets of blood cultures were drawn and viridans group streptococci was isolated from only 1 of the 2 cultures. The third infant made a rapid clinical recovery without antibiotic intervention and was discharged in less than 48 hours, belying microbiological evidence of bacteremia. Therefore, we classified all 3 viridans group streptococci cultures as contaminated specimens. The difference in the frequency with which blood cultures were performed in children younger than (59%) or older than (44%) 6 months of age was not statistically significant (², P = 0.143).

The patient with VRE isolated from blood culture was a 4‐month‐old male who presented with 2 days of vomiting and diarrhea and a fever to 38.7C. The VRE culture, while potentially representing bacterial translocation in the setting of RGE, was presumed to be a contaminant when a repeat peripheral culture was negative. The patient had received amoxicillin for the treatment of otitis media prior to presentation and acquisition of cultures. The susceptibility testing results for ampicillin or amoxicillin were not available; however, the patient did not receive antibiotics for treatment of the VRE blood culture isolate.

Multiple microbiological assays were performed (Figure 1). Many of the detected organisms were considered nonpathogenic. Stool bacterial cultures were obtained in 76 patients (80.9%) with only 1 (1.3%) positive isolate, Proteus mirabilis, considered nonpathogenic. Urine cultures from 41 patients (43.6%) yielded only 1 (2.4%) positive result, Staphylococcus aureus, deemed a contaminant. Nasopharyngeal washes from 15 patients (16%) revealed 3 (20%) positive results (respiratory syncytial virus in 2 patients and influenza virus in 1). Stool assayed for ova and parasites in 9 patients (9.6%) was negative. CSF cultured in 9 patients was also negative, although 3 samples demonstrated pleocytosis. Nonmicrobiological assays included 4 normal chest radiographs, 2 normal urinalyses, and 3 arterial blood gases revealing metabolic acidosis.

Figure 1

A complete blood count was obtained in 77 patients (81.9%). The median peripheral white blood cell count was 8800/mm³ (IQR, 6800 to 11,800). There were no differences between those with and without prolonged LOS on univariate analysis with regard to vital signs or initial symptoms such as tachypnea, fever, tachycardia, or other features associated with illness severity (eg, extent of dehydration). There were no differences in hematological or chemical parameters or with the performance of any other testing. In bivariate analyses, age 6 months (unadjusted OR, 3.43; 95% CI, 1.26‐9.50; P < 0.01) and collection of peripheral blood culture (OR, 3.12; 95% CI, 1.13‐8.98; P < 0.01) were associated with prolonged LOS. Other variables considered for inclusion in the multivariable model included duration of symptoms, presence of a preexisting medical condition, and performance of a nasopharyngeal wash for respiratory virus detection. In multivariable analysis, age <6 months (adjusted OR, 3.01; 95% CI, 1.17‐7.74; P = 0.022) and the performance of a blood culture (adjusted OR, 2.71; 95% CI, 1.03‐7.13; P = 0.043) were independently associated with a prolonged LOS.

Discussion

The absence of SBI in our relatively small cohort of children admitted to a community hospital with laboratory‐confirmed RGE supports earlier estimates of an incidence of less than 1%,5, 7 an incidence similar to that of occult bacteremia in febrile children 2 to 36 months of age following introduction of the heptavalent pneumococcal conjugate vaccine in 2000.10, 11 We found 13 cases reported in the English literature (Table 1). Several salient features are noted when comparing these case reports. All cases of SBI following laboratory‐confirmed RGE were characterized by the development of a second fever after the resolution of initial symptoms. These fevers presented at a mean day of hospitalization of 2.8 (range, 2‐5). Second fevers were high (mean, 39.2C; range, 38.2C to 40C). Cultures obtained other than peripheral blood cultures were only positive in 1 patient; this patient also had cellulitis and Escherichia coli was isolated from both blood and wound cultures.3 One of the reported children with bacteremia died, 2 cases of SBI following RGE were complicated by disseminated intravascular coagulopathy, and 1 case by acute renal failure. Enterobacter cloacae (n = 4) and Klebsiella pneumoniae (n = 3) were the most commonly isolated organisms from peripheral blood culture.

Many children in our study had ancillary laboratory testing performed. The results of these tests were typically normal and rarely affected clinical management in a positive manner. Bacteria and parasites are relatively rare causes of gastroenteritis in the United States in comparison with rotavirus, particularly during the winter months. However, stool was sent for bacterial culture in over 80% of patients and for ova and parasite detection in almost 10% of patients ultimately diagnosed with RGE. Furthermore, despite the relatively low prevalence of bacteremia since licensure of the Haemophilus influenzae type b vaccine, a majority of children had a complete blood count performed while one‐half also had blood obtained for culture. In our cohort, children 6 months and younger and those from whom a blood culture was collected were at an increased risk for prolonged LOS. It was not clear from medical record review whether children with prolonged LOS were kept in the hospital longer for the sole purpose of awaiting the results of blood cultures.

SBI rarely occurs in the context of RGE. While secondary fever seems to be a common manifestation, the sensitivity of secondary fever as a marker for SBI after RGE in this population is unknown. However, given the very low incidence, the potentially serious complications of SBI following laboratory confirmed RGE, and the likely successful management of these complications in the hospital setting, slightly longer hospitalizations for children under 1 year of age must be weighed against earlier discharges with instructions from clinicians to caregivers for careful monitoring of fever and outpatient follow‐up shortly after discharge.

This study has several limitations. First, the timing of the availability of the results of rotavirus antigen testing is not known. It is possible that the rapid availability of rotavirus test results in some circumstances encouraged clinicians to abandon tests seeking other sources of infection. Conversely, children with gastroenteritis in the context of a concurrent bacterial infection may have been less likely to undergo rotavirus stool antigen testing. This latter possibility would bias our findings toward underestimating the prevalence of concurrent bacterial infection among children with RGE. Second, this study was performed prior to licensure and widespread use of the currently‐licensed vaccine against rotavirus (Rotateq; Merck and Company, Whitehouse Station, NJ). Reductions in the burden of gastroenteritis caused by rotavirus may have a much more dramatic impact on resource utilization in the treatment of gastroenteritis than reductions in ancillary testing. Finally, this study was performed at a single urban community hospital and therefore cannot be generalized to other settings such as academic tertiary care centers. Furthermore, test ordering patterns may be local or regional and other community hospitals may exhibit different patterns. Further clarification of the role of ancillary testing in children presenting with diarrhea during the winter months is warranted since reducing the extent of such testing would dramatically reduce resource utilization for this illness. Finally, a blood culture was not obtained from all patients. Therefore, occult bacteremia attributable to RGE could not be detected. Since no patient in our study underwent subsequent clinical deterioration, we presume that any case of occult bacteremia resolved spontaneously and was not of clinical consequence, although such occurrences would cause us to underestimate the prevalence of SBI in this population.

Resource utilization in our cohort was high, while yield from microbiological investigations was low. This finding challenges the need to perform invasive, costly assays to exclude concurrent SBI in this population. It is possible that children with viral gastroenteritis caused by pathogens other than rotavirus are also at low risk of SBI. However, the diagnostic strategy that best identifies patients at risk for SBI following acute gastroenteritis is unknown. Further studies are needed to determine an ideal clinical approach to the infant with RGE.

Rotavirus gastroenteritis (RGE) accounts for approximately 70,000 pediatric hospitalizations annually in the United States.1 Costly microbiological assays are frequently performed in these patients to exclude concurrent serious bacterial infection (SBI), though the actual incidence of SBI is quite low.28 Our objectives were to describe the incidence of SBI in children evaluated at a community hospital and subsequently diagnosed with laboratory‐confirmed RGE and to determine whether ancillary testing was associated with prolonged length of stay (LOS) in hospitalized patients.

Materials and Methods

Results

One hundred cases of RGE were initially identified; 6 patients were excluded4 with negative rotavirus stool antigen tests and 2 because the infection was nosocomially‐acquired. The remaining 94 cases were included in the analysis. Fifty‐eight (61.7%) of the patients were male, and 80 (85.1%) were African‐American. The median age was 8 months (IQR, 1 month to 16 years) and 83 patients (88.3%) were admitted to hospital. Fifty patients (53.2%) were febrile at presentation. The median length of stay was 2 days (IQR, 1‐3 days).

There were no patients with SBI (95% confidence interval [CI], 0%‐3.8%). Ten patients (12%) had received antibiotics in the 72 hours prior to evaluation; 6 of these 10 patients had blood cultures obtained. Peripheral blood cultures were drawn from 47 patients (50%). Of these, 43 (91.5%) were negative. Three cultures yielded viridans group streptococci, and 1 culture yielded vancomycin‐resistant Enterococcus species (VRE). The cultures yielding viridans group streptococci were drawn from 3 infants aged 42 days, 4 months, and 12 months. All 3 infants were febrile at presentation. In 2 of the 3 infants, 2 sets of blood cultures were drawn and viridans group streptococci was isolated from only 1 of the 2 cultures. The third infant made a rapid clinical recovery without antibiotic intervention and was discharged in less than 48 hours, belying microbiological evidence of bacteremia. Therefore, we classified all 3 viridans group streptococci cultures as contaminated specimens. The difference in the frequency with which blood cultures were performed in children younger than (59%) or older than (44%) 6 months of age was not statistically significant (², P = 0.143).

The patient with VRE isolated from blood culture was a 4‐month‐old male who presented with 2 days of vomiting and diarrhea and a fever to 38.7C. The VRE culture, while potentially representing bacterial translocation in the setting of RGE, was presumed to be a contaminant when a repeat peripheral culture was negative. The patient had received amoxicillin for the treatment of otitis media prior to presentation and acquisition of cultures. The susceptibility testing results for ampicillin or amoxicillin were not available; however, the patient did not receive antibiotics for treatment of the VRE blood culture isolate.

Multiple microbiological assays were performed (Figure 1). Many of the detected organisms were considered nonpathogenic. Stool bacterial cultures were obtained in 76 patients (80.9%) with only 1 (1.3%) positive isolate, Proteus mirabilis, considered nonpathogenic. Urine cultures from 41 patients (43.6%) yielded only 1 (2.4%) positive result, Staphylococcus aureus, deemed a contaminant. Nasopharyngeal washes from 15 patients (16%) revealed 3 (20%) positive results (respiratory syncytial virus in 2 patients and influenza virus in 1). Stool assayed for ova and parasites in 9 patients (9.6%) was negative. CSF cultured in 9 patients was also negative, although 3 samples demonstrated pleocytosis. Nonmicrobiological assays included 4 normal chest radiographs, 2 normal urinalyses, and 3 arterial blood gases revealing metabolic acidosis.

Figure 1

A complete blood count was obtained in 77 patients (81.9%). The median peripheral white blood cell count was 8800/mm³ (IQR, 6800 to 11,800). There were no differences between those with and without prolonged LOS on univariate analysis with regard to vital signs or initial symptoms such as tachypnea, fever, tachycardia, or other features associated with illness severity (eg, extent of dehydration). There were no differences in hematological or chemical parameters or with the performance of any other testing. In bivariate analyses, age 6 months (unadjusted OR, 3.43; 95% CI, 1.26‐9.50; P < 0.01) and collection of peripheral blood culture (OR, 3.12; 95% CI, 1.13‐8.98; P < 0.01) were associated with prolonged LOS. Other variables considered for inclusion in the multivariable model included duration of symptoms, presence of a preexisting medical condition, and performance of a nasopharyngeal wash for respiratory virus detection. In multivariable analysis, age <6 months (adjusted OR, 3.01; 95% CI, 1.17‐7.74; P = 0.022) and the performance of a blood culture (adjusted OR, 2.71; 95% CI, 1.03‐7.13; P = 0.043) were independently associated with a prolonged LOS.

Discussion

The absence of SBI in our relatively small cohort of children admitted to a community hospital with laboratory‐confirmed RGE supports earlier estimates of an incidence of less than 1%,5, 7 an incidence similar to that of occult bacteremia in febrile children 2 to 36 months of age following introduction of the heptavalent pneumococcal conjugate vaccine in 2000.10, 11 We found 13 cases reported in the English literature (Table 1). Several salient features are noted when comparing these case reports. All cases of SBI following laboratory‐confirmed RGE were characterized by the development of a second fever after the resolution of initial symptoms. These fevers presented at a mean day of hospitalization of 2.8 (range, 2‐5). Second fevers were high (mean, 39.2C; range, 38.2C to 40C). Cultures obtained other than peripheral blood cultures were only positive in 1 patient; this patient also had cellulitis and Escherichia coli was isolated from both blood and wound cultures.3 One of the reported children with bacteremia died, 2 cases of SBI following RGE were complicated by disseminated intravascular coagulopathy, and 1 case by acute renal failure. Enterobacter cloacae (n = 4) and Klebsiella pneumoniae (n = 3) were the most commonly isolated organisms from peripheral blood culture.

Many children in our study had ancillary laboratory testing performed. The results of these tests were typically normal and rarely affected clinical management in a positive manner. Bacteria and parasites are relatively rare causes of gastroenteritis in the United States in comparison with rotavirus, particularly during the winter months. However, stool was sent for bacterial culture in over 80% of patients and for ova and parasite detection in almost 10% of patients ultimately diagnosed with RGE. Furthermore, despite the relatively low prevalence of bacteremia since licensure of the Haemophilus influenzae type b vaccine, a majority of children had a complete blood count performed while one‐half also had blood obtained for culture. In our cohort, children 6 months and younger and those from whom a blood culture was collected were at an increased risk for prolonged LOS. It was not clear from medical record review whether children with prolonged LOS were kept in the hospital longer for the sole purpose of awaiting the results of blood cultures.

SBI rarely occurs in the context of RGE. While secondary fever seems to be a common manifestation, the sensitivity of secondary fever as a marker for SBI after RGE in this population is unknown. However, given the very low incidence, the potentially serious complications of SBI following laboratory confirmed RGE, and the likely successful management of these complications in the hospital setting, slightly longer hospitalizations for children under 1 year of age must be weighed against earlier discharges with instructions from clinicians to caregivers for careful monitoring of fever and outpatient follow‐up shortly after discharge.

This study has several limitations. First, the timing of the availability of the results of rotavirus antigen testing is not known. It is possible that the rapid availability of rotavirus test results in some circumstances encouraged clinicians to abandon tests seeking other sources of infection. Conversely, children with gastroenteritis in the context of a concurrent bacterial infection may have been less likely to undergo rotavirus stool antigen testing. This latter possibility would bias our findings toward underestimating the prevalence of concurrent bacterial infection among children with RGE. Second, this study was performed prior to licensure and widespread use of the currently‐licensed vaccine against rotavirus (Rotateq; Merck and Company, Whitehouse Station, NJ). Reductions in the burden of gastroenteritis caused by rotavirus may have a much more dramatic impact on resource utilization in the treatment of gastroenteritis than reductions in ancillary testing. Finally, this study was performed at a single urban community hospital and therefore cannot be generalized to other settings such as academic tertiary care centers. Furthermore, test ordering patterns may be local or regional and other community hospitals may exhibit different patterns. Further clarification of the role of ancillary testing in children presenting with diarrhea during the winter months is warranted since reducing the extent of such testing would dramatically reduce resource utilization for this illness. Finally, a blood culture was not obtained from all patients. Therefore, occult bacteremia attributable to RGE could not be detected. Since no patient in our study underwent subsequent clinical deterioration, we presume that any case of occult bacteremia resolved spontaneously and was not of clinical consequence, although such occurrences would cause us to underestimate the prevalence of SBI in this population.

Resource utilization in our cohort was high, while yield from microbiological investigations was low. This finding challenges the need to perform invasive, costly assays to exclude concurrent SBI in this population. It is possible that children with viral gastroenteritis caused by pathogens other than rotavirus are also at low risk of SBI. However, the diagnostic strategy that best identifies patients at risk for SBI following acute gastroenteritis is unknown. Further studies are needed to determine an ideal clinical approach to the infant with RGE.

Rotavirus gastroenteritis (RGE) accounts for approximately 70,000 pediatric hospitalizations annually in the United States.1 Costly microbiological assays are frequently performed in these patients to exclude concurrent serious bacterial infection (SBI), though the actual incidence of SBI is quite low.28 Our objectives were to describe the incidence of SBI in children evaluated at a community hospital and subsequently diagnosed with laboratory‐confirmed RGE and to determine whether ancillary testing was associated with prolonged length of stay (LOS) in hospitalized patients.

Materials and Methods

Results

One hundred cases of RGE were initially identified; 6 patients were excluded4 with negative rotavirus stool antigen tests and 2 because the infection was nosocomially‐acquired. The remaining 94 cases were included in the analysis. Fifty‐eight (61.7%) of the patients were male, and 80 (85.1%) were African‐American. The median age was 8 months (IQR, 1 month to 16 years) and 83 patients (88.3%) were admitted to hospital. Fifty patients (53.2%) were febrile at presentation. The median length of stay was 2 days (IQR, 1‐3 days).

There were no patients with SBI (95% confidence interval [CI], 0%‐3.8%). Ten patients (12%) had received antibiotics in the 72 hours prior to evaluation; 6 of these 10 patients had blood cultures obtained. Peripheral blood cultures were drawn from 47 patients (50%). Of these, 43 (91.5%) were negative. Three cultures yielded viridans group streptococci, and 1 culture yielded vancomycin‐resistant Enterococcus species (VRE). The cultures yielding viridans group streptococci were drawn from 3 infants aged 42 days, 4 months, and 12 months. All 3 infants were febrile at presentation. In 2 of the 3 infants, 2 sets of blood cultures were drawn and viridans group streptococci was isolated from only 1 of the 2 cultures. The third infant made a rapid clinical recovery without antibiotic intervention and was discharged in less than 48 hours, belying microbiological evidence of bacteremia. Therefore, we classified all 3 viridans group streptococci cultures as contaminated specimens. The difference in the frequency with which blood cultures were performed in children younger than (59%) or older than (44%) 6 months of age was not statistically significant (², P = 0.143).

The patient with VRE isolated from blood culture was a 4‐month‐old male who presented with 2 days of vomiting and diarrhea and a fever to 38.7C. The VRE culture, while potentially representing bacterial translocation in the setting of RGE, was presumed to be a contaminant when a repeat peripheral culture was negative. The patient had received amoxicillin for the treatment of otitis media prior to presentation and acquisition of cultures. The susceptibility testing results for ampicillin or amoxicillin were not available; however, the patient did not receive antibiotics for treatment of the VRE blood culture isolate.

Multiple microbiological assays were performed (Figure 1). Many of the detected organisms were considered nonpathogenic. Stool bacterial cultures were obtained in 76 patients (80.9%) with only 1 (1.3%) positive isolate, Proteus mirabilis, considered nonpathogenic. Urine cultures from 41 patients (43.6%) yielded only 1 (2.4%) positive result, Staphylococcus aureus, deemed a contaminant. Nasopharyngeal washes from 15 patients (16%) revealed 3 (20%) positive results (respiratory syncytial virus in 2 patients and influenza virus in 1). Stool assayed for ova and parasites in 9 patients (9.6%) was negative. CSF cultured in 9 patients was also negative, although 3 samples demonstrated pleocytosis. Nonmicrobiological assays included 4 normal chest radiographs, 2 normal urinalyses, and 3 arterial blood gases revealing metabolic acidosis.

Figure 1

A complete blood count was obtained in 77 patients (81.9%). The median peripheral white blood cell count was 8800/mm³ (IQR, 6800 to 11,800). There were no differences between those with and without prolonged LOS on univariate analysis with regard to vital signs or initial symptoms such as tachypnea, fever, tachycardia, or other features associated with illness severity (eg, extent of dehydration). There were no differences in hematological or chemical parameters or with the performance of any other testing. In bivariate analyses, age 6 months (unadjusted OR, 3.43; 95% CI, 1.26‐9.50; P < 0.01) and collection of peripheral blood culture (OR, 3.12; 95% CI, 1.13‐8.98; P < 0.01) were associated with prolonged LOS. Other variables considered for inclusion in the multivariable model included duration of symptoms, presence of a preexisting medical condition, and performance of a nasopharyngeal wash for respiratory virus detection. In multivariable analysis, age <6 months (adjusted OR, 3.01; 95% CI, 1.17‐7.74; P = 0.022) and the performance of a blood culture (adjusted OR, 2.71; 95% CI, 1.03‐7.13; P = 0.043) were independently associated with a prolonged LOS.

Discussion

The absence of SBI in our relatively small cohort of children admitted to a community hospital with laboratory‐confirmed RGE supports earlier estimates of an incidence of less than 1%,5, 7 an incidence similar to that of occult bacteremia in febrile children 2 to 36 months of age following introduction of the heptavalent pneumococcal conjugate vaccine in 2000.10, 11 We found 13 cases reported in the English literature (Table 1). Several salient features are noted when comparing these case reports. All cases of SBI following laboratory‐confirmed RGE were characterized by the development of a second fever after the resolution of initial symptoms. These fevers presented at a mean day of hospitalization of 2.8 (range, 2‐5). Second fevers were high (mean, 39.2C; range, 38.2C to 40C). Cultures obtained other than peripheral blood cultures were only positive in 1 patient; this patient also had cellulitis and Escherichia coli was isolated from both blood and wound cultures.3 One of the reported children with bacteremia died, 2 cases of SBI following RGE were complicated by disseminated intravascular coagulopathy, and 1 case by acute renal failure. Enterobacter cloacae (n = 4) and Klebsiella pneumoniae (n = 3) were the most commonly isolated organisms from peripheral blood culture.

Many children in our study had ancillary laboratory testing performed. The results of these tests were typically normal and rarely affected clinical management in a positive manner. Bacteria and parasites are relatively rare causes of gastroenteritis in the United States in comparison with rotavirus, particularly during the winter months. However, stool was sent for bacterial culture in over 80% of patients and for ova and parasite detection in almost 10% of patients ultimately diagnosed with RGE. Furthermore, despite the relatively low prevalence of bacteremia since licensure of the Haemophilus influenzae type b vaccine, a majority of children had a complete blood count performed while one‐half also had blood obtained for culture. In our cohort, children 6 months and younger and those from whom a blood culture was collected were at an increased risk for prolonged LOS. It was not clear from medical record review whether children with prolonged LOS were kept in the hospital longer for the sole purpose of awaiting the results of blood cultures.

SBI rarely occurs in the context of RGE. While secondary fever seems to be a common manifestation, the sensitivity of secondary fever as a marker for SBI after RGE in this population is unknown. However, given the very low incidence, the potentially serious complications of SBI following laboratory confirmed RGE, and the likely successful management of these complications in the hospital setting, slightly longer hospitalizations for children under 1 year of age must be weighed against earlier discharges with instructions from clinicians to caregivers for careful monitoring of fever and outpatient follow‐up shortly after discharge.

This study has several limitations. First, the timing of the availability of the results of rotavirus antigen testing is not known. It is possible that the rapid availability of rotavirus test results in some circumstances encouraged clinicians to abandon tests seeking other sources of infection. Conversely, children with gastroenteritis in the context of a concurrent bacterial infection may have been less likely to undergo rotavirus stool antigen testing. This latter possibility would bias our findings toward underestimating the prevalence of concurrent bacterial infection among children with RGE. Second, this study was performed prior to licensure and widespread use of the currently‐licensed vaccine against rotavirus (Rotateq; Merck and Company, Whitehouse Station, NJ). Reductions in the burden of gastroenteritis caused by rotavirus may have a much more dramatic impact on resource utilization in the treatment of gastroenteritis than reductions in ancillary testing. Finally, this study was performed at a single urban community hospital and therefore cannot be generalized to other settings such as academic tertiary care centers. Furthermore, test ordering patterns may be local or regional and other community hospitals may exhibit different patterns. Further clarification of the role of ancillary testing in children presenting with diarrhea during the winter months is warranted since reducing the extent of such testing would dramatically reduce resource utilization for this illness. Finally, a blood culture was not obtained from all patients. Therefore, occult bacteremia attributable to RGE could not be detected. Since no patient in our study underwent subsequent clinical deterioration, we presume that any case of occult bacteremia resolved spontaneously and was not of clinical consequence, although such occurrences would cause us to underestimate the prevalence of SBI in this population.

Resource utilization in our cohort was high, while yield from microbiological investigations was low. This finding challenges the need to perform invasive, costly assays to exclude concurrent SBI in this population. It is possible that children with viral gastroenteritis caused by pathogens other than rotavirus are also at low risk of SBI. However, the diagnostic strategy that best identifies patients at risk for SBI following acute gastroenteritis is unknown. Further studies are needed to determine an ideal clinical approach to the infant with RGE.

Upper gastrointestinal hemorrhage (UGH) is a common cause of acute admission for hospitalization.13 However, recent advances in our understanding of erosive disease (ED) and peptic ulcer disease (PUD), 2 of the most common etiologies of UGH, have led to effective strategies to reduce the risk of UGH. Successful implementation of these strategies, such as treatment of Helicobacter pylori (H. pylori) and the use of proton pump inhibitors (PPIs) and selective cyclooxygenase‐2 inhibitors (COX‐2s) in place of traditional nonselective nonsteroidal antiinflammatory drugs (NSAIDs), may be able to significantly reduce rates of UGH caused by ED and PUD.47

Prior to these preventive treatments, PUD and ED, both acid‐related disorders, were the most common causes of UGH requiring admission to the hospital, accounting for 62% and 14% of all UGHs, respectively.2 Given the widespread treatment of H. pylori and use of PPIs and COX‐2s, we might expect that the distribution of etiologies of UGH may have changed. However, there are limited data on the distribution of etiologies of UGH in the era of effective preventive therapy.8 If the distribution of etiologies causing patients to present with UGH has fundamentally changed with these new treatments, established strategies of managing acute UGH may need to be reevaluated. Given that well‐established guidelines exist and that many hospitals use a protocol‐driven management strategy to decide on the need for admission and/or intensive care unit (ICU) admission, changes in the distribution of etiologies since the widespread use of these new pharmacologic approaches may affect the appropriateness of these protocols.9, 10 For example, if the eradication of H. pylori has dramatically reduced the proportion of UGH caused by PUD, then risk stratification studies developed when PUD was far more common may need to be revisited. This would be particularly important if bleeding from PUD was of significantly different severity than bleeding from other causes.

While patients with H. pylori‐related UGH from PUD should be treated for H. pylori eradication, several important questions remain surrounding the use of newer therapeutics that may mitigate the risk of UGH in some patients. It is unclear what proportion of patients admitted with UGH in this new era developed bleeding despite using preventive therapy. These treatment failures are known to occur, but it is not well known how much of the burden of UGH today is due to this breakthrough bleeding.5, 6, 11, 12 Contrastingly, there are also patients who are admitted with UGH who are not on preventive treatment. Current guidelines suggest that high‐risk patients requiring NSAIDs be given COX‐2s or traditional NSAIDs with a PPI.1315 However, there is significant disagreement between these national guidelines about what constitutes a high‐risk profile.1315 For example, some guidelines recommend that elderly patients requiring NSAIDs should be on a PPI while others do not make that recommendation. Similarly, while prior UGH is a well‐recognized risk factor for future bleeding risk even without NSAIDs, current guidelines do not provide guidance toward the use of preventive therapy in these patients. If there are few patients who present with UGH related to acid disease that are not on a preventive therapy, then these unanswered questions or conflicts within current guidelines become less important. However, if a large portion of UGH is due to acid‐related disease in patients not on preventive therapy, then these unanswered questions may become important for future research.

In contrast to previous studies, the current study examines the distribution of etiologies of UGH in the era of widespread use of effective preventive therapy for ED and PUD in 2 U.S. academic medical centers. Prior studies were done before the advent of new therapeutics and did not compare different sites, which may be important.16, 17

PATIENTS AND METHODS

RESULTS

From the entire cohort of 12,091 admitted to the 2 inpatient medical services, 227 (1.9%) patients were identified as having UGH; 138 (61%) were from center 1, where 87% of patients were black and 89 (39%) were from center 2, where 89% of patients were white. Overall, the mean age was 59 years, 45% were female, and 41% were white (Table 1).

The most common etiologies of UGH were ED (44%), PUD (33%), and varices (17%) in the overall population. These same 3 etiologies were also the most common in both of the medical centers, although there were significant differences in the rates of etiologies between the 2 centers. ED was more common among subjects from center 1 (59%) than from center 2 (19%) (P < 0.001), while variceal bleeding was more common among subjects from center 2 (34%) than from center 1 (6.5%) (P = 0.009) (Table 2).

In multivariate logistic regression analyses, only age and site remained independent predictors of etiologies. Advancing age was associated with a higher risk of arteriovenous malformations (AVMs) with the odds of AVMs increasing 6% for every additional year of life (P = 0.007). Site was associated with both ED and variceal bleeding. Patients from center 1 were significantly more likely to have UGH caused by ED, with an OR = 7.10 (P < 0.001), compared to subjects from center 2. However, subjects from center 1 had a significantly lower OR (OR = 0.12) than those subjects at center 2 (P = 0.009) of having UGH caused by a variceal bleed (Table 2).

Risk factors for UGH were common among the patients, including use of aspirin (25.1%), NSAIDs (22.9%), COX‐2s (4.9%), or prior history of UGH (43%). Additionally, 6.6% of patients were taking both an NSAID and aspirin. Differences between the 2 sites were seen only in aspirin use, with 34.8% of patients in the center 1 population using aspirin compared to 10.1% in center 2 (P < 0.001) (Table 3).

Among the overall population, 68.7% of patients had identifiable risk factors (prior history of UGH or preadmission use of aspirin, NSAIDs, or COX‐2s). Of all subjects, 18.5% were on PPIs and 4.9% were taking COX‐2s while 21.1% of at risk subjects were on PPIs and 6.5% of these subjects were on a COX‐2.

Finally, we examined the effects of variations in preadmission medication use between the sites on the etiologies of UGH. None of the site‐based differences in etiologies could be explained by differences in preadmission medication patterns.

DISCUSSION

Despite the emergence of effective therapies for lowering the risk of ED and PUD, these remain the most common etiologies of UGH in our cohort of patients. In a dramatic change from historically reported patterns, ED was more common than PUD. In prior studies, PUD accounted for almost two‐thirds of all UGH.2 While some of the newer therapeutics such as PPIs and COX‐2s reduce the risk for acid‐related bleeding of all types, H. pylori eradication is effective primarily for PUD. Therefore, it may be that widespread testing and treatment of H. pylori have dramatically decreased rates of PUD. Unfortunately, this study does not allow us to directly evaluate the effect of H. pylori treatment on the changing epidemiology of UGH, as that would require a population‐based study.

While decreasing rates of PUD could explain a portion of the change in the distribution of etiologies, increasing rates of ED could also be playing a role. Prior studies have suggested that African Americans and the elderly are more susceptible to ED, particularly in the setting of NSAIDs and/or aspirin use, and less susceptible to cirrhosis.13, 16, 17, 2023 Our finding of a higher rate of ED and lower rates of cirrhosis in center 1 with a higher proportion of African Americans and greater aspirin use is consistent with these prior findings. However, in multivariate analyses, neither race nor preadmission medication use patterns explained the differences in etiologies seen. This suggests that some other factors must play a role in the differences between the 2 centers studied. These results emphasize the importance of local site characteristics in the interpretation and implementation of national guidelines and recommendations. This finding may be particularly important in diseases and clinical presentations that rely on protocol‐driven pathways, such as UGH. Current recommendations on implementing clinical pathways derived from national guidelines emphasize the fact that national development and local implementation optimization is probably the best approach for effective pathway utilization.24

It is important to understand why ED and PUD, for which we now have effective pharmacologic therapies, continue to account for such a large percentage of the burden of UGH. In this study, we found that a majority of subjects were known to have significant risk factors for UGH (aspirin use, NSAID use, COX‐2s, or prior UGH) and only 31% of the subjects could not have been identified as at‐risk prior to admission. PPIs or COX‐2s should not be used universally as preventive therapy, and they are not completely effective at preventing UGH in at‐risk patients. In this study, two‐thirds of patients with risk factors were not on preventive therapy, but almost one‐third of patients with risk factors had bleeding despite being on preventive therapy. A better understanding of why these treatment failures (bleeding despite preventive therapy) occur may be helpful in our future ability to prevent UGH. This study was not designed to determine if the two‐thirds of patients not taking preventive therapy were being treated consistent with established guidelines. However, current guidelines have significant variation in recommendations as to which patients are at high enough risk to warrant preventive therapy,1315 and there is no consensus as to which patients are at high enough risk to warrant preventive therapy. Our data suggest that additional studies will be required to determine the optimal recommendations for preventive therapy among at‐risk patients.

There are several limitations to this study. First, it only included 2 academic institutions. However, these institutions represented very different patient populations. Second, the study design is not a population‐based study. This limitation prevents us from addressing questions such as the effectiveness or cost‐effectiveness of interventions to prevent admission for UGH. Although we analyzed preadmission PPI or COX‐2 use in at‐risk patients as preventive therapy, we are unable to determine the actual intent of the physician in prescribing these drugs. Finally, although the mechanisms by which PPIs and COX‐2 affect the risk of UGH are fundamentally different and should not be considered equivalent choices, we chose to analyze either option as representing a preventive strategy in order to provide the most conservative estimate possible of preventive therapy utilization rates. However, our assumptions would generally overestimate the use of preventive therapy (as opposed to PPI use for symptom control), as we assumed all potentially preventive therapy was intended as such.

This study highlights several unanswered questions that may be important in the management of UGH. First, identifying factors that affect local patters of UGH may better inform local implementation of nationally developed guidelines. Second, a more complete understanding of the impact positive and negative risk factors for UGH have on specific patient populations may allow for a more consistent targeted approach to using preventive therapy in at‐risk patients.

Finally, and perhaps most importantly, is to determine if the change in distribution of etiologies is in fact related to a decline in bleeding related to PUD. In addition to this being a marker of the success of the H. pylori story, it may have important implications on our understanding of the acute management of UGH. If PUD is of a different severity than other common causes of UGH, such as ED, current risk stratification prediction models may need to be revalidated. For example, if UGH secondary to PUD results in greater morbidity and mortality than UGH secondary to ED, our current models identifying who requires ICU admission, urgent endoscopy, and other therapeutic interventions may result in overutilization of these resource intensive interventions. However, if larger studies do not confirm this decline in PUD it suggests the need for additional studies to identify why PUD remains so prevalent despite the major advances in treatment and prevention of PUD through H. pylori identification and eradication.

Upper gastrointestinal hemorrhage (UGH) is a common cause of acute admission for hospitalization.13 However, recent advances in our understanding of erosive disease (ED) and peptic ulcer disease (PUD), 2 of the most common etiologies of UGH, have led to effective strategies to reduce the risk of UGH. Successful implementation of these strategies, such as treatment of Helicobacter pylori (H. pylori) and the use of proton pump inhibitors (PPIs) and selective cyclooxygenase‐2 inhibitors (COX‐2s) in place of traditional nonselective nonsteroidal antiinflammatory drugs (NSAIDs), may be able to significantly reduce rates of UGH caused by ED and PUD.47

Prior to these preventive treatments, PUD and ED, both acid‐related disorders, were the most common causes of UGH requiring admission to the hospital, accounting for 62% and 14% of all UGHs, respectively.2 Given the widespread treatment of H. pylori and use of PPIs and COX‐2s, we might expect that the distribution of etiologies of UGH may have changed. However, there are limited data on the distribution of etiologies of UGH in the era of effective preventive therapy.8 If the distribution of etiologies causing patients to present with UGH has fundamentally changed with these new treatments, established strategies of managing acute UGH may need to be reevaluated. Given that well‐established guidelines exist and that many hospitals use a protocol‐driven management strategy to decide on the need for admission and/or intensive care unit (ICU) admission, changes in the distribution of etiologies since the widespread use of these new pharmacologic approaches may affect the appropriateness of these protocols.9, 10 For example, if the eradication of H. pylori has dramatically reduced the proportion of UGH caused by PUD, then risk stratification studies developed when PUD was far more common may need to be revisited. This would be particularly important if bleeding from PUD was of significantly different severity than bleeding from other causes.

While patients with H. pylori‐related UGH from PUD should be treated for H. pylori eradication, several important questions remain surrounding the use of newer therapeutics that may mitigate the risk of UGH in some patients. It is unclear what proportion of patients admitted with UGH in this new era developed bleeding despite using preventive therapy. These treatment failures are known to occur, but it is not well known how much of the burden of UGH today is due to this breakthrough bleeding.5, 6, 11, 12 Contrastingly, there are also patients who are admitted with UGH who are not on preventive treatment. Current guidelines suggest that high‐risk patients requiring NSAIDs be given COX‐2s or traditional NSAIDs with a PPI.1315 However, there is significant disagreement between these national guidelines about what constitutes a high‐risk profile.1315 For example, some guidelines recommend that elderly patients requiring NSAIDs should be on a PPI while others do not make that recommendation. Similarly, while prior UGH is a well‐recognized risk factor for future bleeding risk even without NSAIDs, current guidelines do not provide guidance toward the use of preventive therapy in these patients. If there are few patients who present with UGH related to acid disease that are not on a preventive therapy, then these unanswered questions or conflicts within current guidelines become less important. However, if a large portion of UGH is due to acid‐related disease in patients not on preventive therapy, then these unanswered questions may become important for future research.

In contrast to previous studies, the current study examines the distribution of etiologies of UGH in the era of widespread use of effective preventive therapy for ED and PUD in 2 U.S. academic medical centers. Prior studies were done before the advent of new therapeutics and did not compare different sites, which may be important.16, 17

PATIENTS AND METHODS

RESULTS

From the entire cohort of 12,091 admitted to the 2 inpatient medical services, 227 (1.9%) patients were identified as having UGH; 138 (61%) were from center 1, where 87% of patients were black and 89 (39%) were from center 2, where 89% of patients were white. Overall, the mean age was 59 years, 45% were female, and 41% were white (Table 1).

The most common etiologies of UGH were ED (44%), PUD (33%), and varices (17%) in the overall population. These same 3 etiologies were also the most common in both of the medical centers, although there were significant differences in the rates of etiologies between the 2 centers. ED was more common among subjects from center 1 (59%) than from center 2 (19%) (P < 0.001), while variceal bleeding was more common among subjects from center 2 (34%) than from center 1 (6.5%) (P = 0.009) (Table 2).

In multivariate logistic regression analyses, only age and site remained independent predictors of etiologies. Advancing age was associated with a higher risk of arteriovenous malformations (AVMs) with the odds of AVMs increasing 6% for every additional year of life (P = 0.007). Site was associated with both ED and variceal bleeding. Patients from center 1 were significantly more likely to have UGH caused by ED, with an OR = 7.10 (P < 0.001), compared to subjects from center 2. However, subjects from center 1 had a significantly lower OR (OR = 0.12) than those subjects at center 2 (P = 0.009) of having UGH caused by a variceal bleed (Table 2).

Risk factors for UGH were common among the patients, including use of aspirin (25.1%), NSAIDs (22.9%), COX‐2s (4.9%), or prior history of UGH (43%). Additionally, 6.6% of patients were taking both an NSAID and aspirin. Differences between the 2 sites were seen only in aspirin use, with 34.8% of patients in the center 1 population using aspirin compared to 10.1% in center 2 (P < 0.001) (Table 3).

Among the overall population, 68.7% of patients had identifiable risk factors (prior history of UGH or preadmission use of aspirin, NSAIDs, or COX‐2s). Of all subjects, 18.5% were on PPIs and 4.9% were taking COX‐2s while 21.1% of at risk subjects were on PPIs and 6.5% of these subjects were on a COX‐2.

Finally, we examined the effects of variations in preadmission medication use between the sites on the etiologies of UGH. None of the site‐based differences in etiologies could be explained by differences in preadmission medication patterns.

DISCUSSION

Despite the emergence of effective therapies for lowering the risk of ED and PUD, these remain the most common etiologies of UGH in our cohort of patients. In a dramatic change from historically reported patterns, ED was more common than PUD. In prior studies, PUD accounted for almost two‐thirds of all UGH.2 While some of the newer therapeutics such as PPIs and COX‐2s reduce the risk for acid‐related bleeding of all types, H. pylori eradication is effective primarily for PUD. Therefore, it may be that widespread testing and treatment of H. pylori have dramatically decreased rates of PUD. Unfortunately, this study does not allow us to directly evaluate the effect of H. pylori treatment on the changing epidemiology of UGH, as that would require a population‐based study.

While decreasing rates of PUD could explain a portion of the change in the distribution of etiologies, increasing rates of ED could also be playing a role. Prior studies have suggested that African Americans and the elderly are more susceptible to ED, particularly in the setting of NSAIDs and/or aspirin use, and less susceptible to cirrhosis.13, 16, 17, 2023 Our finding of a higher rate of ED and lower rates of cirrhosis in center 1 with a higher proportion of African Americans and greater aspirin use is consistent with these prior findings. However, in multivariate analyses, neither race nor preadmission medication use patterns explained the differences in etiologies seen. This suggests that some other factors must play a role in the differences between the 2 centers studied. These results emphasize the importance of local site characteristics in the interpretation and implementation of national guidelines and recommendations. This finding may be particularly important in diseases and clinical presentations that rely on protocol‐driven pathways, such as UGH. Current recommendations on implementing clinical pathways derived from national guidelines emphasize the fact that national development and local implementation optimization is probably the best approach for effective pathway utilization.24

It is important to understand why ED and PUD, for which we now have effective pharmacologic therapies, continue to account for such a large percentage of the burden of UGH. In this study, we found that a majority of subjects were known to have significant risk factors for UGH (aspirin use, NSAID use, COX‐2s, or prior UGH) and only 31% of the subjects could not have been identified as at‐risk prior to admission. PPIs or COX‐2s should not be used universally as preventive therapy, and they are not completely effective at preventing UGH in at‐risk patients. In this study, two‐thirds of patients with risk factors were not on preventive therapy, but almost one‐third of patients with risk factors had bleeding despite being on preventive therapy. A better understanding of why these treatment failures (bleeding despite preventive therapy) occur may be helpful in our future ability to prevent UGH. This study was not designed to determine if the two‐thirds of patients not taking preventive therapy were being treated consistent with established guidelines. However, current guidelines have significant variation in recommendations as to which patients are at high enough risk to warrant preventive therapy,1315 and there is no consensus as to which patients are at high enough risk to warrant preventive therapy. Our data suggest that additional studies will be required to determine the optimal recommendations for preventive therapy among at‐risk patients.

There are several limitations to this study. First, it only included 2 academic institutions. However, these institutions represented very different patient populations. Second, the study design is not a population‐based study. This limitation prevents us from addressing questions such as the effectiveness or cost‐effectiveness of interventions to prevent admission for UGH. Although we analyzed preadmission PPI or COX‐2 use in at‐risk patients as preventive therapy, we are unable to determine the actual intent of the physician in prescribing these drugs. Finally, although the mechanisms by which PPIs and COX‐2 affect the risk of UGH are fundamentally different and should not be considered equivalent choices, we chose to analyze either option as representing a preventive strategy in order to provide the most conservative estimate possible of preventive therapy utilization rates. However, our assumptions would generally overestimate the use of preventive therapy (as opposed to PPI use for symptom control), as we assumed all potentially preventive therapy was intended as such.

This study highlights several unanswered questions that may be important in the management of UGH. First, identifying factors that affect local patters of UGH may better inform local implementation of nationally developed guidelines. Second, a more complete understanding of the impact positive and negative risk factors for UGH have on specific patient populations may allow for a more consistent targeted approach to using preventive therapy in at‐risk patients.

Finally, and perhaps most importantly, is to determine if the change in distribution of etiologies is in fact related to a decline in bleeding related to PUD. In addition to this being a marker of the success of the H. pylori story, it may have important implications on our understanding of the acute management of UGH. If PUD is of a different severity than other common causes of UGH, such as ED, current risk stratification prediction models may need to be revalidated. For example, if UGH secondary to PUD results in greater morbidity and mortality than UGH secondary to ED, our current models identifying who requires ICU admission, urgent endoscopy, and other therapeutic interventions may result in overutilization of these resource intensive interventions. However, if larger studies do not confirm this decline in PUD it suggests the need for additional studies to identify why PUD remains so prevalent despite the major advances in treatment and prevention of PUD through H. pylori identification and eradication.

Upper gastrointestinal hemorrhage (UGH) is a common cause of acute admission for hospitalization.13 However, recent advances in our understanding of erosive disease (ED) and peptic ulcer disease (PUD), 2 of the most common etiologies of UGH, have led to effective strategies to reduce the risk of UGH. Successful implementation of these strategies, such as treatment of Helicobacter pylori (H. pylori) and the use of proton pump inhibitors (PPIs) and selective cyclooxygenase‐2 inhibitors (COX‐2s) in place of traditional nonselective nonsteroidal antiinflammatory drugs (NSAIDs), may be able to significantly reduce rates of UGH caused by ED and PUD.47

Prior to these preventive treatments, PUD and ED, both acid‐related disorders, were the most common causes of UGH requiring admission to the hospital, accounting for 62% and 14% of all UGHs, respectively.2 Given the widespread treatment of H. pylori and use of PPIs and COX‐2s, we might expect that the distribution of etiologies of UGH may have changed. However, there are limited data on the distribution of etiologies of UGH in the era of effective preventive therapy.8 If the distribution of etiologies causing patients to present with UGH has fundamentally changed with these new treatments, established strategies of managing acute UGH may need to be reevaluated. Given that well‐established guidelines exist and that many hospitals use a protocol‐driven management strategy to decide on the need for admission and/or intensive care unit (ICU) admission, changes in the distribution of etiologies since the widespread use of these new pharmacologic approaches may affect the appropriateness of these protocols.9, 10 For example, if the eradication of H. pylori has dramatically reduced the proportion of UGH caused by PUD, then risk stratification studies developed when PUD was far more common may need to be revisited. This would be particularly important if bleeding from PUD was of significantly different severity than bleeding from other causes.

While patients with H. pylori‐related UGH from PUD should be treated for H. pylori eradication, several important questions remain surrounding the use of newer therapeutics that may mitigate the risk of UGH in some patients. It is unclear what proportion of patients admitted with UGH in this new era developed bleeding despite using preventive therapy. These treatment failures are known to occur, but it is not well known how much of the burden of UGH today is due to this breakthrough bleeding.5, 6, 11, 12 Contrastingly, there are also patients who are admitted with UGH who are not on preventive treatment. Current guidelines suggest that high‐risk patients requiring NSAIDs be given COX‐2s or traditional NSAIDs with a PPI.1315 However, there is significant disagreement between these national guidelines about what constitutes a high‐risk profile.1315 For example, some guidelines recommend that elderly patients requiring NSAIDs should be on a PPI while others do not make that recommendation. Similarly, while prior UGH is a well‐recognized risk factor for future bleeding risk even without NSAIDs, current guidelines do not provide guidance toward the use of preventive therapy in these patients. If there are few patients who present with UGH related to acid disease that are not on a preventive therapy, then these unanswered questions or conflicts within current guidelines become less important. However, if a large portion of UGH is due to acid‐related disease in patients not on preventive therapy, then these unanswered questions may become important for future research.

In contrast to previous studies, the current study examines the distribution of etiologies of UGH in the era of widespread use of effective preventive therapy for ED and PUD in 2 U.S. academic medical centers. Prior studies were done before the advent of new therapeutics and did not compare different sites, which may be important.16, 17

PATIENTS AND METHODS

RESULTS

From the entire cohort of 12,091 admitted to the 2 inpatient medical services, 227 (1.9%) patients were identified as having UGH; 138 (61%) were from center 1, where 87% of patients were black and 89 (39%) were from center 2, where 89% of patients were white. Overall, the mean age was 59 years, 45% were female, and 41% were white (Table 1).

The most common etiologies of UGH were ED (44%), PUD (33%), and varices (17%) in the overall population. These same 3 etiologies were also the most common in both of the medical centers, although there were significant differences in the rates of etiologies between the 2 centers. ED was more common among subjects from center 1 (59%) than from center 2 (19%) (P < 0.001), while variceal bleeding was more common among subjects from center 2 (34%) than from center 1 (6.5%) (P = 0.009) (Table 2).

In multivariate logistic regression analyses, only age and site remained independent predictors of etiologies. Advancing age was associated with a higher risk of arteriovenous malformations (AVMs) with the odds of AVMs increasing 6% for every additional year of life (P = 0.007). Site was associated with both ED and variceal bleeding. Patients from center 1 were significantly more likely to have UGH caused by ED, with an OR = 7.10 (P < 0.001), compared to subjects from center 2. However, subjects from center 1 had a significantly lower OR (OR = 0.12) than those subjects at center 2 (P = 0.009) of having UGH caused by a variceal bleed (Table 2).

Risk factors for UGH were common among the patients, including use of aspirin (25.1%), NSAIDs (22.9%), COX‐2s (4.9%), or prior history of UGH (43%). Additionally, 6.6% of patients were taking both an NSAID and aspirin. Differences between the 2 sites were seen only in aspirin use, with 34.8% of patients in the center 1 population using aspirin compared to 10.1% in center 2 (P < 0.001) (Table 3).

Among the overall population, 68.7% of patients had identifiable risk factors (prior history of UGH or preadmission use of aspirin, NSAIDs, or COX‐2s). Of all subjects, 18.5% were on PPIs and 4.9% were taking COX‐2s while 21.1% of at risk subjects were on PPIs and 6.5% of these subjects were on a COX‐2.

Finally, we examined the effects of variations in preadmission medication use between the sites on the etiologies of UGH. None of the site‐based differences in etiologies could be explained by differences in preadmission medication patterns.

DISCUSSION

Despite the emergence of effective therapies for lowering the risk of ED and PUD, these remain the most common etiologies of UGH in our cohort of patients. In a dramatic change from historically reported patterns, ED was more common than PUD. In prior studies, PUD accounted for almost two‐thirds of all UGH.2 While some of the newer therapeutics such as PPIs and COX‐2s reduce the risk for acid‐related bleeding of all types, H. pylori eradication is effective primarily for PUD. Therefore, it may be that widespread testing and treatment of H. pylori have dramatically decreased rates of PUD. Unfortunately, this study does not allow us to directly evaluate the effect of H. pylori treatment on the changing epidemiology of UGH, as that would require a population‐based study.

While decreasing rates of PUD could explain a portion of the change in the distribution of etiologies, increasing rates of ED could also be playing a role. Prior studies have suggested that African Americans and the elderly are more susceptible to ED, particularly in the setting of NSAIDs and/or aspirin use, and less susceptible to cirrhosis.13, 16, 17, 2023 Our finding of a higher rate of ED and lower rates of cirrhosis in center 1 with a higher proportion of African Americans and greater aspirin use is consistent with these prior findings. However, in multivariate analyses, neither race nor preadmission medication use patterns explained the differences in etiologies seen. This suggests that some other factors must play a role in the differences between the 2 centers studied. These results emphasize the importance of local site characteristics in the interpretation and implementation of national guidelines and recommendations. This finding may be particularly important in diseases and clinical presentations that rely on protocol‐driven pathways, such as UGH. Current recommendations on implementing clinical pathways derived from national guidelines emphasize the fact that national development and local implementation optimization is probably the best approach for effective pathway utilization.24

It is important to understand why ED and PUD, for which we now have effective pharmacologic therapies, continue to account for such a large percentage of the burden of UGH. In this study, we found that a majority of subjects were known to have significant risk factors for UGH (aspirin use, NSAID use, COX‐2s, or prior UGH) and only 31% of the subjects could not have been identified as at‐risk prior to admission. PPIs or COX‐2s should not be used universally as preventive therapy, and they are not completely effective at preventing UGH in at‐risk patients. In this study, two‐thirds of patients with risk factors were not on preventive therapy, but almost one‐third of patients with risk factors had bleeding despite being on preventive therapy. A better understanding of why these treatment failures (bleeding despite preventive therapy) occur may be helpful in our future ability to prevent UGH. This study was not designed to determine if the two‐thirds of patients not taking preventive therapy were being treated consistent with established guidelines. However, current guidelines have significant variation in recommendations as to which patients are at high enough risk to warrant preventive therapy,1315 and there is no consensus as to which patients are at high enough risk to warrant preventive therapy. Our data suggest that additional studies will be required to determine the optimal recommendations for preventive therapy among at‐risk patients.

There are several limitations to this study. First, it only included 2 academic institutions. However, these institutions represented very different patient populations. Second, the study design is not a population‐based study. This limitation prevents us from addressing questions such as the effectiveness or cost‐effectiveness of interventions to prevent admission for UGH. Although we analyzed preadmission PPI or COX‐2 use in at‐risk patients as preventive therapy, we are unable to determine the actual intent of the physician in prescribing these drugs. Finally, although the mechanisms by which PPIs and COX‐2 affect the risk of UGH are fundamentally different and should not be considered equivalent choices, we chose to analyze either option as representing a preventive strategy in order to provide the most conservative estimate possible of preventive therapy utilization rates. However, our assumptions would generally overestimate the use of preventive therapy (as opposed to PPI use for symptom control), as we assumed all potentially preventive therapy was intended as such.

This study highlights several unanswered questions that may be important in the management of UGH. First, identifying factors that affect local patters of UGH may better inform local implementation of nationally developed guidelines. Second, a more complete understanding of the impact positive and negative risk factors for UGH have on specific patient populations may allow for a more consistent targeted approach to using preventive therapy in at‐risk patients.

Finally, and perhaps most importantly, is to determine if the change in distribution of etiologies is in fact related to a decline in bleeding related to PUD. In addition to this being a marker of the success of the H. pylori story, it may have important implications on our understanding of the acute management of UGH. If PUD is of a different severity than other common causes of UGH, such as ED, current risk stratification prediction models may need to be revalidated. For example, if UGH secondary to PUD results in greater morbidity and mortality than UGH secondary to ED, our current models identifying who requires ICU admission, urgent endoscopy, and other therapeutic interventions may result in overutilization of these resource intensive interventions. However, if larger studies do not confirm this decline in PUD it suggests the need for additional studies to identify why PUD remains so prevalent despite the major advances in treatment and prevention of PUD through H. pylori identification and eradication.

The Accreditation Counsel for Graduate Medical Education (ACGME) states in its Program Requirements for Residency Education in Internal Medicine that all residents must develop technical proficiency in several procedures, including central venous line placement.1 Developing competency in common procedural skills has long been a part of medical training. The philosophy of see‐1, do‐1, teach‐1 is still the most common means by which most residents seek to obtain this proficiency, even though serious concerns have been raised about this approach.2 A typical first experience in central line placement usually involves an eager (or terrified) trainee making several clumsy attempts on an actual patient, under the hurried guidance of a senior resident who themselves received an unknown degree of training. In this scenario, rarely does standardized instruction, formal evaluation, or structured follow‐up occur.

A revitalized emphasis is now being placed on patient safety in healthcare, including an industry‐wide commitment to minimizing procedural complications. The most common complications associated to central line placement include vascular damage and catheter‐related bloodstream infections. A number of creative approaches are being developed to improve the quality of instruction on proper procedural techniques, all varying considerably in sophistication, scope, and rigor. Examples include the use of computer‐assisted methods for training ultrasound‐guided needle insertion techniques and ureteroscopy training, hands‐on training with synthetic models for thoracentesis training, video training, and uterine aspiration using papayas.311 Implicit in this trend is recognition that we, as educators, healthcare providers, and patient advocates, must design more cost effective and efficient ways to teach medical and surgical procedural techniques to clinicians.

Our approach was previously described in phase I of the Procedural Patient Safety Initiative (PPSI).12 In PPSI‐I, we introduced a nonhuman tissue model (NHTM; Figure 1) as the basis for teaching physicians a more rigorous curriculum of essential central line placement skills. By way of brief review, the NHTMs were constructed by tunneling 0.2‐mm‐thick rubber tubing (vessels) lengthwise through raw, whole chickens purchased at the grocery store. The vessels were filled with colored water to simulate blood. The NHTM has several unique features, including: (1) realistic‐appearing vessels when viewed under ultrasound, which mimic the appearance of human internal jugular veins and carotid arteries (Figure 2); (2) tissue turgor and vessel composition that produce realistic pops and flashes upon puncture and allow for multiple cannulations; (3) the ability to perform a complete central line placement (including wire advancement, dilation, line insertion, suturing, and sterile dressing placement); (4) cost effectiveness relative to other commercially‐available products (each NHTM costs $120 and can withstand multiple cannulations over 2 days).1317 During the training sessions of Phase I, participants were oriented to the ultrasound machines, shown the contents of our central line kit, and taught the principles of wide sterile barriers (WSBs), sharps safety, and vascular access under real‐time ultrasound guidance. A self‐completed survey tool was filled out by the participants before and after the session that contained questions about their precourse baseline procedural experience, and their subjective comfort level with specific skills after the course. The results of our intervention, as measured by the responses to the survey, were significantly positive. We recognized the limitations of these results based on using subjective criteria to measure efficacy, a lack of follow‐up on participants' skill retention, and with no ultimate evaluation of procedural competency evaluations on actual patients (compared to an untrained control group).

Figure 1

Figure 2

Our ultimate goal is to validate a curriculum that will give trainees the necessary education and skills that enables them to make a smooth, competent, and complication‐free transition to live patient procedures. Phase II of PPSI is our next step toward this goal. In this study, we sought to measure the impact of intensive, 1:1 central line placement training with a proceduralist, objectively validate the efficacy of the NHTM and our training curriculum using a standardized 6‐point scoring scale and skilled evaluators, and to evaluate the degree of skill retention over time (decay). Our hypothesis was that the depth of skills' imprinting from a single, standardized training session would result in a significant improvement in measured procedure skills immediately after the trainee is taught the skills, and that the retention of these skills would be demonstrable when participants were reevaluated at a future date.

Methods

PPSI‐II was an observational, prospective study conducted by The Procedure Center at Cedars‐Sinai Medical Center, a 900+‐bed, community‐based teaching hospital. The Procedure Center is staffed by dedicated Proceduralists who perform a number of common medical procedures on a daily basis and are facile with both real‐time ultrasound guidance and proper procedural techniques.18, 19 Our target population was the incoming Internal Medicine residents. Subjects were recruited by email prior to orientation week and were offered the option of participating in our study. Our only exclusion criterion was the prior observation or placement of 10 or more central lines. The study was approved by our hospital's Institutional Review Board prior to initiating recruitment. Those who chose not to participate underwent the standard orientation training required by our institution, which included a brief overview lecture on the topic of central lines and ultrasound‐guidance, a group viewing of the New England Journal of Medicine (NEJM) video on central line placement,20 and small‐group (4 participants/group) hands‐on practice sessions lasting 45 minutes with NHTMs and a trained Proceduralist.

All of the evaluations for Phase II were done using the Central Line Placement Skill Assessment Tool depicted in Figure 3. This tool, which was developed by Cedars‐Sinai Proceduralists solely for the purposes of this study, is a comprehensive step‐by‐step checklist delineating the specific steps necessary to place a sterile, ultrasound‐guided central venous catheter. It was closely derived from a central line insertion checklist that was created by the Procedure Center 3 years ago to help guide novice clinicians through the procedure, and has since been widely used throughout the institution during the placement of hundreds of central lines. The scoring system, also devised by Cedars‐Sinai Proceduralists, was based on over 15 years of experience supervising and teaching hundreds of residents on proper central line insertion techniques. It consists of clear definitions for each score that were agreed upon via consensus amongst study coordinators. Prior to any evaluations being conducted, we put our 2 evaluators (both senior medicine residents) through identical and simultaneous scoring training with the Proceduralist trainer to standardize procedural knowledge and scoring methodology.

Figure 3

A total of 20 incoming interns (trainees) out of a possible 54 invitations (37%) volunteered to participate. Each trainee was randomly assigned a number from 1 to 20. The study began for each trainee with a brief, 5‐question survey to determine their prior procedural experience (Table 1). Next, each trainee watched the NEJM online training video on central line placement.20 They were then brought into a training room that contained an NHTM sitting on a Mayo stand, an ultrasound machine, and all the materials required to place a central line insertion under ultrasound‐guidance. The trainee's baseline central line insertion skills were evaluated on 22 unique procedure steps, with each score being given by 1 of the 2 evaluators (initial evaluation). The trainee did not receive any guidance or suggestions during this initial evaluation unless the trainee reached an impasse. In these cases, the evaluator completed that single step on the trainee's behalf and then allowed the session to resume. The identity of the evaluator was indicated on each evaluation form, and after each of the evaluations the completed assessment tool was given to our blinded assistant for data recording.

Each trainee was then given a personalized, hands‐on training session by a proceduralist, using the checklist as a guide to take them through all the steps of a central line insertion. The trainee was allowed to observe and practice each skill for an unlimited period of time with the proceduralist present, until he or she demonstrated competency and felt confident enough with their independent skills (in both trainee's and proceduralist's judgment) to move forward. The entire session ended only when all steps had been taught and practiced to the proceduralist's satisfaction, the trainee felt comfortable independently performing each step (and in proper sequence), and all questions had been answered. At no time was there an imposition of time constraints or external pressure from study coordinators.

Immediately following this training session the proceduralist and trainee left the room, the procedure room was reset by an evaluator (taking approximately 5‐10 minutes), and then the trainee submitted themselves to an immediate posttraining evaluation (immediate evaluation). As with the initial assessment, the evaluator did not interfere or make any comments or suggestions during the evaluation periods, unless the trainee reached an impasse at any step. In that case, the trainee would receive a 0 for that step, the evaluator would assist them to complete that step only, and then the session would continue. No time limits were imposed.

The final part of the study required each trainee to return for follow‐up assessment (delayed evaluation), a process that was identical to the immediate posttraining evaluation. This delayed evaluation was intended to occur between 3 to 4 weeks after the immediate posttraining session, based on trainees' schedules and availability. No refresher or practice time was permitted prior to the delayed evaluation: upon arrival, trainees wrote down on a separate piece of paper (not seen by the evaluator) the number of interim line experiences they had experienced, then they were brought directly into a fully‐prepared room, and instructed to begin. The evaluator was also blind to the trainee's scores from the 2 previous evaluation sessions.

The primary endpoints were the degree of changes in overall average scores (from the 22 steps on the assessment tool) from the initial to the immediate evaluations and from the immediate to delayed evaluations. The secondary endpoints were also based on changes in average scores from the initial to immediate and immediate to delayed evaluations, and looked at 5 essential elements (steps in the assessment tool that we deemed critical to the safe and successful placement of a central line). These essential elements included (1) hand washing; (2) creation of a WSB; (3) ultrasound‐guided vessel cannulation; (4) proper catheter placement; and (5) sharps safety. Of note, the creation of a WSB element consisted of 4 steps, each of which was analyzed separately. The average scores are reported as means standard deviations (SDs).

To determine the type of analysis that would be performed, we started by assessing the changes using paired t tests. The Kolmogorov‐Smirnov and Anderson‐Darling normality tests revealed no evidence of violations of the normality assumption, confirming that using paired t tests was valid.

To address potential contamination from residents' real experiences on the rate of their knowledge decay between the immediate evaluation and delayed follow‐up, each participant completed a brief survey before their delayed evaluation asking about interim experiences. All calculations were performed including and excluding from participants' scores with affirmative answers to control for this contamination. Last, a post‐hoc analysis was performed on participants' scores using a scatterplot and statistical analyses to control for the varying time‐to‐follow‐up.

Results

All 20 individuals completed the study, for a total of 60 evaluations (20 each of initial, immediate, and delayed). The actual training time (not including the viewing of the video) ranged between 45 to 120 minutes, depending on the trainee. Our primary endpoints are depicted in Table 2. The mean overall score on the initial evaluation was 1.0 0.8. The mean overall score for the immediate posttraining evaluation was 4.4 0.3. This improvement of 3.4 points was significant (P < 0.001; 95% CI, 3.0‐3.7). The delayed evaluations took place an average of 22 days after the training session (range, 5‐101 days), and produced an overall mean score of 4.2 0.3. This decay of 0.2 was not significant (P = 0.14; 95% CI, 0.31 to 0.05). With regard to the amount of skills decay, additional calculations were performed from the scatterplot that depicted scores and the variability in time‐to‐follow‐up. We found that even after controlling this variable, the amount of decay for the overall score remained insignificant.

The results of the secondary endpoint calculations (essential elements) are depicted in Table 3. Ultrasound‐guided vessel cannulations improved from an initial average score of 0.9 1.0 to an immediate average score of 4.2 0.5 (P < 0.001; 95% CI, 3.0‐3.7); the delayed score of 4.3 0.6 was statistically unchanged from immediate (P = 0.77; 95% CI, 0.4 to 0.3). Catheter placement skills improved from 1.1 1.1 to 4.2 0.5 (P < 0.001; 95% CI, 2.6‐3.7), and the delayed score of 4.3 0.7 was unchanged from immediate (P < 0.58; 95% CI, 0.5 to 0.3). Sharps safety also improved significantly from initial (2.0 2.3) to immediate (4.9 0.5) (P < 0.0001; 95% CI, 1.9‐3.9), and the delayed scores dropped insignificantly to 4.6 0.8 (P = 0.08; 95% CI, 0.0‐0.6). Hand washing improved significantly from an initial score of 0.9 1.9 to an immediate score of 3.5 2.2 (P < 0.001; 95% CI, 1.4‐3.7), and decayed insignificantly on the delayed evaluation to 3.0 2.3 (P = 0.53; 95% CI, 0.9 to 1.7). WSB skills consisted of 4 individual steps, all of which all improved significantly from initial to immediate scores, and had insignificant decays on the delayed evaluations (see Table 3 WSB for details).

We performed validation exercises to determine the degree of interrater agreement. Of the 60 total evaluations that were eventually performed, 11 evaluations had been performed simultaneously and independently by evaluators A and B. An analysis of the scores assigned by each evaluator to these 11 trainees revealed a high level of interrater agreement (96%). Further, we performed independent analyses of the trainees' scores as assessed by evaluator A (22 sessions) or evaluator B (27 sessions) across the initial, immediate, and delayed sessions, and we detected no statistical differences in the changes in scores (which mirrored the overall results above).

With regards to real‐life contamination between immediate scores and delayed scores, we identified 3 trainees who had placed central lines on actual patients during the interim period (2 trainees placed 1 line each, and 1 trainee placed 2 lines). We repeated all of the calculations without these participants' delayed scores and determined that the removal of their scores did not change the statistical significance of any of the study results. With regard to knowledge decay, the scatterplot comparing delayed scores to varying time‐to‐follow up revealed no correlation.

Discussion

Our study was designed to determine whether novice trainees could learn and retain proper central line placement skills on the NHTM by receiving personalized training in a relaxed, 1‐on‐1 learning environment. Success was measured by trained evaluators using a detailed evaluation tool with a 6‐point scoring scale. The results of our primary endpoints (changes in overall average scores across the 3 evaluation periods) confirmed that this type of training could quickly improve novice practitioners' skill levels from very low (initial evaluation) to significantly higher (immediate posttraining). The dropoff (decay) in skill levels was found to not be statistically significant over a period of several weeks, although we recognize that further study should be performed to establish the degree of skill decay over a longer period of time.

Because some steps in a central line insertion are more critical to the procedure's success than others (ie, a skin nick with a scalpel is less critical than vessel cannulation under ultrasound‐guidance), we analyzed 5 essential elements individually as secondary endpoints. This secondary analysis was designed to unmask any critical skill deficiencies that might otherwise have been lost in the overall analysis. For each individual essential elements step, this subanalysis similarly revealed a significant improvement from initial to immediate posttraining, and an insignificant score decay on the delayed evaluation.

We recognize a number of limitations to this study. First, the n is relatively small. A larger sample size would have allowed for greater statistical power. In addition, the scoring system used for this study was created by our Procedure Center staff and had never been truly validated elsewhere. The scoring system was transparent and logical, but we recognize that any attempt to use an interval scoring system to quantify procedural skills will be inherently imperfect; the difference between 1 and 2 is not necessarily the same as a difference between 4 and 5. Great efforts were taken to mitigate the impact of this limitation: explicit definitions were established for each score, and we put our evaluators through a rigorous scoring orientation at the outset to standardize their interpretation and use of the scoring system and assessment tool.

The variability in the amount of training time spent in each session could be considered to be a confounder. Our prior experiences training interns in small groups, however, suggested that individuals learn these skills at different paces and in different ways, and so we consider our customized approach to be an essential part of this training experience. We do recognize the practical limitations inherent in rolling out such an open‐ended approach, and program directors may face time and/or resource limitations if attempting to replicate this training strategy.

We were also aware of potential interrater variability between the evaluators. Our approach to addressing this was multifactorial: we went to great lengths to standardize evaluators' understanding of the intended scoring methodology prior to the initiation of the study. We also assessed the degree of interrater reliability once all data was collected. This analysis reinforced that both evaluators were scoring trainees in a virtually identical fashion. We attribute this consistency to the quality of the scoring system, the effectiveness of the prestudy evaluator orientation with a proceduralist, and the high degree of teamwork between the 2 evaluators that kept them closely in sync with one another throughout the study.

Evaluator bias was also a concern. While each evaluator was blinded to the trainees' prior scores, the setup associated with the different training sessions, as well as the obvious differences in performance between the trainees' initial and immediate/delayed performances, made full blinding of the evaluators difficult. The theoretical risk of evaluator bias in this study would have led to evaluators rating trainees higher in the immediate and delayed performances in order to demonstrate more dramatic results. We believe that, since the evaluators themselves did not perform the actual training, and since they did not know the previous scores for the trainee, they were less inclined to skew the scores. Video recording each performance and submitting this recording to a fully‐blinded, third‐party evaluator would have more rigorously ensured blinding than we were able to accomplish. This approach could be considered in future studies of this type.

An addition limitation involved the time‐to‐follow‐up. While a longer time interval between the immediate and delayed evaluations may have better evaluated the impact of the training and potential decay, we sought to balance this with the growing risk of contamination from real central line placement experiences as more time passed. With this issue in mind, the removal of the delayed scores from the 3 trainees who had placed central lines on actual patients in between the immediate and delayed evaluations (2 trainees placed 1 line each, and 1 trainee placed 2 lines) did not change the statistical significance of any of the study results.

One practical concern has to do with the reproducibility of this approach at other institutions. Each trainee received up to 2 hours of individualized attention, and each session consumed fresh supplies and required a proceduralist's and an evaluator's time. This represents a significant commitment of materials and manpower. A careful cost/benefit analysis is therefore warranted before implementing this kind of rigorous training program. As mentioned, the cost of the NHTM is approximately $120 and can withstand several cannulations over a 2‐day period; the sterile supplies and central line add up to approximately another $75/evaluation. Depending on the number of interns and residents at a given institution, these costs could prove prohibitive to cash‐poor residency training programs. In the larger picture, however, catheter‐related bloodstream infections have been estimated to result in a mortality rate of 4% to 20%, and a single catheter‐related bloodstream infection can cost up to $45,000.2124 In addition, new Medicare reimbursement policies are now beginning to limit hospital reimbursement for these types of iatrogenic events; hence, narrowing the margin of error and putting even greater financial pressures on hospitals.25 It is our belief, therefore, that an up‐front investment in NHTMs (or an alternative simulator), basic supplies, and the necessary trainer time will prove to be cost‐effective and enlightened investments from forward‐thinking leadership.

Last, we are also aware that our study did not look at whether our trainees' improved performance on the NHTM actually translated into better patient outcomes. Since patient safety is our ultimate goal, and this phase of PPSI limited all of our training and evaluations to the NHTMs, future studies must ultimately evaluate how well these learned skills translate into procedure performance on actual patients. This controlled study (possibly with a see‐1 do‐1 teach‐1 control group) will be logistically challenging, but will be the most definitive manner with which to demonstrate the true value of personalized training sessions using the NHTM (or another nonhuman simulator).

PPSI‐II demonstrated that using the NHTM as the basis for training novice practitioners in a personalized, 1‐on‐1 training session led to significant improvements in measured procedural skills. Further, these skills were retained over time. This positive study contributes to the growing body of literature pointing towards the role of intensive 1‐on‐1 training with simulators to advance procedural education for clinicians. Ultimately, we aim to demonstrate that providing trainees this type of training prior to having them perform procedures on actual patients will translate into superior patient care, greater success rates, fewer minor and major complications, and lower overall patient care costs. Rather than clinging to the classic but never‐validated see‐1, do‐1, teach‐1 approach, we believe that procedural training must adapt new curricula and technologies that will help us achieve the goals of maximizing the safety and quality of care for our patients.

The Accreditation Counsel for Graduate Medical Education (ACGME) states in its Program Requirements for Residency Education in Internal Medicine that all residents must develop technical proficiency in several procedures, including central venous line placement.1 Developing competency in common procedural skills has long been a part of medical training. The philosophy of see‐1, do‐1, teach‐1 is still the most common means by which most residents seek to obtain this proficiency, even though serious concerns have been raised about this approach.2 A typical first experience in central line placement usually involves an eager (or terrified) trainee making several clumsy attempts on an actual patient, under the hurried guidance of a senior resident who themselves received an unknown degree of training. In this scenario, rarely does standardized instruction, formal evaluation, or structured follow‐up occur.

A revitalized emphasis is now being placed on patient safety in healthcare, including an industry‐wide commitment to minimizing procedural complications. The most common complications associated to central line placement include vascular damage and catheter‐related bloodstream infections. A number of creative approaches are being developed to improve the quality of instruction on proper procedural techniques, all varying considerably in sophistication, scope, and rigor. Examples include the use of computer‐assisted methods for training ultrasound‐guided needle insertion techniques and ureteroscopy training, hands‐on training with synthetic models for thoracentesis training, video training, and uterine aspiration using papayas.311 Implicit in this trend is recognition that we, as educators, healthcare providers, and patient advocates, must design more cost effective and efficient ways to teach medical and surgical procedural techniques to clinicians.

Our approach was previously described in phase I of the Procedural Patient Safety Initiative (PPSI).12 In PPSI‐I, we introduced a nonhuman tissue model (NHTM; Figure 1) as the basis for teaching physicians a more rigorous curriculum of essential central line placement skills. By way of brief review, the NHTMs were constructed by tunneling 0.2‐mm‐thick rubber tubing (vessels) lengthwise through raw, whole chickens purchased at the grocery store. The vessels were filled with colored water to simulate blood. The NHTM has several unique features, including: (1) realistic‐appearing vessels when viewed under ultrasound, which mimic the appearance of human internal jugular veins and carotid arteries (Figure 2); (2) tissue turgor and vessel composition that produce realistic pops and flashes upon puncture and allow for multiple cannulations; (3) the ability to perform a complete central line placement (including wire advancement, dilation, line insertion, suturing, and sterile dressing placement); (4) cost effectiveness relative to other commercially‐available products (each NHTM costs $120 and can withstand multiple cannulations over 2 days).1317 During the training sessions of Phase I, participants were oriented to the ultrasound machines, shown the contents of our central line kit, and taught the principles of wide sterile barriers (WSBs), sharps safety, and vascular access under real‐time ultrasound guidance. A self‐completed survey tool was filled out by the participants before and after the session that contained questions about their precourse baseline procedural experience, and their subjective comfort level with specific skills after the course. The results of our intervention, as measured by the responses to the survey, were significantly positive. We recognized the limitations of these results based on using subjective criteria to measure efficacy, a lack of follow‐up on participants' skill retention, and with no ultimate evaluation of procedural competency evaluations on actual patients (compared to an untrained control group).

Figure 1

Figure 2

Our ultimate goal is to validate a curriculum that will give trainees the necessary education and skills that enables them to make a smooth, competent, and complication‐free transition to live patient procedures. Phase II of PPSI is our next step toward this goal. In this study, we sought to measure the impact of intensive, 1:1 central line placement training with a proceduralist, objectively validate the efficacy of the NHTM and our training curriculum using a standardized 6‐point scoring scale and skilled evaluators, and to evaluate the degree of skill retention over time (decay). Our hypothesis was that the depth of skills' imprinting from a single, standardized training session would result in a significant improvement in measured procedure skills immediately after the trainee is taught the skills, and that the retention of these skills would be demonstrable when participants were reevaluated at a future date.

Methods

PPSI‐II was an observational, prospective study conducted by The Procedure Center at Cedars‐Sinai Medical Center, a 900+‐bed, community‐based teaching hospital. The Procedure Center is staffed by dedicated Proceduralists who perform a number of common medical procedures on a daily basis and are facile with both real‐time ultrasound guidance and proper procedural techniques.18, 19 Our target population was the incoming Internal Medicine residents. Subjects were recruited by email prior to orientation week and were offered the option of participating in our study. Our only exclusion criterion was the prior observation or placement of 10 or more central lines. The study was approved by our hospital's Institutional Review Board prior to initiating recruitment. Those who chose not to participate underwent the standard orientation training required by our institution, which included a brief overview lecture on the topic of central lines and ultrasound‐guidance, a group viewing of the New England Journal of Medicine (NEJM) video on central line placement,20 and small‐group (4 participants/group) hands‐on practice sessions lasting 45 minutes with NHTMs and a trained Proceduralist.

All of the evaluations for Phase II were done using the Central Line Placement Skill Assessment Tool depicted in Figure 3. This tool, which was developed by Cedars‐Sinai Proceduralists solely for the purposes of this study, is a comprehensive step‐by‐step checklist delineating the specific steps necessary to place a sterile, ultrasound‐guided central venous catheter. It was closely derived from a central line insertion checklist that was created by the Procedure Center 3 years ago to help guide novice clinicians through the procedure, and has since been widely used throughout the institution during the placement of hundreds of central lines. The scoring system, also devised by Cedars‐Sinai Proceduralists, was based on over 15 years of experience supervising and teaching hundreds of residents on proper central line insertion techniques. It consists of clear definitions for each score that were agreed upon via consensus amongst study coordinators. Prior to any evaluations being conducted, we put our 2 evaluators (both senior medicine residents) through identical and simultaneous scoring training with the Proceduralist trainer to standardize procedural knowledge and scoring methodology.

Figure 3

A total of 20 incoming interns (trainees) out of a possible 54 invitations (37%) volunteered to participate. Each trainee was randomly assigned a number from 1 to 20. The study began for each trainee with a brief, 5‐question survey to determine their prior procedural experience (Table 1). Next, each trainee watched the NEJM online training video on central line placement.20 They were then brought into a training room that contained an NHTM sitting on a Mayo stand, an ultrasound machine, and all the materials required to place a central line insertion under ultrasound‐guidance. The trainee's baseline central line insertion skills were evaluated on 22 unique procedure steps, with each score being given by 1 of the 2 evaluators (initial evaluation). The trainee did not receive any guidance or suggestions during this initial evaluation unless the trainee reached an impasse. In these cases, the evaluator completed that single step on the trainee's behalf and then allowed the session to resume. The identity of the evaluator was indicated on each evaluation form, and after each of the evaluations the completed assessment tool was given to our blinded assistant for data recording.

Each trainee was then given a personalized, hands‐on training session by a proceduralist, using the checklist as a guide to take them through all the steps of a central line insertion. The trainee was allowed to observe and practice each skill for an unlimited period of time with the proceduralist present, until he or she demonstrated competency and felt confident enough with their independent skills (in both trainee's and proceduralist's judgment) to move forward. The entire session ended only when all steps had been taught and practiced to the proceduralist's satisfaction, the trainee felt comfortable independently performing each step (and in proper sequence), and all questions had been answered. At no time was there an imposition of time constraints or external pressure from study coordinators.

Immediately following this training session the proceduralist and trainee left the room, the procedure room was reset by an evaluator (taking approximately 5‐10 minutes), and then the trainee submitted themselves to an immediate posttraining evaluation (immediate evaluation). As with the initial assessment, the evaluator did not interfere or make any comments or suggestions during the evaluation periods, unless the trainee reached an impasse at any step. In that case, the trainee would receive a 0 for that step, the evaluator would assist them to complete that step only, and then the session would continue. No time limits were imposed.

The final part of the study required each trainee to return for follow‐up assessment (delayed evaluation), a process that was identical to the immediate posttraining evaluation. This delayed evaluation was intended to occur between 3 to 4 weeks after the immediate posttraining session, based on trainees' schedules and availability. No refresher or practice time was permitted prior to the delayed evaluation: upon arrival, trainees wrote down on a separate piece of paper (not seen by the evaluator) the number of interim line experiences they had experienced, then they were brought directly into a fully‐prepared room, and instructed to begin. The evaluator was also blind to the trainee's scores from the 2 previous evaluation sessions.

The primary endpoints were the degree of changes in overall average scores (from the 22 steps on the assessment tool) from the initial to the immediate evaluations and from the immediate to delayed evaluations. The secondary endpoints were also based on changes in average scores from the initial to immediate and immediate to delayed evaluations, and looked at 5 essential elements (steps in the assessment tool that we deemed critical to the safe and successful placement of a central line). These essential elements included (1) hand washing; (2) creation of a WSB; (3) ultrasound‐guided vessel cannulation; (4) proper catheter placement; and (5) sharps safety. Of note, the creation of a WSB element consisted of 4 steps, each of which was analyzed separately. The average scores are reported as means standard deviations (SDs).

To determine the type of analysis that would be performed, we started by assessing the changes using paired t tests. The Kolmogorov‐Smirnov and Anderson‐Darling normality tests revealed no evidence of violations of the normality assumption, confirming that using paired t tests was valid.

To address potential contamination from residents' real experiences on the rate of their knowledge decay between the immediate evaluation and delayed follow‐up, each participant completed a brief survey before their delayed evaluation asking about interim experiences. All calculations were performed including and excluding from participants' scores with affirmative answers to control for this contamination. Last, a post‐hoc analysis was performed on participants' scores using a scatterplot and statistical analyses to control for the varying time‐to‐follow‐up.

Results

All 20 individuals completed the study, for a total of 60 evaluations (20 each of initial, immediate, and delayed). The actual training time (not including the viewing of the video) ranged between 45 to 120 minutes, depending on the trainee. Our primary endpoints are depicted in Table 2. The mean overall score on the initial evaluation was 1.0 0.8. The mean overall score for the immediate posttraining evaluation was 4.4 0.3. This improvement of 3.4 points was significant (P < 0.001; 95% CI, 3.0‐3.7). The delayed evaluations took place an average of 22 days after the training session (range, 5‐101 days), and produced an overall mean score of 4.2 0.3. This decay of 0.2 was not significant (P = 0.14; 95% CI, 0.31 to 0.05). With regard to the amount of skills decay, additional calculations were performed from the scatterplot that depicted scores and the variability in time‐to‐follow‐up. We found that even after controlling this variable, the amount of decay for the overall score remained insignificant.

The results of the secondary endpoint calculations (essential elements) are depicted in Table 3. Ultrasound‐guided vessel cannulations improved from an initial average score of 0.9 1.0 to an immediate average score of 4.2 0.5 (P < 0.001; 95% CI, 3.0‐3.7); the delayed score of 4.3 0.6 was statistically unchanged from immediate (P = 0.77; 95% CI, 0.4 to 0.3). Catheter placement skills improved from 1.1 1.1 to 4.2 0.5 (P < 0.001; 95% CI, 2.6‐3.7), and the delayed score of 4.3 0.7 was unchanged from immediate (P < 0.58; 95% CI, 0.5 to 0.3). Sharps safety also improved significantly from initial (2.0 2.3) to immediate (4.9 0.5) (P < 0.0001; 95% CI, 1.9‐3.9), and the delayed scores dropped insignificantly to 4.6 0.8 (P = 0.08; 95% CI, 0.0‐0.6). Hand washing improved significantly from an initial score of 0.9 1.9 to an immediate score of 3.5 2.2 (P < 0.001; 95% CI, 1.4‐3.7), and decayed insignificantly on the delayed evaluation to 3.0 2.3 (P = 0.53; 95% CI, 0.9 to 1.7). WSB skills consisted of 4 individual steps, all of which all improved significantly from initial to immediate scores, and had insignificant decays on the delayed evaluations (see Table 3 WSB for details).

We performed validation exercises to determine the degree of interrater agreement. Of the 60 total evaluations that were eventually performed, 11 evaluations had been performed simultaneously and independently by evaluators A and B. An analysis of the scores assigned by each evaluator to these 11 trainees revealed a high level of interrater agreement (96%). Further, we performed independent analyses of the trainees' scores as assessed by evaluator A (22 sessions) or evaluator B (27 sessions) across the initial, immediate, and delayed sessions, and we detected no statistical differences in the changes in scores (which mirrored the overall results above).

With regards to real‐life contamination between immediate scores and delayed scores, we identified 3 trainees who had placed central lines on actual patients during the interim period (2 trainees placed 1 line each, and 1 trainee placed 2 lines). We repeated all of the calculations without these participants' delayed scores and determined that the removal of their scores did not change the statistical significance of any of the study results. With regard to knowledge decay, the scatterplot comparing delayed scores to varying time‐to‐follow up revealed no correlation.

Discussion

Our study was designed to determine whether novice trainees could learn and retain proper central line placement skills on the NHTM by receiving personalized training in a relaxed, 1‐on‐1 learning environment. Success was measured by trained evaluators using a detailed evaluation tool with a 6‐point scoring scale. The results of our primary endpoints (changes in overall average scores across the 3 evaluation periods) confirmed that this type of training could quickly improve novice practitioners' skill levels from very low (initial evaluation) to significantly higher (immediate posttraining). The dropoff (decay) in skill levels was found to not be statistically significant over a period of several weeks, although we recognize that further study should be performed to establish the degree of skill decay over a longer period of time.

Because some steps in a central line insertion are more critical to the procedure's success than others (ie, a skin nick with a scalpel is less critical than vessel cannulation under ultrasound‐guidance), we analyzed 5 essential elements individually as secondary endpoints. This secondary analysis was designed to unmask any critical skill deficiencies that might otherwise have been lost in the overall analysis. For each individual essential elements step, this subanalysis similarly revealed a significant improvement from initial to immediate posttraining, and an insignificant score decay on the delayed evaluation.

We recognize a number of limitations to this study. First, the n is relatively small. A larger sample size would have allowed for greater statistical power. In addition, the scoring system used for this study was created by our Procedure Center staff and had never been truly validated elsewhere. The scoring system was transparent and logical, but we recognize that any attempt to use an interval scoring system to quantify procedural skills will be inherently imperfect; the difference between 1 and 2 is not necessarily the same as a difference between 4 and 5. Great efforts were taken to mitigate the impact of this limitation: explicit definitions were established for each score, and we put our evaluators through a rigorous scoring orientation at the outset to standardize their interpretation and use of the scoring system and assessment tool.

The variability in the amount of training time spent in each session could be considered to be a confounder. Our prior experiences training interns in small groups, however, suggested that individuals learn these skills at different paces and in different ways, and so we consider our customized approach to be an essential part of this training experience. We do recognize the practical limitations inherent in rolling out such an open‐ended approach, and program directors may face time and/or resource limitations if attempting to replicate this training strategy.

We were also aware of potential interrater variability between the evaluators. Our approach to addressing this was multifactorial: we went to great lengths to standardize evaluators' understanding of the intended scoring methodology prior to the initiation of the study. We also assessed the degree of interrater reliability once all data was collected. This analysis reinforced that both evaluators were scoring trainees in a virtually identical fashion. We attribute this consistency to the quality of the scoring system, the effectiveness of the prestudy evaluator orientation with a proceduralist, and the high degree of teamwork between the 2 evaluators that kept them closely in sync with one another throughout the study.

Evaluator bias was also a concern. While each evaluator was blinded to the trainees' prior scores, the setup associated with the different training sessions, as well as the obvious differences in performance between the trainees' initial and immediate/delayed performances, made full blinding of the evaluators difficult. The theoretical risk of evaluator bias in this study would have led to evaluators rating trainees higher in the immediate and delayed performances in order to demonstrate more dramatic results. We believe that, since the evaluators themselves did not perform the actual training, and since they did not know the previous scores for the trainee, they were less inclined to skew the scores. Video recording each performance and submitting this recording to a fully‐blinded, third‐party evaluator would have more rigorously ensured blinding than we were able to accomplish. This approach could be considered in future studies of this type.

An addition limitation involved the time‐to‐follow‐up. While a longer time interval between the immediate and delayed evaluations may have better evaluated the impact of the training and potential decay, we sought to balance this with the growing risk of contamination from real central line placement experiences as more time passed. With this issue in mind, the removal of the delayed scores from the 3 trainees who had placed central lines on actual patients in between the immediate and delayed evaluations (2 trainees placed 1 line each, and 1 trainee placed 2 lines) did not change the statistical significance of any of the study results.

One practical concern has to do with the reproducibility of this approach at other institutions. Each trainee received up to 2 hours of individualized attention, and each session consumed fresh supplies and required a proceduralist's and an evaluator's time. This represents a significant commitment of materials and manpower. A careful cost/benefit analysis is therefore warranted before implementing this kind of rigorous training program. As mentioned, the cost of the NHTM is approximately $120 and can withstand several cannulations over a 2‐day period; the sterile supplies and central line add up to approximately another $75/evaluation. Depending on the number of interns and residents at a given institution, these costs could prove prohibitive to cash‐poor residency training programs. In the larger picture, however, catheter‐related bloodstream infections have been estimated to result in a mortality rate of 4% to 20%, and a single catheter‐related bloodstream infection can cost up to $45,000.2124 In addition, new Medicare reimbursement policies are now beginning to limit hospital reimbursement for these types of iatrogenic events; hence, narrowing the margin of error and putting even greater financial pressures on hospitals.25 It is our belief, therefore, that an up‐front investment in NHTMs (or an alternative simulator), basic supplies, and the necessary trainer time will prove to be cost‐effective and enlightened investments from forward‐thinking leadership.

Last, we are also aware that our study did not look at whether our trainees' improved performance on the NHTM actually translated into better patient outcomes. Since patient safety is our ultimate goal, and this phase of PPSI limited all of our training and evaluations to the NHTMs, future studies must ultimately evaluate how well these learned skills translate into procedure performance on actual patients. This controlled study (possibly with a see‐1 do‐1 teach‐1 control group) will be logistically challenging, but will be the most definitive manner with which to demonstrate the true value of personalized training sessions using the NHTM (or another nonhuman simulator).

PPSI‐II demonstrated that using the NHTM as the basis for training novice practitioners in a personalized, 1‐on‐1 training session led to significant improvements in measured procedural skills. Further, these skills were retained over time. This positive study contributes to the growing body of literature pointing towards the role of intensive 1‐on‐1 training with simulators to advance procedural education for clinicians. Ultimately, we aim to demonstrate that providing trainees this type of training prior to having them perform procedures on actual patients will translate into superior patient care, greater success rates, fewer minor and major complications, and lower overall patient care costs. Rather than clinging to the classic but never‐validated see‐1, do‐1, teach‐1 approach, we believe that procedural training must adapt new curricula and technologies that will help us achieve the goals of maximizing the safety and quality of care for our patients.

The Accreditation Counsel for Graduate Medical Education (ACGME) states in its Program Requirements for Residency Education in Internal Medicine that all residents must develop technical proficiency in several procedures, including central venous line placement.1 Developing competency in common procedural skills has long been a part of medical training. The philosophy of see‐1, do‐1, teach‐1 is still the most common means by which most residents seek to obtain this proficiency, even though serious concerns have been raised about this approach.2 A typical first experience in central line placement usually involves an eager (or terrified) trainee making several clumsy attempts on an actual patient, under the hurried guidance of a senior resident who themselves received an unknown degree of training. In this scenario, rarely does standardized instruction, formal evaluation, or structured follow‐up occur.

A revitalized emphasis is now being placed on patient safety in healthcare, including an industry‐wide commitment to minimizing procedural complications. The most common complications associated to central line placement include vascular damage and catheter‐related bloodstream infections. A number of creative approaches are being developed to improve the quality of instruction on proper procedural techniques, all varying considerably in sophistication, scope, and rigor. Examples include the use of computer‐assisted methods for training ultrasound‐guided needle insertion techniques and ureteroscopy training, hands‐on training with synthetic models for thoracentesis training, video training, and uterine aspiration using papayas.311 Implicit in this trend is recognition that we, as educators, healthcare providers, and patient advocates, must design more cost effective and efficient ways to teach medical and surgical procedural techniques to clinicians.

Our approach was previously described in phase I of the Procedural Patient Safety Initiative (PPSI).12 In PPSI‐I, we introduced a nonhuman tissue model (NHTM; Figure 1) as the basis for teaching physicians a more rigorous curriculum of essential central line placement skills. By way of brief review, the NHTMs were constructed by tunneling 0.2‐mm‐thick rubber tubing (vessels) lengthwise through raw, whole chickens purchased at the grocery store. The vessels were filled with colored water to simulate blood. The NHTM has several unique features, including: (1) realistic‐appearing vessels when viewed under ultrasound, which mimic the appearance of human internal jugular veins and carotid arteries (Figure 2); (2) tissue turgor and vessel composition that produce realistic pops and flashes upon puncture and allow for multiple cannulations; (3) the ability to perform a complete central line placement (including wire advancement, dilation, line insertion, suturing, and sterile dressing placement); (4) cost effectiveness relative to other commercially‐available products (each NHTM costs $120 and can withstand multiple cannulations over 2 days).1317 During the training sessions of Phase I, participants were oriented to the ultrasound machines, shown the contents of our central line kit, and taught the principles of wide sterile barriers (WSBs), sharps safety, and vascular access under real‐time ultrasound guidance. A self‐completed survey tool was filled out by the participants before and after the session that contained questions about their precourse baseline procedural experience, and their subjective comfort level with specific skills after the course. The results of our intervention, as measured by the responses to the survey, were significantly positive. We recognized the limitations of these results based on using subjective criteria to measure efficacy, a lack of follow‐up on participants' skill retention, and with no ultimate evaluation of procedural competency evaluations on actual patients (compared to an untrained control group).

Figure 1

Figure 2

Our ultimate goal is to validate a curriculum that will give trainees the necessary education and skills that enables them to make a smooth, competent, and complication‐free transition to live patient procedures. Phase II of PPSI is our next step toward this goal. In this study, we sought to measure the impact of intensive, 1:1 central line placement training with a proceduralist, objectively validate the efficacy of the NHTM and our training curriculum using a standardized 6‐point scoring scale and skilled evaluators, and to evaluate the degree of skill retention over time (decay). Our hypothesis was that the depth of skills' imprinting from a single, standardized training session would result in a significant improvement in measured procedure skills immediately after the trainee is taught the skills, and that the retention of these skills would be demonstrable when participants were reevaluated at a future date.

Methods

PPSI‐II was an observational, prospective study conducted by The Procedure Center at Cedars‐Sinai Medical Center, a 900+‐bed, community‐based teaching hospital. The Procedure Center is staffed by dedicated Proceduralists who perform a number of common medical procedures on a daily basis and are facile with both real‐time ultrasound guidance and proper procedural techniques.18, 19 Our target population was the incoming Internal Medicine residents. Subjects were recruited by email prior to orientation week and were offered the option of participating in our study. Our only exclusion criterion was the prior observation or placement of 10 or more central lines. The study was approved by our hospital's Institutional Review Board prior to initiating recruitment. Those who chose not to participate underwent the standard orientation training required by our institution, which included a brief overview lecture on the topic of central lines and ultrasound‐guidance, a group viewing of the New England Journal of Medicine (NEJM) video on central line placement,20 and small‐group (4 participants/group) hands‐on practice sessions lasting 45 minutes with NHTMs and a trained Proceduralist.

All of the evaluations for Phase II were done using the Central Line Placement Skill Assessment Tool depicted in Figure 3. This tool, which was developed by Cedars‐Sinai Proceduralists solely for the purposes of this study, is a comprehensive step‐by‐step checklist delineating the specific steps necessary to place a sterile, ultrasound‐guided central venous catheter. It was closely derived from a central line insertion checklist that was created by the Procedure Center 3 years ago to help guide novice clinicians through the procedure, and has since been widely used throughout the institution during the placement of hundreds of central lines. The scoring system, also devised by Cedars‐Sinai Proceduralists, was based on over 15 years of experience supervising and teaching hundreds of residents on proper central line insertion techniques. It consists of clear definitions for each score that were agreed upon via consensus amongst study coordinators. Prior to any evaluations being conducted, we put our 2 evaluators (both senior medicine residents) through identical and simultaneous scoring training with the Proceduralist trainer to standardize procedural knowledge and scoring methodology.

Figure 3

A total of 20 incoming interns (trainees) out of a possible 54 invitations (37%) volunteered to participate. Each trainee was randomly assigned a number from 1 to 20. The study began for each trainee with a brief, 5‐question survey to determine their prior procedural experience (Table 1). Next, each trainee watched the NEJM online training video on central line placement.20 They were then brought into a training room that contained an NHTM sitting on a Mayo stand, an ultrasound machine, and all the materials required to place a central line insertion under ultrasound‐guidance. The trainee's baseline central line insertion skills were evaluated on 22 unique procedure steps, with each score being given by 1 of the 2 evaluators (initial evaluation). The trainee did not receive any guidance or suggestions during this initial evaluation unless the trainee reached an impasse. In these cases, the evaluator completed that single step on the trainee's behalf and then allowed the session to resume. The identity of the evaluator was indicated on each evaluation form, and after each of the evaluations the completed assessment tool was given to our blinded assistant for data recording.

Each trainee was then given a personalized, hands‐on training session by a proceduralist, using the checklist as a guide to take them through all the steps of a central line insertion. The trainee was allowed to observe and practice each skill for an unlimited period of time with the proceduralist present, until he or she demonstrated competency and felt confident enough with their independent skills (in both trainee's and proceduralist's judgment) to move forward. The entire session ended only when all steps had been taught and practiced to the proceduralist's satisfaction, the trainee felt comfortable independently performing each step (and in proper sequence), and all questions had been answered. At no time was there an imposition of time constraints or external pressure from study coordinators.

Immediately following this training session the proceduralist and trainee left the room, the procedure room was reset by an evaluator (taking approximately 5‐10 minutes), and then the trainee submitted themselves to an immediate posttraining evaluation (immediate evaluation). As with the initial assessment, the evaluator did not interfere or make any comments or suggestions during the evaluation periods, unless the trainee reached an impasse at any step. In that case, the trainee would receive a 0 for that step, the evaluator would assist them to complete that step only, and then the session would continue. No time limits were imposed.

The final part of the study required each trainee to return for follow‐up assessment (delayed evaluation), a process that was identical to the immediate posttraining evaluation. This delayed evaluation was intended to occur between 3 to 4 weeks after the immediate posttraining session, based on trainees' schedules and availability. No refresher or practice time was permitted prior to the delayed evaluation: upon arrival, trainees wrote down on a separate piece of paper (not seen by the evaluator) the number of interim line experiences they had experienced, then they were brought directly into a fully‐prepared room, and instructed to begin. The evaluator was also blind to the trainee's scores from the 2 previous evaluation sessions.

The primary endpoints were the degree of changes in overall average scores (from the 22 steps on the assessment tool) from the initial to the immediate evaluations and from the immediate to delayed evaluations. The secondary endpoints were also based on changes in average scores from the initial to immediate and immediate to delayed evaluations, and looked at 5 essential elements (steps in the assessment tool that we deemed critical to the safe and successful placement of a central line). These essential elements included (1) hand washing; (2) creation of a WSB; (3) ultrasound‐guided vessel cannulation; (4) proper catheter placement; and (5) sharps safety. Of note, the creation of a WSB element consisted of 4 steps, each of which was analyzed separately. The average scores are reported as means standard deviations (SDs).

To determine the type of analysis that would be performed, we started by assessing the changes using paired t tests. The Kolmogorov‐Smirnov and Anderson‐Darling normality tests revealed no evidence of violations of the normality assumption, confirming that using paired t tests was valid.

To address potential contamination from residents' real experiences on the rate of their knowledge decay between the immediate evaluation and delayed follow‐up, each participant completed a brief survey before their delayed evaluation asking about interim experiences. All calculations were performed including and excluding from participants' scores with affirmative answers to control for this contamination. Last, a post‐hoc analysis was performed on participants' scores using a scatterplot and statistical analyses to control for the varying time‐to‐follow‐up.

Results

All 20 individuals completed the study, for a total of 60 evaluations (20 each of initial, immediate, and delayed). The actual training time (not including the viewing of the video) ranged between 45 to 120 minutes, depending on the trainee. Our primary endpoints are depicted in Table 2. The mean overall score on the initial evaluation was 1.0 0.8. The mean overall score for the immediate posttraining evaluation was 4.4 0.3. This improvement of 3.4 points was significant (P < 0.001; 95% CI, 3.0‐3.7). The delayed evaluations took place an average of 22 days after the training session (range, 5‐101 days), and produced an overall mean score of 4.2 0.3. This decay of 0.2 was not significant (P = 0.14; 95% CI, 0.31 to 0.05). With regard to the amount of skills decay, additional calculations were performed from the scatterplot that depicted scores and the variability in time‐to‐follow‐up. We found that even after controlling this variable, the amount of decay for the overall score remained insignificant.

The results of the secondary endpoint calculations (essential elements) are depicted in Table 3. Ultrasound‐guided vessel cannulations improved from an initial average score of 0.9 1.0 to an immediate average score of 4.2 0.5 (P < 0.001; 95% CI, 3.0‐3.7); the delayed score of 4.3 0.6 was statistically unchanged from immediate (P = 0.77; 95% CI, 0.4 to 0.3). Catheter placement skills improved from 1.1 1.1 to 4.2 0.5 (P < 0.001; 95% CI, 2.6‐3.7), and the delayed score of 4.3 0.7 was unchanged from immediate (P < 0.58; 95% CI, 0.5 to 0.3). Sharps safety also improved significantly from initial (2.0 2.3) to immediate (4.9 0.5) (P < 0.0001; 95% CI, 1.9‐3.9), and the delayed scores dropped insignificantly to 4.6 0.8 (P = 0.08; 95% CI, 0.0‐0.6). Hand washing improved significantly from an initial score of 0.9 1.9 to an immediate score of 3.5 2.2 (P < 0.001; 95% CI, 1.4‐3.7), and decayed insignificantly on the delayed evaluation to 3.0 2.3 (P = 0.53; 95% CI, 0.9 to 1.7). WSB skills consisted of 4 individual steps, all of which all improved significantly from initial to immediate scores, and had insignificant decays on the delayed evaluations (see Table 3 WSB for details).

We performed validation exercises to determine the degree of interrater agreement. Of the 60 total evaluations that were eventually performed, 11 evaluations had been performed simultaneously and independently by evaluators A and B. An analysis of the scores assigned by each evaluator to these 11 trainees revealed a high level of interrater agreement (96%). Further, we performed independent analyses of the trainees' scores as assessed by evaluator A (22 sessions) or evaluator B (27 sessions) across the initial, immediate, and delayed sessions, and we detected no statistical differences in the changes in scores (which mirrored the overall results above).

With regards to real‐life contamination between immediate scores and delayed scores, we identified 3 trainees who had placed central lines on actual patients during the interim period (2 trainees placed 1 line each, and 1 trainee placed 2 lines). We repeated all of the calculations without these participants' delayed scores and determined that the removal of their scores did not change the statistical significance of any of the study results. With regard to knowledge decay, the scatterplot comparing delayed scores to varying time‐to‐follow up revealed no correlation.

Discussion

Our study was designed to determine whether novice trainees could learn and retain proper central line placement skills on the NHTM by receiving personalized training in a relaxed, 1‐on‐1 learning environment. Success was measured by trained evaluators using a detailed evaluation tool with a 6‐point scoring scale. The results of our primary endpoints (changes in overall average scores across the 3 evaluation periods) confirmed that this type of training could quickly improve novice practitioners' skill levels from very low (initial evaluation) to significantly higher (immediate posttraining). The dropoff (decay) in skill levels was found to not be statistically significant over a period of several weeks, although we recognize that further study should be performed to establish the degree of skill decay over a longer period of time.

Because some steps in a central line insertion are more critical to the procedure's success than others (ie, a skin nick with a scalpel is less critical than vessel cannulation under ultrasound‐guidance), we analyzed 5 essential elements individually as secondary endpoints. This secondary analysis was designed to unmask any critical skill deficiencies that might otherwise have been lost in the overall analysis. For each individual essential elements step, this subanalysis similarly revealed a significant improvement from initial to immediate posttraining, and an insignificant score decay on the delayed evaluation.

We recognize a number of limitations to this study. First, the n is relatively small. A larger sample size would have allowed for greater statistical power. In addition, the scoring system used for this study was created by our Procedure Center staff and had never been truly validated elsewhere. The scoring system was transparent and logical, but we recognize that any attempt to use an interval scoring system to quantify procedural skills will be inherently imperfect; the difference between 1 and 2 is not necessarily the same as a difference between 4 and 5. Great efforts were taken to mitigate the impact of this limitation: explicit definitions were established for each score, and we put our evaluators through a rigorous scoring orientation at the outset to standardize their interpretation and use of the scoring system and assessment tool.

The variability in the amount of training time spent in each session could be considered to be a confounder. Our prior experiences training interns in small groups, however, suggested that individuals learn these skills at different paces and in different ways, and so we consider our customized approach to be an essential part of this training experience. We do recognize the practical limitations inherent in rolling out such an open‐ended approach, and program directors may face time and/or resource limitations if attempting to replicate this training strategy.

We were also aware of potential interrater variability between the evaluators. Our approach to addressing this was multifactorial: we went to great lengths to standardize evaluators' understanding of the intended scoring methodology prior to the initiation of the study. We also assessed the degree of interrater reliability once all data was collected. This analysis reinforced that both evaluators were scoring trainees in a virtually identical fashion. We attribute this consistency to the quality of the scoring system, the effectiveness of the prestudy evaluator orientation with a proceduralist, and the high degree of teamwork between the 2 evaluators that kept them closely in sync with one another throughout the study.

Evaluator bias was also a concern. While each evaluator was blinded to the trainees' prior scores, the setup associated with the different training sessions, as well as the obvious differences in performance between the trainees' initial and immediate/delayed performances, made full blinding of the evaluators difficult. The theoretical risk of evaluator bias in this study would have led to evaluators rating trainees higher in the immediate and delayed performances in order to demonstrate more dramatic results. We believe that, since the evaluators themselves did not perform the actual training, and since they did not know the previous scores for the trainee, they were less inclined to skew the scores. Video recording each performance and submitting this recording to a fully‐blinded, third‐party evaluator would have more rigorously ensured blinding than we were able to accomplish. This approach could be considered in future studies of this type.

An addition limitation involved the time‐to‐follow‐up. While a longer time interval between the immediate and delayed evaluations may have better evaluated the impact of the training and potential decay, we sought to balance this with the growing risk of contamination from real central line placement experiences as more time passed. With this issue in mind, the removal of the delayed scores from the 3 trainees who had placed central lines on actual patients in between the immediate and delayed evaluations (2 trainees placed 1 line each, and 1 trainee placed 2 lines) did not change the statistical significance of any of the study results.

One practical concern has to do with the reproducibility of this approach at other institutions. Each trainee received up to 2 hours of individualized attention, and each session consumed fresh supplies and required a proceduralist's and an evaluator's time. This represents a significant commitment of materials and manpower. A careful cost/benefit analysis is therefore warranted before implementing this kind of rigorous training program. As mentioned, the cost of the NHTM is approximately $120 and can withstand several cannulations over a 2‐day period; the sterile supplies and central line add up to approximately another $75/evaluation. Depending on the number of interns and residents at a given institution, these costs could prove prohibitive to cash‐poor residency training programs. In the larger picture, however, catheter‐related bloodstream infections have been estimated to result in a mortality rate of 4% to 20%, and a single catheter‐related bloodstream infection can cost up to $45,000.2124 In addition, new Medicare reimbursement policies are now beginning to limit hospital reimbursement for these types of iatrogenic events; hence, narrowing the margin of error and putting even greater financial pressures on hospitals.25 It is our belief, therefore, that an up‐front investment in NHTMs (or an alternative simulator), basic supplies, and the necessary trainer time will prove to be cost‐effective and enlightened investments from forward‐thinking leadership.

Last, we are also aware that our study did not look at whether our trainees' improved performance on the NHTM actually translated into better patient outcomes. Since patient safety is our ultimate goal, and this phase of PPSI limited all of our training and evaluations to the NHTMs, future studies must ultimately evaluate how well these learned skills translate into procedure performance on actual patients. This controlled study (possibly with a see‐1 do‐1 teach‐1 control group) will be logistically challenging, but will be the most definitive manner with which to demonstrate the true value of personalized training sessions using the NHTM (or another nonhuman simulator).

PPSI‐II demonstrated that using the NHTM as the basis for training novice practitioners in a personalized, 1‐on‐1 training session led to significant improvements in measured procedural skills. Further, these skills were retained over time. This positive study contributes to the growing body of literature pointing towards the role of intensive 1‐on‐1 training with simulators to advance procedural education for clinicians. Ultimately, we aim to demonstrate that providing trainees this type of training prior to having them perform procedures on actual patients will translate into superior patient care, greater success rates, fewer minor and major complications, and lower overall patient care costs. Rather than clinging to the classic but never‐validated see‐1, do‐1, teach‐1 approach, we believe that procedural training must adapt new curricula and technologies that will help us achieve the goals of maximizing the safety and quality of care for our patients.

Resuscitation status and patient wishes in terms of advanced cardiopulmonary support must be addressed during inpatient hospital admissions. However, the lack of clarity of the patients' wishes and the variability in physicians' comfort addressing these issues often leads to ambiguity in an emergency setting. This may result in inappropriately aggressive management, and conversely, it may also lead to withholding potentially lifesaving therapy due to Do Not Resuscitate (DNR) designation. We report a case of hemodynamic instability due to acute supraventricular tachycardia (SVT) in a patient with a DNR designation. He was successfully treated according to the advanced cardiac life support (ACLS) protocol for SVT. We also discuss some of the ethical challenges of providing potential life‐sustaining interventions in palliative medicine, as well as the dilemma of whether or not to provide such interventions to patients who have DNR status.

Case Presentation

A 45‐year‐old man with advanced tonsillar cancer was admitted to an inpatient palliative care unit for evaluation and treatment of anorexia, progressive pain, and asthenia. He had undergone tumor debulking and neck dissection followed by adjuvant chemotherapy and external beam radiation therapy. Despite maximal therapy, the patient developed locally recurrent disease (leading to more surgery) and later, progressive metastatic disease (treated with palliative radiation therapy). With ongoing weight loss and failure to thrive, a percutaneous gastrostomy tube was placed for nutritional support. Still, the patient suffered from significant stomatitis, esophagitis, and diarrhea consistent with radiation‐induced injury, and had several admissions for dehydration and pain control.

During this and prior admissions, the patient clearly articulated his preference for DNR status. The patient was clinically declining, but was still functional, with and estimated survival of weeks to a few months. As with previous admissions, he was given intravenous fluids and parenteral opioids, and his electrolytes and vital signs normalized to his baseline. On the day of anticipated discharge, the patient was at his hemodynamic baseline (pulse of 100 beats per minute, blood pressure of 98/60 mmHg). Upon returning to bed after a shower, the patient developed acute dyspnea, weakness, and diaphoresis. Heart rate was 170 beats per minute and blood pressure was 70/50 mmHg. Intravenous normal saline boluses were given while electrocardiogram (EKG) was obtained. EKG revealed SVT with changes suggestive of demand myocardial ischemia. Carotid massage and Valsalva maneuvers were unsuccessful in converting the rhythm to sinus.

At that point, consideration was given to his DNR designation. The treating physician and patient briefly discussed the alternatives of no treatment of his arrhythmia, or alternatively, more aggressive treatment options on the Palliative Care Unit, including intravenous (IV) adenosine and direct current cardioversion. He did not have a detailed advanced directive discussing similar scenarios; he had only completed a commonly‐used, state‐issued Durable DNR form. All decided the SVT was potentially reversible and appeared to be causing many of the patient's acute symptoms; hence, aggressive treatment of the arrhythmia was in his best interest.

Despite absence of telemetry monitoring, consideration was given to IV diltiazem or metoprolol, either of which could precipitate worsening hypotension. However, the goals were to restore his previous rhythm, to relieve symptoms with a minimum of side effects and unintended effects, and to avoid intensive care unit (ICU) transfer. Intravenous adenosine and esmolol were also considered, given their shorter half‐life, potentially lower side effect profile, and ability to produce relief of the patient's distress without further complication. The pros and cons of the situation were discussed with the patient. While he desperately wanted to feel better, he wished to stay with his family where he was. He consented to a trial of adenosine, and agreed to remain on the Palliative Care Unit. The therapeutic plan was a trial of IV adenosine, and then metoprolol if necessary. He was assured that if this was unsuccessful, we would do all we could to keep him comfortable without ICU transfer. While the patient was monitored with a portable 12‐lead EKG machine, the Palliative Medicine fellow administered adenosine 6 mg IV. Predictably, the patient noted flushing, a sense of impending doom, and a short pause of asystole. This was followed by electrocardiographic conversion to sinus tachycardia at a rate of 100 beats per minute and hemodynamic and symptomatic improvement. The patient noted that his dyspnea and generalized sense of not feeling well resolved, and he was monitored for about 30 minutes without return of the SVT. The remainder of his hospitalization was uneventful, and he was discharged to home hospice the following day. He survived for another 3 weeks without return of symptoms of arrhythmia.

Discussion

Patient preferences in terms of advanced cardiopulmonary support must be addressed during hospital admission. This is in accord with recommendations from the Patient Self‐Determination Act of 1990, as well as the Joint Commission on Accreditation of Healthcare Organizations.1 Nevertheless, the number of U.S. adults with completed advance directives to guide care providers and families with preferences if personally unable to articulate them is estimated at 5% to 25%.2 Clearly‐documented wishes are particularly important in patients with advanced cancer; however, early studies show that this happens as little as 27% of the time3 in seriously ill cancer patients. In fact, oncology physicians report direct discussions about death with only 37% of their dying patients4 and cancer patients are found to have discussions at far lower rates than patients with amyotrophic lateral sclerosis despite worse survival.5

Cardiopulmonary resuscitation (CPR) and the advanced cardiac life support (ACLS) algorithms were established to treat life‐threatening arrhythmias (namely ventricular tachycardia/fibrillation) in otherwise healthy patients who experienced witnessed intraoperative arrest. Original reports of closed chest compressions were in the intraoperative or perioperative setting.6 However, benefits of rapid initiation of CPR in witnessed out‐of‐hospital cardiac arrest were later noted as providing the only reasonable hope for reduced mortality and improved neurologic outcomes.7, 8

While CPR has shown this marginal but significant difference in outcomes of witnessed out‐of‐hospital cardiac arrest, patient with advanced life‐limiting or life threatening illness tend to have even worse outcomes even if cardiac arrest is witnessed. Survival of all cardiac arrest patients to discharge has been estimated at 3% to 14% if cardiac arrest occurs outside of the hospital and 10% to 20% for witnessed, in‐hospital cardiac arrest.912 However, a recent meta‐analysis of resuscitation for cancer patients estimates overall survival to discharge at 6.2%, and less when factoring in metastatic disease (5.6%), or ICU care at time of arrest (2.2%).13

Multiple reasons have been cited regarding why patients choose to forego resuscitation or proceed with full resuscitation status despite advanced life‐threatening illness. Factors associated with refusal of CPR include being older, female, living in a nursing home and having a worsening functional status, depression, and/or an expected poor outcome.14, 15 One can speculate that fear of no longer being cared for or being abandoned may be inferred or directly stated, and this may or may not be related to socioeconomic factors, stressors outside of the medical system, or underlying depressive symptomatology, especially hopelessness. Alternatively, 1 study revealed that an unclear expectation of outcome and prognosis after cardiopulmonary arrest led some to proceed with full resuscitative measures.15

Reports differ regarding the advanced care trajectory based on patient wishes. One study of 872 critically ill cancer patients found no significant difference in application of life‐sustaining therapies regardless of presence of an advance directive.3 The SUPPORT study mentioned above was specifically designed to understand preferences for CPR.14 While SUPPORT found that foregoing CPR may be associated with a small reduction in intensity of care, there was no difference in overall hospital survival.14 Last, although advance directives are static in terms of patient's stated wishes, a patient with decision‐making capacity is able to request a shift in goals of care at any time. However, a case‐based survey of 241 responding physicians concluded that a DNR order may indeed be associated with less aggressive and/or life‐prolonging interventions, CPR notwithstanding.16 This concept of treating those with DNR status less aggressively is often born out in terms of popular perception.17 A recent study has demonstrated that patients who discuss these issues with physicians and elect a DNR status not only have fewer aggressive interventions, but also report a higher quality of life.4

A particular nidus for this confusion may be how one interprets the DNR directive. Although DNR is specifically associated with 3 basic tenets (no endotracheal intubation, no chest compressions, and no defibrillation in the setting of cardiopulmonary arrest), this designation does not substitute for intact patient decision‐making capacity in considering other supportive measures. Intermediate steps such as limited aggressive therapy orders have been suggested to provide time‐limited and goal‐limited advanced care.9 While this offers a broader array of scenarios to be considered prior to and during clinical encounters, this may also muddy the picture with impractical options and further lack of clarity in already complex situations. The Physician Orders for Life‐Sustaining Treatment (POLST) movement has taken roots in several states, targeting seriously ill patients such as the frail and elderly. The POLST provides more explicit information regarding limited advanced measures such as nutrition or antibiotics, and may be particularly useful as a prehospital decision aid.18 While the POLST, just as the traditional advance directive, may provide clinical guidance outside of situations described explicitly therein,, it may not provide further information about goals of care, (ie, Is there a situation when 1 of these measures may be acceptable?). To reiterate what was stated about traditional directives, the POLST also applies only in situations where a patient is lacking decision‐making capacity at the time of an acute event.

The designation of DNR may indeed allow for introduction of advanced care measures that may be in accord with the patient's overall wishes and clinical prognosis. Several interventions may be appropriate on a time‐limited basis. In addition to administration of adenosine or antiarrhythmics, as in the case of our patient, the use of broad‐spectrum antimicrobial therapy, vasoactive medications, and consideration for intensive monitoring may all be appropriate on a time‐limited basis. Nevertheless, without a clear understanding of the goals of limited aggressive therapy, some would argue there is always a slippery slope in terms of technology and the implementation of advanced care measures. Hence, expectations regarding perceived outcomes, goals to be achieved by the therapy, and reasonable time lines may further clarify the patient's wishes.

In this patient scenario, the administration of adenosine is generally safe, but may lead to prolonged asystole, atrial fibrillation, and ventricular tachyarrhythmias.1921 This may lead one to consider further downstream ACLS interventions, including defibrillation or atropine. From an ethical standpoint, it is valuable to consider what would have been the next step beyond this step, in terms of advanced care measures. In the case of our patient, these measures were considered, and all accepted the goals of our intervention and its limitations. While virtually all treatments provided by physicians may predispose patients to iatrogenesis, the risks and benefits of interventions are particularly important considerations in the seriously ill patient with limited life expectancy.

Iatrogenic adverse events can be serious and fatal, and occur in 4% to 9% of hospitalized patients.2224 There has been much debate about what to do for iatrogenic adverse events, particularly when patients have clearly articulated advanced directives and DNR requests. While some argue there is a higher moral duty to reverse complications resulting from physician error or treatment‐induced complication, others would feel that the fiduciary obligation is to the patient's request.25, 26 Again, in the setting of our clinical scenario, having clear, up‐front expectations about goals of care and limitation inherent were articulated as much as able.

With increasing complexity of inpatient care and team‐based models of care becoming the norm, discerning patient's wishes continuously throughout a hospital course is critical. While this responsibility previously would have fallen to the 1 coordinating clinician (ie, the primary care physician, or the patient's subspecialist), it is increasingly becoming the responsibility of all members of the team. While provider's level of prior education, exposure, and comfort may vary, several resources have attempted to address these concerns and attempted to lay a framework for overcoming barriers to these discussion and tips on empathetic and effective communication.17, 2729

Skills notwithstanding, hospitalists particularly face a challenge in communicating these tenuous issues with patients. While there is intrinsic value in having an standardized approach to these situations, hospitalists are often thrown into these difficult situations in a fragmented, nonlongitudinal fashion, further heightening the clinical and ethical tension.28, 30 However, hospitalists are also is an area where they can truly make an impact in these patients' lives at a critical juncture. Evidence suggests that regardless of the provider who broaches the subject, patients have a desire to talk about these issues.4, 14 Hospitalists may be in an advantageous position compared to their primary care or subspecialist colleagues, in that they can offer a fresh perspective and the ability to have a dialog with the patient about these issues.

Implications

While patients are entitled to die free from the intrusion of chest compression and endotracheal tubes, they are also entitled to have symptoms aggressively managed. Advanced care measures may be appropriate for symptom palliation in complex clinical situations. A careful understanding of the patient's wishes and goals of care, after thoughtful exploration, may include therapies that in isolation, appear to be extraordinary or excessive. SVT is often quickly and successfully treated at the bedside. Despite a firm DNR status, treatment with IV adenosine allowed our patient time to return home with his family.

Resuscitation status and patient wishes in terms of advanced cardiopulmonary support must be addressed during inpatient hospital admissions. However, the lack of clarity of the patients' wishes and the variability in physicians' comfort addressing these issues often leads to ambiguity in an emergency setting. This may result in inappropriately aggressive management, and conversely, it may also lead to withholding potentially lifesaving therapy due to Do Not Resuscitate (DNR) designation. We report a case of hemodynamic instability due to acute supraventricular tachycardia (SVT) in a patient with a DNR designation. He was successfully treated according to the advanced cardiac life support (ACLS) protocol for SVT. We also discuss some of the ethical challenges of providing potential life‐sustaining interventions in palliative medicine, as well as the dilemma of whether or not to provide such interventions to patients who have DNR status.

Case Presentation

A 45‐year‐old man with advanced tonsillar cancer was admitted to an inpatient palliative care unit for evaluation and treatment of anorexia, progressive pain, and asthenia. He had undergone tumor debulking and neck dissection followed by adjuvant chemotherapy and external beam radiation therapy. Despite maximal therapy, the patient developed locally recurrent disease (leading to more surgery) and later, progressive metastatic disease (treated with palliative radiation therapy). With ongoing weight loss and failure to thrive, a percutaneous gastrostomy tube was placed for nutritional support. Still, the patient suffered from significant stomatitis, esophagitis, and diarrhea consistent with radiation‐induced injury, and had several admissions for dehydration and pain control.

During this and prior admissions, the patient clearly articulated his preference for DNR status. The patient was clinically declining, but was still functional, with and estimated survival of weeks to a few months. As with previous admissions, he was given intravenous fluids and parenteral opioids, and his electrolytes and vital signs normalized to his baseline. On the day of anticipated discharge, the patient was at his hemodynamic baseline (pulse of 100 beats per minute, blood pressure of 98/60 mmHg). Upon returning to bed after a shower, the patient developed acute dyspnea, weakness, and diaphoresis. Heart rate was 170 beats per minute and blood pressure was 70/50 mmHg. Intravenous normal saline boluses were given while electrocardiogram (EKG) was obtained. EKG revealed SVT with changes suggestive of demand myocardial ischemia. Carotid massage and Valsalva maneuvers were unsuccessful in converting the rhythm to sinus.

At that point, consideration was given to his DNR designation. The treating physician and patient briefly discussed the alternatives of no treatment of his arrhythmia, or alternatively, more aggressive treatment options on the Palliative Care Unit, including intravenous (IV) adenosine and direct current cardioversion. He did not have a detailed advanced directive discussing similar scenarios; he had only completed a commonly‐used, state‐issued Durable DNR form. All decided the SVT was potentially reversible and appeared to be causing many of the patient's acute symptoms; hence, aggressive treatment of the arrhythmia was in his best interest.

Despite absence of telemetry monitoring, consideration was given to IV diltiazem or metoprolol, either of which could precipitate worsening hypotension. However, the goals were to restore his previous rhythm, to relieve symptoms with a minimum of side effects and unintended effects, and to avoid intensive care unit (ICU) transfer. Intravenous adenosine and esmolol were also considered, given their shorter half‐life, potentially lower side effect profile, and ability to produce relief of the patient's distress without further complication. The pros and cons of the situation were discussed with the patient. While he desperately wanted to feel better, he wished to stay with his family where he was. He consented to a trial of adenosine, and agreed to remain on the Palliative Care Unit. The therapeutic plan was a trial of IV adenosine, and then metoprolol if necessary. He was assured that if this was unsuccessful, we would do all we could to keep him comfortable without ICU transfer. While the patient was monitored with a portable 12‐lead EKG machine, the Palliative Medicine fellow administered adenosine 6 mg IV. Predictably, the patient noted flushing, a sense of impending doom, and a short pause of asystole. This was followed by electrocardiographic conversion to sinus tachycardia at a rate of 100 beats per minute and hemodynamic and symptomatic improvement. The patient noted that his dyspnea and generalized sense of not feeling well resolved, and he was monitored for about 30 minutes without return of the SVT. The remainder of his hospitalization was uneventful, and he was discharged to home hospice the following day. He survived for another 3 weeks without return of symptoms of arrhythmia.

Discussion

Patient preferences in terms of advanced cardiopulmonary support must be addressed during hospital admission. This is in accord with recommendations from the Patient Self‐Determination Act of 1990, as well as the Joint Commission on Accreditation of Healthcare Organizations.1 Nevertheless, the number of U.S. adults with completed advance directives to guide care providers and families with preferences if personally unable to articulate them is estimated at 5% to 25%.2 Clearly‐documented wishes are particularly important in patients with advanced cancer; however, early studies show that this happens as little as 27% of the time3 in seriously ill cancer patients. In fact, oncology physicians report direct discussions about death with only 37% of their dying patients4 and cancer patients are found to have discussions at far lower rates than patients with amyotrophic lateral sclerosis despite worse survival.5

Cardiopulmonary resuscitation (CPR) and the advanced cardiac life support (ACLS) algorithms were established to treat life‐threatening arrhythmias (namely ventricular tachycardia/fibrillation) in otherwise healthy patients who experienced witnessed intraoperative arrest. Original reports of closed chest compressions were in the intraoperative or perioperative setting.6 However, benefits of rapid initiation of CPR in witnessed out‐of‐hospital cardiac arrest were later noted as providing the only reasonable hope for reduced mortality and improved neurologic outcomes.7, 8

While CPR has shown this marginal but significant difference in outcomes of witnessed out‐of‐hospital cardiac arrest, patient with advanced life‐limiting or life threatening illness tend to have even worse outcomes even if cardiac arrest is witnessed. Survival of all cardiac arrest patients to discharge has been estimated at 3% to 14% if cardiac arrest occurs outside of the hospital and 10% to 20% for witnessed, in‐hospital cardiac arrest.912 However, a recent meta‐analysis of resuscitation for cancer patients estimates overall survival to discharge at 6.2%, and less when factoring in metastatic disease (5.6%), or ICU care at time of arrest (2.2%).13

Multiple reasons have been cited regarding why patients choose to forego resuscitation or proceed with full resuscitation status despite advanced life‐threatening illness. Factors associated with refusal of CPR include being older, female, living in a nursing home and having a worsening functional status, depression, and/or an expected poor outcome.14, 15 One can speculate that fear of no longer being cared for or being abandoned may be inferred or directly stated, and this may or may not be related to socioeconomic factors, stressors outside of the medical system, or underlying depressive symptomatology, especially hopelessness. Alternatively, 1 study revealed that an unclear expectation of outcome and prognosis after cardiopulmonary arrest led some to proceed with full resuscitative measures.15

Reports differ regarding the advanced care trajectory based on patient wishes. One study of 872 critically ill cancer patients found no significant difference in application of life‐sustaining therapies regardless of presence of an advance directive.3 The SUPPORT study mentioned above was specifically designed to understand preferences for CPR.14 While SUPPORT found that foregoing CPR may be associated with a small reduction in intensity of care, there was no difference in overall hospital survival.14 Last, although advance directives are static in terms of patient's stated wishes, a patient with decision‐making capacity is able to request a shift in goals of care at any time. However, a case‐based survey of 241 responding physicians concluded that a DNR order may indeed be associated with less aggressive and/or life‐prolonging interventions, CPR notwithstanding.16 This concept of treating those with DNR status less aggressively is often born out in terms of popular perception.17 A recent study has demonstrated that patients who discuss these issues with physicians and elect a DNR status not only have fewer aggressive interventions, but also report a higher quality of life.4

A particular nidus for this confusion may be how one interprets the DNR directive. Although DNR is specifically associated with 3 basic tenets (no endotracheal intubation, no chest compressions, and no defibrillation in the setting of cardiopulmonary arrest), this designation does not substitute for intact patient decision‐making capacity in considering other supportive measures. Intermediate steps such as limited aggressive therapy orders have been suggested to provide time‐limited and goal‐limited advanced care.9 While this offers a broader array of scenarios to be considered prior to and during clinical encounters, this may also muddy the picture with impractical options and further lack of clarity in already complex situations. The Physician Orders for Life‐Sustaining Treatment (POLST) movement has taken roots in several states, targeting seriously ill patients such as the frail and elderly. The POLST provides more explicit information regarding limited advanced measures such as nutrition or antibiotics, and may be particularly useful as a prehospital decision aid.18 While the POLST, just as the traditional advance directive, may provide clinical guidance outside of situations described explicitly therein,, it may not provide further information about goals of care, (ie, Is there a situation when 1 of these measures may be acceptable?). To reiterate what was stated about traditional directives, the POLST also applies only in situations where a patient is lacking decision‐making capacity at the time of an acute event.

The designation of DNR may indeed allow for introduction of advanced care measures that may be in accord with the patient's overall wishes and clinical prognosis. Several interventions may be appropriate on a time‐limited basis. In addition to administration of adenosine or antiarrhythmics, as in the case of our patient, the use of broad‐spectrum antimicrobial therapy, vasoactive medications, and consideration for intensive monitoring may all be appropriate on a time‐limited basis. Nevertheless, without a clear understanding of the goals of limited aggressive therapy, some would argue there is always a slippery slope in terms of technology and the implementation of advanced care measures. Hence, expectations regarding perceived outcomes, goals to be achieved by the therapy, and reasonable time lines may further clarify the patient's wishes.

In this patient scenario, the administration of adenosine is generally safe, but may lead to prolonged asystole, atrial fibrillation, and ventricular tachyarrhythmias.1921 This may lead one to consider further downstream ACLS interventions, including defibrillation or atropine. From an ethical standpoint, it is valuable to consider what would have been the next step beyond this step, in terms of advanced care measures. In the case of our patient, these measures were considered, and all accepted the goals of our intervention and its limitations. While virtually all treatments provided by physicians may predispose patients to iatrogenesis, the risks and benefits of interventions are particularly important considerations in the seriously ill patient with limited life expectancy.

Iatrogenic adverse events can be serious and fatal, and occur in 4% to 9% of hospitalized patients.2224 There has been much debate about what to do for iatrogenic adverse events, particularly when patients have clearly articulated advanced directives and DNR requests. While some argue there is a higher moral duty to reverse complications resulting from physician error or treatment‐induced complication, others would feel that the fiduciary obligation is to the patient's request.25, 26 Again, in the setting of our clinical scenario, having clear, up‐front expectations about goals of care and limitation inherent were articulated as much as able.

With increasing complexity of inpatient care and team‐based models of care becoming the norm, discerning patient's wishes continuously throughout a hospital course is critical. While this responsibility previously would have fallen to the 1 coordinating clinician (ie, the primary care physician, or the patient's subspecialist), it is increasingly becoming the responsibility of all members of the team. While provider's level of prior education, exposure, and comfort may vary, several resources have attempted to address these concerns and attempted to lay a framework for overcoming barriers to these discussion and tips on empathetic and effective communication.17, 2729

Skills notwithstanding, hospitalists particularly face a challenge in communicating these tenuous issues with patients. While there is intrinsic value in having an standardized approach to these situations, hospitalists are often thrown into these difficult situations in a fragmented, nonlongitudinal fashion, further heightening the clinical and ethical tension.28, 30 However, hospitalists are also is an area where they can truly make an impact in these patients' lives at a critical juncture. Evidence suggests that regardless of the provider who broaches the subject, patients have a desire to talk about these issues.4, 14 Hospitalists may be in an advantageous position compared to their primary care or subspecialist colleagues, in that they can offer a fresh perspective and the ability to have a dialog with the patient about these issues.

Implications

While patients are entitled to die free from the intrusion of chest compression and endotracheal tubes, they are also entitled to have symptoms aggressively managed. Advanced care measures may be appropriate for symptom palliation in complex clinical situations. A careful understanding of the patient's wishes and goals of care, after thoughtful exploration, may include therapies that in isolation, appear to be extraordinary or excessive. SVT is often quickly and successfully treated at the bedside. Despite a firm DNR status, treatment with IV adenosine allowed our patient time to return home with his family.

Resuscitation status and patient wishes in terms of advanced cardiopulmonary support must be addressed during inpatient hospital admissions. However, the lack of clarity of the patients' wishes and the variability in physicians' comfort addressing these issues often leads to ambiguity in an emergency setting. This may result in inappropriately aggressive management, and conversely, it may also lead to withholding potentially lifesaving therapy due to Do Not Resuscitate (DNR) designation. We report a case of hemodynamic instability due to acute supraventricular tachycardia (SVT) in a patient with a DNR designation. He was successfully treated according to the advanced cardiac life support (ACLS) protocol for SVT. We also discuss some of the ethical challenges of providing potential life‐sustaining interventions in palliative medicine, as well as the dilemma of whether or not to provide such interventions to patients who have DNR status.

Case Presentation

A 45‐year‐old man with advanced tonsillar cancer was admitted to an inpatient palliative care unit for evaluation and treatment of anorexia, progressive pain, and asthenia. He had undergone tumor debulking and neck dissection followed by adjuvant chemotherapy and external beam radiation therapy. Despite maximal therapy, the patient developed locally recurrent disease (leading to more surgery) and later, progressive metastatic disease (treated with palliative radiation therapy). With ongoing weight loss and failure to thrive, a percutaneous gastrostomy tube was placed for nutritional support. Still, the patient suffered from significant stomatitis, esophagitis, and diarrhea consistent with radiation‐induced injury, and had several admissions for dehydration and pain control.

During this and prior admissions, the patient clearly articulated his preference for DNR status. The patient was clinically declining, but was still functional, with and estimated survival of weeks to a few months. As with previous admissions, he was given intravenous fluids and parenteral opioids, and his electrolytes and vital signs normalized to his baseline. On the day of anticipated discharge, the patient was at his hemodynamic baseline (pulse of 100 beats per minute, blood pressure of 98/60 mmHg). Upon returning to bed after a shower, the patient developed acute dyspnea, weakness, and diaphoresis. Heart rate was 170 beats per minute and blood pressure was 70/50 mmHg. Intravenous normal saline boluses were given while electrocardiogram (EKG) was obtained. EKG revealed SVT with changes suggestive of demand myocardial ischemia. Carotid massage and Valsalva maneuvers were unsuccessful in converting the rhythm to sinus.

At that point, consideration was given to his DNR designation. The treating physician and patient briefly discussed the alternatives of no treatment of his arrhythmia, or alternatively, more aggressive treatment options on the Palliative Care Unit, including intravenous (IV) adenosine and direct current cardioversion. He did not have a detailed advanced directive discussing similar scenarios; he had only completed a commonly‐used, state‐issued Durable DNR form. All decided the SVT was potentially reversible and appeared to be causing many of the patient's acute symptoms; hence, aggressive treatment of the arrhythmia was in his best interest.

Despite absence of telemetry monitoring, consideration was given to IV diltiazem or metoprolol, either of which could precipitate worsening hypotension. However, the goals were to restore his previous rhythm, to relieve symptoms with a minimum of side effects and unintended effects, and to avoid intensive care unit (ICU) transfer. Intravenous adenosine and esmolol were also considered, given their shorter half‐life, potentially lower side effect profile, and ability to produce relief of the patient's distress without further complication. The pros and cons of the situation were discussed with the patient. While he desperately wanted to feel better, he wished to stay with his family where he was. He consented to a trial of adenosine, and agreed to remain on the Palliative Care Unit. The therapeutic plan was a trial of IV adenosine, and then metoprolol if necessary. He was assured that if this was unsuccessful, we would do all we could to keep him comfortable without ICU transfer. While the patient was monitored with a portable 12‐lead EKG machine, the Palliative Medicine fellow administered adenosine 6 mg IV. Predictably, the patient noted flushing, a sense of impending doom, and a short pause of asystole. This was followed by electrocardiographic conversion to sinus tachycardia at a rate of 100 beats per minute and hemodynamic and symptomatic improvement. The patient noted that his dyspnea and generalized sense of not feeling well resolved, and he was monitored for about 30 minutes without return of the SVT. The remainder of his hospitalization was uneventful, and he was discharged to home hospice the following day. He survived for another 3 weeks without return of symptoms of arrhythmia.

Discussion

Patient preferences in terms of advanced cardiopulmonary support must be addressed during hospital admission. This is in accord with recommendations from the Patient Self‐Determination Act of 1990, as well as the Joint Commission on Accreditation of Healthcare Organizations.1 Nevertheless, the number of U.S. adults with completed advance directives to guide care providers and families with preferences if personally unable to articulate them is estimated at 5% to 25%.2 Clearly‐documented wishes are particularly important in patients with advanced cancer; however, early studies show that this happens as little as 27% of the time3 in seriously ill cancer patients. In fact, oncology physicians report direct discussions about death with only 37% of their dying patients4 and cancer patients are found to have discussions at far lower rates than patients with amyotrophic lateral sclerosis despite worse survival.5

Cardiopulmonary resuscitation (CPR) and the advanced cardiac life support (ACLS) algorithms were established to treat life‐threatening arrhythmias (namely ventricular tachycardia/fibrillation) in otherwise healthy patients who experienced witnessed intraoperative arrest. Original reports of closed chest compressions were in the intraoperative or perioperative setting.6 However, benefits of rapid initiation of CPR in witnessed out‐of‐hospital cardiac arrest were later noted as providing the only reasonable hope for reduced mortality and improved neurologic outcomes.7, 8

While CPR has shown this marginal but significant difference in outcomes of witnessed out‐of‐hospital cardiac arrest, patient with advanced life‐limiting or life threatening illness tend to have even worse outcomes even if cardiac arrest is witnessed. Survival of all cardiac arrest patients to discharge has been estimated at 3% to 14% if cardiac arrest occurs outside of the hospital and 10% to 20% for witnessed, in‐hospital cardiac arrest.912 However, a recent meta‐analysis of resuscitation for cancer patients estimates overall survival to discharge at 6.2%, and less when factoring in metastatic disease (5.6%), or ICU care at time of arrest (2.2%).13

Multiple reasons have been cited regarding why patients choose to forego resuscitation or proceed with full resuscitation status despite advanced life‐threatening illness. Factors associated with refusal of CPR include being older, female, living in a nursing home and having a worsening functional status, depression, and/or an expected poor outcome.14, 15 One can speculate that fear of no longer being cared for or being abandoned may be inferred or directly stated, and this may or may not be related to socioeconomic factors, stressors outside of the medical system, or underlying depressive symptomatology, especially hopelessness. Alternatively, 1 study revealed that an unclear expectation of outcome and prognosis after cardiopulmonary arrest led some to proceed with full resuscitative measures.15

Reports differ regarding the advanced care trajectory based on patient wishes. One study of 872 critically ill cancer patients found no significant difference in application of life‐sustaining therapies regardless of presence of an advance directive.3 The SUPPORT study mentioned above was specifically designed to understand preferences for CPR.14 While SUPPORT found that foregoing CPR may be associated with a small reduction in intensity of care, there was no difference in overall hospital survival.14 Last, although advance directives are static in terms of patient's stated wishes, a patient with decision‐making capacity is able to request a shift in goals of care at any time. However, a case‐based survey of 241 responding physicians concluded that a DNR order may indeed be associated with less aggressive and/or life‐prolonging interventions, CPR notwithstanding.16 This concept of treating those with DNR status less aggressively is often born out in terms of popular perception.17 A recent study has demonstrated that patients who discuss these issues with physicians and elect a DNR status not only have fewer aggressive interventions, but also report a higher quality of life.4

A particular nidus for this confusion may be how one interprets the DNR directive. Although DNR is specifically associated with 3 basic tenets (no endotracheal intubation, no chest compressions, and no defibrillation in the setting of cardiopulmonary arrest), this designation does not substitute for intact patient decision‐making capacity in considering other supportive measures. Intermediate steps such as limited aggressive therapy orders have been suggested to provide time‐limited and goal‐limited advanced care.9 While this offers a broader array of scenarios to be considered prior to and during clinical encounters, this may also muddy the picture with impractical options and further lack of clarity in already complex situations. The Physician Orders for Life‐Sustaining Treatment (POLST) movement has taken roots in several states, targeting seriously ill patients such as the frail and elderly. The POLST provides more explicit information regarding limited advanced measures such as nutrition or antibiotics, and may be particularly useful as a prehospital decision aid.18 While the POLST, just as the traditional advance directive, may provide clinical guidance outside of situations described explicitly therein,, it may not provide further information about goals of care, (ie, Is there a situation when 1 of these measures may be acceptable?). To reiterate what was stated about traditional directives, the POLST also applies only in situations where a patient is lacking decision‐making capacity at the time of an acute event.

The designation of DNR may indeed allow for introduction of advanced care measures that may be in accord with the patient's overall wishes and clinical prognosis. Several interventions may be appropriate on a time‐limited basis. In addition to administration of adenosine or antiarrhythmics, as in the case of our patient, the use of broad‐spectrum antimicrobial therapy, vasoactive medications, and consideration for intensive monitoring may all be appropriate on a time‐limited basis. Nevertheless, without a clear understanding of the goals of limited aggressive therapy, some would argue there is always a slippery slope in terms of technology and the implementation of advanced care measures. Hence, expectations regarding perceived outcomes, goals to be achieved by the therapy, and reasonable time lines may further clarify the patient's wishes.

In this patient scenario, the administration of adenosine is generally safe, but may lead to prolonged asystole, atrial fibrillation, and ventricular tachyarrhythmias.1921 This may lead one to consider further downstream ACLS interventions, including defibrillation or atropine. From an ethical standpoint, it is valuable to consider what would have been the next step beyond this step, in terms of advanced care measures. In the case of our patient, these measures were considered, and all accepted the goals of our intervention and its limitations. While virtually all treatments provided by physicians may predispose patients to iatrogenesis, the risks and benefits of interventions are particularly important considerations in the seriously ill patient with limited life expectancy.

Iatrogenic adverse events can be serious and fatal, and occur in 4% to 9% of hospitalized patients.2224 There has been much debate about what to do for iatrogenic adverse events, particularly when patients have clearly articulated advanced directives and DNR requests. While some argue there is a higher moral duty to reverse complications resulting from physician error or treatment‐induced complication, others would feel that the fiduciary obligation is to the patient's request.25, 26 Again, in the setting of our clinical scenario, having clear, up‐front expectations about goals of care and limitation inherent were articulated as much as able.

With increasing complexity of inpatient care and team‐based models of care becoming the norm, discerning patient's wishes continuously throughout a hospital course is critical. While this responsibility previously would have fallen to the 1 coordinating clinician (ie, the primary care physician, or the patient's subspecialist), it is increasingly becoming the responsibility of all members of the team. While provider's level of prior education, exposure, and comfort may vary, several resources have attempted to address these concerns and attempted to lay a framework for overcoming barriers to these discussion and tips on empathetic and effective communication.17, 2729

Skills notwithstanding, hospitalists particularly face a challenge in communicating these tenuous issues with patients. While there is intrinsic value in having an standardized approach to these situations, hospitalists are often thrown into these difficult situations in a fragmented, nonlongitudinal fashion, further heightening the clinical and ethical tension.28, 30 However, hospitalists are also is an area where they can truly make an impact in these patients' lives at a critical juncture. Evidence suggests that regardless of the provider who broaches the subject, patients have a desire to talk about these issues.4, 14 Hospitalists may be in an advantageous position compared to their primary care or subspecialist colleagues, in that they can offer a fresh perspective and the ability to have a dialog with the patient about these issues.

Implications

While patients are entitled to die free from the intrusion of chest compression and endotracheal tubes, they are also entitled to have symptoms aggressively managed. Advanced care measures may be appropriate for symptom palliation in complex clinical situations. A careful understanding of the patient's wishes and goals of care, after thoughtful exploration, may include therapies that in isolation, appear to be extraordinary or excessive. SVT is often quickly and successfully treated at the bedside. Despite a firm DNR status, treatment with IV adenosine allowed our patient time to return home with his family.

Internal medicine residency programs, the major pipeline for incoming hospitalists, often provide little hands‐on experience in bedside procedures. Some residents may only insert 1 central venous catheter every 4 months on the general medicine wards,1 and others may gain little more experience during intensive care unit rotations. As seen in the survey presented by Grover et al.2 in this issue of the Journal, after 3 years of training in all types of patient care units, residents often count their accumulated experience on their fingers and toes. Such sparse experience hardly leads to expertise. Recognizing this pervasive lack of training the American Board of Internal Medicine narrowed its certification requirements for bedside procedures in 2006.3 Residents are no longer expected to perform bedside procedures but instead to know them. This important revision acknowledges that manual skills training should neither be assumed nor expendablecontinuing to do so is too risky.4 Yet as internal medicine residency programs focus their bedside procedure training on cognitive competence, the ongoing exodus of bedside procedures to the up‐market hands of subspecialists, surgeons, anesthesiologists, and interventional radiologists5 will likely accelerate.

But why should hospitalists disrupt this trend? Bedside procedures are common and not always conveniently needed during daytime hours. Roughly one‐tenth of general medicine inpatients receive a central venous catheter (CVC) insertion, a lumbar puncture, an abdominal paracentesis, or a thoracentesis.6 Among these patients, about one‐half will urgently need procedures during off‐hours. Outside of the emergency department, hospitalists will likely remain the only group of physicians available at the bedsides of general medicine inpatients 7 days a week, 24 hours per day. Thus, in developing our particular practice system to best serve our patients,7 we believe that hospitalists ought to remain principals in ensuring that inpatients have ready access to expertly performed bedside procedures.

Yet unfortunately, given the limited training in manual skills that today's internal medicine residents receive, hospitalists are increasingly less prepared to provide this access themselves.8 State‐of‐the‐art training methods developed by medical specialties that depend largely on manual skills provide promising potential solutions for both future and practicing hospitalists.9 In particular, patient simulators can provide trainees with the essential hands‐on experience they often lack. In contrast to the ad hoc see‐one, do‐one, teach‐one method in current widespread use, training with simulators has distinct advantages. First, simulators obviate the increasingly awkward consent as patients grow savvier about safety concerns and (understandably) less tolerant of a novice's need to acquire experience.10 Second, training with simulators is controlled so that anatomic variations, comorbidities, patient discomfort, and time pressuresthough important real‐world factorscan be artificially removed in the earlier cognitive and integrative stages of training.11 Third, immediate feedback, which at the bedside of real patients is often empathetically avoided or delivered in cryptic hand signals, can be unmistakably unmuted and honest in the simulator setting. Fourth, and most important to the development of expertise, simulators can be used repeatedly, allowing trainees first to become facile in the mechanics of their performance (eg, holding an ultrasound probe for real‐time guidance or knowing how it feels to enter a vein) before attempting a procedure on a patient.

Three examples of patient simulators used to train internal medicine residents in CVC insertion are presented in this issue of the Journal.1214 Using observers who adhered to objective, a priori assessment criteria, both Rosen et al.13 and Millington et al.14 carefully demonstrate that internal medicine residents' manual skills can improve with patient simulators. Given the understood importance of hands‐on experience in manual skills training,15 these anticipated findings are important validations of simulator theory. The work by Barsuk et al.12 goes further to begin to examine whether or not simulator training actually leads to improved patient outcomesthe holy grail of such research. In this observational study, compared to residents who did not undergo simulator training, those who did undergo such training had 1 fewer needle passes during successful CVC insertions. Given the relative infrequency of periprocedural complications, this study was understandably underpowered to measure true complications, relying instead on the often‐used surrogate of needle passes. Nonetheless, this work will serve as an important initial example of why simulator training may be worth the effort.

To direct participation in simulator training, we endorse selecting trainees who will perform bedside procedures in their future practice.16 Given the trend in manual skills training among internal medicine residency training programs, hospitalist programs may need to shoulder this effort themselves. Thankfully, simulator training need not be expensive. Based on transfer‐of‐learning research,17 the fidelity of the simulator is less important than the accumulated experience it can afford. Even low‐fidelity simulators, such as the store‐bought whole chicken used by Rosen et al.,13 may preserve trainees' manual skills just as effectively as the expensive, bionic, high‐fidelity simulators used by Barsuk et al.12 and Millington et al.14

Beyond the costs of training, however, hospital administrators and hospitalist group leaders have more complex externalities and opportunity costs to weigh when evaluating which physician groups should perform bedside procedures. The intuitively lower‐cost strategy for hospitals, we believe, would be to ask hospitalists to perform bedside procedures at patients' bedsides instead of asking, say, highly‐paid interventional radiologists to perform the same procedures in fully‐staffed fluoroscopy suites. There is, however, very little research to help inform these decisions. As hospitalists, we know firsthand that modern healthcare remuneration is based more on doing than on knowing. Yet, whether or not bedside procedures afford financial incentives for hospitalists is unclearmuch will depend on local factors. Regardless of the finances, we believe that hospitalists skilled in performing common bedside procedures can improve the quality and efficiency of care delivery at patients' bedsides. So, instead of a call to arms for yet another turf battle, let's continue development of state‐of‐the‐art training methods like simulators to ensure that future hospitalists can expertly perform bedside procedures. After all, fighting for improvements in patient safety is a battle that we hospitalists know how to win.

Internal medicine residency programs, the major pipeline for incoming hospitalists, often provide little hands‐on experience in bedside procedures. Some residents may only insert 1 central venous catheter every 4 months on the general medicine wards,1 and others may gain little more experience during intensive care unit rotations. As seen in the survey presented by Grover et al.2 in this issue of the Journal, after 3 years of training in all types of patient care units, residents often count their accumulated experience on their fingers and toes. Such sparse experience hardly leads to expertise. Recognizing this pervasive lack of training the American Board of Internal Medicine narrowed its certification requirements for bedside procedures in 2006.3 Residents are no longer expected to perform bedside procedures but instead to know them. This important revision acknowledges that manual skills training should neither be assumed nor expendablecontinuing to do so is too risky.4 Yet as internal medicine residency programs focus their bedside procedure training on cognitive competence, the ongoing exodus of bedside procedures to the up‐market hands of subspecialists, surgeons, anesthesiologists, and interventional radiologists5 will likely accelerate.

But why should hospitalists disrupt this trend? Bedside procedures are common and not always conveniently needed during daytime hours. Roughly one‐tenth of general medicine inpatients receive a central venous catheter (CVC) insertion, a lumbar puncture, an abdominal paracentesis, or a thoracentesis.6 Among these patients, about one‐half will urgently need procedures during off‐hours. Outside of the emergency department, hospitalists will likely remain the only group of physicians available at the bedsides of general medicine inpatients 7 days a week, 24 hours per day. Thus, in developing our particular practice system to best serve our patients,7 we believe that hospitalists ought to remain principals in ensuring that inpatients have ready access to expertly performed bedside procedures.

Yet unfortunately, given the limited training in manual skills that today's internal medicine residents receive, hospitalists are increasingly less prepared to provide this access themselves.8 State‐of‐the‐art training methods developed by medical specialties that depend largely on manual skills provide promising potential solutions for both future and practicing hospitalists.9 In particular, patient simulators can provide trainees with the essential hands‐on experience they often lack. In contrast to the ad hoc see‐one, do‐one, teach‐one method in current widespread use, training with simulators has distinct advantages. First, simulators obviate the increasingly awkward consent as patients grow savvier about safety concerns and (understandably) less tolerant of a novice's need to acquire experience.10 Second, training with simulators is controlled so that anatomic variations, comorbidities, patient discomfort, and time pressuresthough important real‐world factorscan be artificially removed in the earlier cognitive and integrative stages of training.11 Third, immediate feedback, which at the bedside of real patients is often empathetically avoided or delivered in cryptic hand signals, can be unmistakably unmuted and honest in the simulator setting. Fourth, and most important to the development of expertise, simulators can be used repeatedly, allowing trainees first to become facile in the mechanics of their performance (eg, holding an ultrasound probe for real‐time guidance or knowing how it feels to enter a vein) before attempting a procedure on a patient.

Three examples of patient simulators used to train internal medicine residents in CVC insertion are presented in this issue of the Journal.1214 Using observers who adhered to objective, a priori assessment criteria, both Rosen et al.13 and Millington et al.14 carefully demonstrate that internal medicine residents' manual skills can improve with patient simulators. Given the understood importance of hands‐on experience in manual skills training,15 these anticipated findings are important validations of simulator theory. The work by Barsuk et al.12 goes further to begin to examine whether or not simulator training actually leads to improved patient outcomesthe holy grail of such research. In this observational study, compared to residents who did not undergo simulator training, those who did undergo such training had 1 fewer needle passes during successful CVC insertions. Given the relative infrequency of periprocedural complications, this study was understandably underpowered to measure true complications, relying instead on the often‐used surrogate of needle passes. Nonetheless, this work will serve as an important initial example of why simulator training may be worth the effort.

To direct participation in simulator training, we endorse selecting trainees who will perform bedside procedures in their future practice.16 Given the trend in manual skills training among internal medicine residency training programs, hospitalist programs may need to shoulder this effort themselves. Thankfully, simulator training need not be expensive. Based on transfer‐of‐learning research,17 the fidelity of the simulator is less important than the accumulated experience it can afford. Even low‐fidelity simulators, such as the store‐bought whole chicken used by Rosen et al.,13 may preserve trainees' manual skills just as effectively as the expensive, bionic, high‐fidelity simulators used by Barsuk et al.12 and Millington et al.14

Beyond the costs of training, however, hospital administrators and hospitalist group leaders have more complex externalities and opportunity costs to weigh when evaluating which physician groups should perform bedside procedures. The intuitively lower‐cost strategy for hospitals, we believe, would be to ask hospitalists to perform bedside procedures at patients' bedsides instead of asking, say, highly‐paid interventional radiologists to perform the same procedures in fully‐staffed fluoroscopy suites. There is, however, very little research to help inform these decisions. As hospitalists, we know firsthand that modern healthcare remuneration is based more on doing than on knowing. Yet, whether or not bedside procedures afford financial incentives for hospitalists is unclearmuch will depend on local factors. Regardless of the finances, we believe that hospitalists skilled in performing common bedside procedures can improve the quality and efficiency of care delivery at patients' bedsides. So, instead of a call to arms for yet another turf battle, let's continue development of state‐of‐the‐art training methods like simulators to ensure that future hospitalists can expertly perform bedside procedures. After all, fighting for improvements in patient safety is a battle that we hospitalists know how to win.

Internal medicine residency programs, the major pipeline for incoming hospitalists, often provide little hands‐on experience in bedside procedures. Some residents may only insert 1 central venous catheter every 4 months on the general medicine wards,1 and others may gain little more experience during intensive care unit rotations. As seen in the survey presented by Grover et al.2 in this issue of the Journal, after 3 years of training in all types of patient care units, residents often count their accumulated experience on their fingers and toes. Such sparse experience hardly leads to expertise. Recognizing this pervasive lack of training the American Board of Internal Medicine narrowed its certification requirements for bedside procedures in 2006.3 Residents are no longer expected to perform bedside procedures but instead to know them. This important revision acknowledges that manual skills training should neither be assumed nor expendablecontinuing to do so is too risky.4 Yet as internal medicine residency programs focus their bedside procedure training on cognitive competence, the ongoing exodus of bedside procedures to the up‐market hands of subspecialists, surgeons, anesthesiologists, and interventional radiologists5 will likely accelerate.

But why should hospitalists disrupt this trend? Bedside procedures are common and not always conveniently needed during daytime hours. Roughly one‐tenth of general medicine inpatients receive a central venous catheter (CVC) insertion, a lumbar puncture, an abdominal paracentesis, or a thoracentesis.6 Among these patients, about one‐half will urgently need procedures during off‐hours. Outside of the emergency department, hospitalists will likely remain the only group of physicians available at the bedsides of general medicine inpatients 7 days a week, 24 hours per day. Thus, in developing our particular practice system to best serve our patients,7 we believe that hospitalists ought to remain principals in ensuring that inpatients have ready access to expertly performed bedside procedures.

Yet unfortunately, given the limited training in manual skills that today's internal medicine residents receive, hospitalists are increasingly less prepared to provide this access themselves.8 State‐of‐the‐art training methods developed by medical specialties that depend largely on manual skills provide promising potential solutions for both future and practicing hospitalists.9 In particular, patient simulators can provide trainees with the essential hands‐on experience they often lack. In contrast to the ad hoc see‐one, do‐one, teach‐one method in current widespread use, training with simulators has distinct advantages. First, simulators obviate the increasingly awkward consent as patients grow savvier about safety concerns and (understandably) less tolerant of a novice's need to acquire experience.10 Second, training with simulators is controlled so that anatomic variations, comorbidities, patient discomfort, and time pressuresthough important real‐world factorscan be artificially removed in the earlier cognitive and integrative stages of training.11 Third, immediate feedback, which at the bedside of real patients is often empathetically avoided or delivered in cryptic hand signals, can be unmistakably unmuted and honest in the simulator setting. Fourth, and most important to the development of expertise, simulators can be used repeatedly, allowing trainees first to become facile in the mechanics of their performance (eg, holding an ultrasound probe for real‐time guidance or knowing how it feels to enter a vein) before attempting a procedure on a patient.

Three examples of patient simulators used to train internal medicine residents in CVC insertion are presented in this issue of the Journal.1214 Using observers who adhered to objective, a priori assessment criteria, both Rosen et al.13 and Millington et al.14 carefully demonstrate that internal medicine residents' manual skills can improve with patient simulators. Given the understood importance of hands‐on experience in manual skills training,15 these anticipated findings are important validations of simulator theory. The work by Barsuk et al.12 goes further to begin to examine whether or not simulator training actually leads to improved patient outcomesthe holy grail of such research. In this observational study, compared to residents who did not undergo simulator training, those who did undergo such training had 1 fewer needle passes during successful CVC insertions. Given the relative infrequency of periprocedural complications, this study was understandably underpowered to measure true complications, relying instead on the often‐used surrogate of needle passes. Nonetheless, this work will serve as an important initial example of why simulator training may be worth the effort.

To direct participation in simulator training, we endorse selecting trainees who will perform bedside procedures in their future practice.16 Given the trend in manual skills training among internal medicine residency training programs, hospitalist programs may need to shoulder this effort themselves. Thankfully, simulator training need not be expensive. Based on transfer‐of‐learning research,17 the fidelity of the simulator is less important than the accumulated experience it can afford. Even low‐fidelity simulators, such as the store‐bought whole chicken used by Rosen et al.,13 may preserve trainees' manual skills just as effectively as the expensive, bionic, high‐fidelity simulators used by Barsuk et al.12 and Millington et al.14

Beyond the costs of training, however, hospital administrators and hospitalist group leaders have more complex externalities and opportunity costs to weigh when evaluating which physician groups should perform bedside procedures. The intuitively lower‐cost strategy for hospitals, we believe, would be to ask hospitalists to perform bedside procedures at patients' bedsides instead of asking, say, highly‐paid interventional radiologists to perform the same procedures in fully‐staffed fluoroscopy suites. There is, however, very little research to help inform these decisions. As hospitalists, we know firsthand that modern healthcare remuneration is based more on doing than on knowing. Yet, whether or not bedside procedures afford financial incentives for hospitalists is unclearmuch will depend on local factors. Regardless of the finances, we believe that hospitalists skilled in performing common bedside procedures can improve the quality and efficiency of care delivery at patients' bedsides. So, instead of a call to arms for yet another turf battle, let's continue development of state‐of‐the‐art training methods like simulators to ensure that future hospitalists can expertly perform bedside procedures. After all, fighting for improvements in patient safety is a battle that we hospitalists know how to win.

A 79‐year‐old woman presented from a nursing home with unusual lower extremity skin changes. Her medical history included congestive heart failure, morbid obesity, chronic lymphedema, and deep vein thrombosis with inferior vena cava filter placement. There was no history of cellulitis, filariasis, or travel to endemic areas. The patient was afebrile without adenopathy and had bilateral lower extremity edema with hyperpigmented, cobble‐stoned, hyperkeratotic skin and verrucous nodules on the inner thighs (Figures 1 and 2). Elephantiasis nostras verrucosa secondary to longstanding lymphedema and obesity was diagnosed by the dermatology consultant. The patient was treated with compression stockings and topical emollients. Elephantiasis nostras verrucosa is a rare disorder secondary to chronic noninfectious or recurrent cellulitic lymphedema that results in hyperplastic fibrotic dermal changes.1 Diagnosis is clinical, but biopsy to exclude malignancies such as Stewart‐Treves syndrome is needed in atypical cases.2 Treatment options include compression stockings, limb elevation, topical keratolytics, emollients, retinoids, and surgical debridement.2, 3

Figure 1

Figure 2

A 79‐year‐old woman presented from a nursing home with unusual lower extremity skin changes. Her medical history included congestive heart failure, morbid obesity, chronic lymphedema, and deep vein thrombosis with inferior vena cava filter placement. There was no history of cellulitis, filariasis, or travel to endemic areas. The patient was afebrile without adenopathy and had bilateral lower extremity edema with hyperpigmented, cobble‐stoned, hyperkeratotic skin and verrucous nodules on the inner thighs (Figures 1 and 2). Elephantiasis nostras verrucosa secondary to longstanding lymphedema and obesity was diagnosed by the dermatology consultant. The patient was treated with compression stockings and topical emollients. Elephantiasis nostras verrucosa is a rare disorder secondary to chronic noninfectious or recurrent cellulitic lymphedema that results in hyperplastic fibrotic dermal changes.1 Diagnosis is clinical, but biopsy to exclude malignancies such as Stewart‐Treves syndrome is needed in atypical cases.2 Treatment options include compression stockings, limb elevation, topical keratolytics, emollients, retinoids, and surgical debridement.2, 3

Figure 1

Figure 2

A 79‐year‐old woman presented from a nursing home with unusual lower extremity skin changes. Her medical history included congestive heart failure, morbid obesity, chronic lymphedema, and deep vein thrombosis with inferior vena cava filter placement. There was no history of cellulitis, filariasis, or travel to endemic areas. The patient was afebrile without adenopathy and had bilateral lower extremity edema with hyperpigmented, cobble‐stoned, hyperkeratotic skin and verrucous nodules on the inner thighs (Figures 1 and 2). Elephantiasis nostras verrucosa secondary to longstanding lymphedema and obesity was diagnosed by the dermatology consultant. The patient was treated with compression stockings and topical emollients. Elephantiasis nostras verrucosa is a rare disorder secondary to chronic noninfectious or recurrent cellulitic lymphedema that results in hyperplastic fibrotic dermal changes.1 Diagnosis is clinical, but biopsy to exclude malignancies such as Stewart‐Treves syndrome is needed in atypical cases.2 Treatment options include compression stockings, limb elevation, topical keratolytics, emollients, retinoids, and surgical debridement.2, 3

Figure 1

Figure 2

A 9‐year‐old male, with a history of immune thrombocytopenic purpura (ITP) and hypoplastic left heart syndrome (HLHS) repaired by total cavopulmonary shunt, presented to the Emergency Department with a 4‐day history of crampy abdominal pain with defecation and a 4‐day history of nonbloody diarrhea with intermittent nonbilious vomiting. The abdominal pain was diffuse and nonspecific without any radiation. He had a history of encopresis and constipation over the past 3 months. No anorexia was noted and the pain did not keep him from his activities of daily living. Review of all other systems was noncontributory.

His past medical history consisted of HLHS repaired by total cavopulmonary shunt with excellent results. He was diagnosed with ITP about 6 weeks prior to this presentation and had been treated with intravenous immunoglobulin and oral prednisone. His home medications included prednisone (1.5 mg/kg/day), lansoprazole, digoxin, enalapril, furosemide, and warfarin.

On physical examination, his temperature was 37C, pulse rate 97, respiratory rate 16, unlabored blood pressure was 112/74 mm Hg, oxygen saturation 91% on room air, and weight was 22.8 kg. He had a grade 4/6 holosystolic murmur across the precordium and multiple healed surgical incisions. His abdomen was soft without tenderness or distention with normoactive bowel sounds. The rest of his physical exam was unremarkable. An acute abdominal series was obtained, which showed pneumatosis intestinalis (PI) of the right colon, pneumoperitoneum, and possible portal venous gas (Figure 1A). Initial laboratory evaluation including a complete blood count and a comprehensive metabolic panel; amylase, lipase, lactate, and venous blood gas were all within normal limits, with the exception of a platelet level of 75,000/L (normal, 150,000450,000 cells/mL). Stool was negative for Rotavirus antigen, Clostridium difficile toxin, Helicobacter pylori antigen, and Shiga toxin 1 and 2. Bacterial cultures and trichrome stain for ova and parasites were both negative. Stool analysis for occult blood was negative on admission and became positive during his hospital course. A contrast computed tomography (CT) of the abdomen and pelvis confirmed the findings of pneumatosis intestinalis in the cecum and ascending colon with intraperitoneal and retroperitoneal air, but did not reveal any portal venous gas (Figure 1B).

Figure 1

The patient was admitted to the Children's Hospital and made nil per os (NPO; ie, nothing by mouth), placed on ampicillin/sulbactam and metronidazole prophylaxis, and observed with serial abdominal examinations. Total parenteral nutrition (TPN) was begun on hospital day 2 and continued for 10 days. Surgical intervention was not required at the initial presentation secondary to his clinical and hemodynamic stability. Since immunosuppression from chronic steroid therapy is a known risk factor for the development of PI,1 a slow steroid taper with intravenous methylprednisolone was initiated and was transitioned to oral prednisone after he resumed oral nutrition. He remained NPO for 10 days until the PI radiographically resolved. Oral feeds were reintroduced slowly without complications or recurrence of PI.

Discussion

Three major hypotheses for the origin of bowel wall gas have been proposed: intraluminal gastrointestinal (GI) gas; bacterial production of gas; and pulmonary gas.1 The intralumenal GI gas hypothesis states that intralumenal gas translocates to the bowel wall due to increased intralumenal pressure, mucosal injury from direct trauma, reduction is size of Peyer's patches from immunosuppressive medications, or a combination of factors.1, 2 The bacterial theory proposes direct invasion of the bowel wall by gas‐producing bacteria; this hypothesis is not adequately supported by bacteriologic data.2 The pulmonary gas hypothesis states that alveolar rupture could result in dissection of air through the mediastinum to the retroperitoneum and eventually along vascular channels to the gut.2 Increased intralumenal gut pressure due to coughing also drives this dissection of gas into the bowel wall.2

Chronic immunosuppression with steroids and congenital heart disease are both known risk factors for the development of PI.3 Our patient presented with common complaints of abdominal pain, encopresis, and vomiting, with a benign exam. However, he had a radiographic finding of PI. With bowel rest, antibiotics, and TPN, our patient made a full recovery without requiring surgical intervention. With patients at higher risk for PI, there needs to be a higher index of suspicion for PI in the setting of GI complaints and hemodynamic instability. It has been reported in the literature that patients at higher risk can include those with inflammatory bowel disease, chronic pulmonary disease, immunosuppressive states such as leukemia or acquired immunodeficiency syndrome, short gut syndrome, and malignancies.1, 3 Kurbegov and Sondheimer4 published a series of 32 nonneonatal cases of PI, looking for characteristics that predicted higher risk of poor outcome. Their findings showed low serum bicarbonate and PI with free air and portal venous gas as significant predictors of poor outcome.4 A recent study by Morris et al.5 showed similar results: lactic acidosis is a predictor of poor patient outcome. The clinical examination of this patient, both initially and longitudinally, and the lack of laboratory abnormalities were the key factors in the disposition of this patient in the setting of alarming abdominal radiographs.

A 9‐year‐old male, with a history of immune thrombocytopenic purpura (ITP) and hypoplastic left heart syndrome (HLHS) repaired by total cavopulmonary shunt, presented to the Emergency Department with a 4‐day history of crampy abdominal pain with defecation and a 4‐day history of nonbloody diarrhea with intermittent nonbilious vomiting. The abdominal pain was diffuse and nonspecific without any radiation. He had a history of encopresis and constipation over the past 3 months. No anorexia was noted and the pain did not keep him from his activities of daily living. Review of all other systems was noncontributory.

His past medical history consisted of HLHS repaired by total cavopulmonary shunt with excellent results. He was diagnosed with ITP about 6 weeks prior to this presentation and had been treated with intravenous immunoglobulin and oral prednisone. His home medications included prednisone (1.5 mg/kg/day), lansoprazole, digoxin, enalapril, furosemide, and warfarin.

On physical examination, his temperature was 37C, pulse rate 97, respiratory rate 16, unlabored blood pressure was 112/74 mm Hg, oxygen saturation 91% on room air, and weight was 22.8 kg. He had a grade 4/6 holosystolic murmur across the precordium and multiple healed surgical incisions. His abdomen was soft without tenderness or distention with normoactive bowel sounds. The rest of his physical exam was unremarkable. An acute abdominal series was obtained, which showed pneumatosis intestinalis (PI) of the right colon, pneumoperitoneum, and possible portal venous gas (Figure 1A). Initial laboratory evaluation including a complete blood count and a comprehensive metabolic panel; amylase, lipase, lactate, and venous blood gas were all within normal limits, with the exception of a platelet level of 75,000/L (normal, 150,000450,000 cells/mL). Stool was negative for Rotavirus antigen, Clostridium difficile toxin, Helicobacter pylori antigen, and Shiga toxin 1 and 2. Bacterial cultures and trichrome stain for ova and parasites were both negative. Stool analysis for occult blood was negative on admission and became positive during his hospital course. A contrast computed tomography (CT) of the abdomen and pelvis confirmed the findings of pneumatosis intestinalis in the cecum and ascending colon with intraperitoneal and retroperitoneal air, but did not reveal any portal venous gas (Figure 1B).

Figure 1

The patient was admitted to the Children's Hospital and made nil per os (NPO; ie, nothing by mouth), placed on ampicillin/sulbactam and metronidazole prophylaxis, and observed with serial abdominal examinations. Total parenteral nutrition (TPN) was begun on hospital day 2 and continued for 10 days. Surgical intervention was not required at the initial presentation secondary to his clinical and hemodynamic stability. Since immunosuppression from chronic steroid therapy is a known risk factor for the development of PI,1 a slow steroid taper with intravenous methylprednisolone was initiated and was transitioned to oral prednisone after he resumed oral nutrition. He remained NPO for 10 days until the PI radiographically resolved. Oral feeds were reintroduced slowly without complications or recurrence of PI.

Discussion

Three major hypotheses for the origin of bowel wall gas have been proposed: intraluminal gastrointestinal (GI) gas; bacterial production of gas; and pulmonary gas.1 The intralumenal GI gas hypothesis states that intralumenal gas translocates to the bowel wall due to increased intralumenal pressure, mucosal injury from direct trauma, reduction is size of Peyer's patches from immunosuppressive medications, or a combination of factors.1, 2 The bacterial theory proposes direct invasion of the bowel wall by gas‐producing bacteria; this hypothesis is not adequately supported by bacteriologic data.2 The pulmonary gas hypothesis states that alveolar rupture could result in dissection of air through the mediastinum to the retroperitoneum and eventually along vascular channels to the gut.2 Increased intralumenal gut pressure due to coughing also drives this dissection of gas into the bowel wall.2

Chronic immunosuppression with steroids and congenital heart disease are both known risk factors for the development of PI.3 Our patient presented with common complaints of abdominal pain, encopresis, and vomiting, with a benign exam. However, he had a radiographic finding of PI. With bowel rest, antibiotics, and TPN, our patient made a full recovery without requiring surgical intervention. With patients at higher risk for PI, there needs to be a higher index of suspicion for PI in the setting of GI complaints and hemodynamic instability. It has been reported in the literature that patients at higher risk can include those with inflammatory bowel disease, chronic pulmonary disease, immunosuppressive states such as leukemia or acquired immunodeficiency syndrome, short gut syndrome, and malignancies.1, 3 Kurbegov and Sondheimer4 published a series of 32 nonneonatal cases of PI, looking for characteristics that predicted higher risk of poor outcome. Their findings showed low serum bicarbonate and PI with free air and portal venous gas as significant predictors of poor outcome.4 A recent study by Morris et al.5 showed similar results: lactic acidosis is a predictor of poor patient outcome. The clinical examination of this patient, both initially and longitudinally, and the lack of laboratory abnormalities were the key factors in the disposition of this patient in the setting of alarming abdominal radiographs.

A 9‐year‐old male, with a history of immune thrombocytopenic purpura (ITP) and hypoplastic left heart syndrome (HLHS) repaired by total cavopulmonary shunt, presented to the Emergency Department with a 4‐day history of crampy abdominal pain with defecation and a 4‐day history of nonbloody diarrhea with intermittent nonbilious vomiting. The abdominal pain was diffuse and nonspecific without any radiation. He had a history of encopresis and constipation over the past 3 months. No anorexia was noted and the pain did not keep him from his activities of daily living. Review of all other systems was noncontributory.

His past medical history consisted of HLHS repaired by total cavopulmonary shunt with excellent results. He was diagnosed with ITP about 6 weeks prior to this presentation and had been treated with intravenous immunoglobulin and oral prednisone. His home medications included prednisone (1.5 mg/kg/day), lansoprazole, digoxin, enalapril, furosemide, and warfarin.

On physical examination, his temperature was 37C, pulse rate 97, respiratory rate 16, unlabored blood pressure was 112/74 mm Hg, oxygen saturation 91% on room air, and weight was 22.8 kg. He had a grade 4/6 holosystolic murmur across the precordium and multiple healed surgical incisions. His abdomen was soft without tenderness or distention with normoactive bowel sounds. The rest of his physical exam was unremarkable. An acute abdominal series was obtained, which showed pneumatosis intestinalis (PI) of the right colon, pneumoperitoneum, and possible portal venous gas (Figure 1A). Initial laboratory evaluation including a complete blood count and a comprehensive metabolic panel; amylase, lipase, lactate, and venous blood gas were all within normal limits, with the exception of a platelet level of 75,000/L (normal, 150,000450,000 cells/mL). Stool was negative for Rotavirus antigen, Clostridium difficile toxin, Helicobacter pylori antigen, and Shiga toxin 1 and 2. Bacterial cultures and trichrome stain for ova and parasites were both negative. Stool analysis for occult blood was negative on admission and became positive during his hospital course. A contrast computed tomography (CT) of the abdomen and pelvis confirmed the findings of pneumatosis intestinalis in the cecum and ascending colon with intraperitoneal and retroperitoneal air, but did not reveal any portal venous gas (Figure 1B).

Figure 1

The patient was admitted to the Children's Hospital and made nil per os (NPO; ie, nothing by mouth), placed on ampicillin/sulbactam and metronidazole prophylaxis, and observed with serial abdominal examinations. Total parenteral nutrition (TPN) was begun on hospital day 2 and continued for 10 days. Surgical intervention was not required at the initial presentation secondary to his clinical and hemodynamic stability. Since immunosuppression from chronic steroid therapy is a known risk factor for the development of PI,1 a slow steroid taper with intravenous methylprednisolone was initiated and was transitioned to oral prednisone after he resumed oral nutrition. He remained NPO for 10 days until the PI radiographically resolved. Oral feeds were reintroduced slowly without complications or recurrence of PI.

Discussion

Three major hypotheses for the origin of bowel wall gas have been proposed: intraluminal gastrointestinal (GI) gas; bacterial production of gas; and pulmonary gas.1 The intralumenal GI gas hypothesis states that intralumenal gas translocates to the bowel wall due to increased intralumenal pressure, mucosal injury from direct trauma, reduction is size of Peyer's patches from immunosuppressive medications, or a combination of factors.1, 2 The bacterial theory proposes direct invasion of the bowel wall by gas‐producing bacteria; this hypothesis is not adequately supported by bacteriologic data.2 The pulmonary gas hypothesis states that alveolar rupture could result in dissection of air through the mediastinum to the retroperitoneum and eventually along vascular channels to the gut.2 Increased intralumenal gut pressure due to coughing also drives this dissection of gas into the bowel wall.2

Chronic immunosuppression with steroids and congenital heart disease are both known risk factors for the development of PI.3 Our patient presented with common complaints of abdominal pain, encopresis, and vomiting, with a benign exam. However, he had a radiographic finding of PI. With bowel rest, antibiotics, and TPN, our patient made a full recovery without requiring surgical intervention. With patients at higher risk for PI, there needs to be a higher index of suspicion for PI in the setting of GI complaints and hemodynamic instability. It has been reported in the literature that patients at higher risk can include those with inflammatory bowel disease, chronic pulmonary disease, immunosuppressive states such as leukemia or acquired immunodeficiency syndrome, short gut syndrome, and malignancies.1, 3 Kurbegov and Sondheimer4 published a series of 32 nonneonatal cases of PI, looking for characteristics that predicted higher risk of poor outcome. Their findings showed low serum bicarbonate and PI with free air and portal venous gas as significant predictors of poor outcome.4 A recent study by Morris et al.5 showed similar results: lactic acidosis is a predictor of poor patient outcome. The clinical examination of this patient, both initially and longitudinally, and the lack of laboratory abnormalities were the key factors in the disposition of this patient in the setting of alarming abdominal radiographs.

The use of peripherally inserted central catheters (PICCs) to facilitate the administration of intravenous medications and fluids has become commonplace in hospitalized patients. Clinicians often prefer PICCs over other central venous catheters due to their ease of insertion and the perception that PICCs may have lower risks than other central venous catheters. However, recent studies1, 2 have begun to suggest that the benefit derived from these devices can be offset by the development of complications such as upper extremity deep vein thrombosis (UEDVT). Since venous thromboembolism (VTE) in hospitalized patients is associated with increased morbidity, mortality, length of stay, and costs, we sought to determine the rate of VTE in a population of patients who received a PICC solely during their hospital stay.

Methods

Results