Hospital Teaching Status and the Effect on Endoscopic Retrograde Cholangiopancreatography Mortality and Complications in the United States

Abstract

Introduction: There is limited data in the literature on the outcomes of patients that undergo Endoscopic Retrograde Cholangiopancreatography (ERCP) in teaching versus non-teaching hospitals. Since ERCP has the potential for numerous complications, our aim was to assess the differences in post-ERCP outcomes of infection rates, pancreatitis, gastrointestinal (GI) bleed, perforation and mortality in teaching versus nonteaching hospitals nationwide. We hypothesized that ERCP complication rates would be higher in teaching hospitals given the involvement of trainees and the greater complexity of patient populations. Methods: Data from the National Inpatient Sample (NIS), the largest national inpatient database, was used from the years 2008 to 2012. Individuals that underwent inpatient ERCPs were identified using the International Classification of Diseases, Ninth Edition (ICD-9) codes. NIS classifies hospitals as teaching urban, non-teaching urban, and rural. AMA-approved residency programs are required to qualify as teaching hospitals. Non-teaching urban and rural hospitals were grouped together for the analysis. We then identified hospital stays complicated by pancreatitis, infection, GI hemorrhage, perforation and mortality. Using ANOVA analysis in STATA, the differences in infection rates, pancreatitis, GI hemorrhage, perforation and mortality between teaching and non-teaching hospitals were calculated. Results: A total of 296,835 inpatient ERCPs was included in this study. 75,268 ERCPs were diagnostic, while 267,990 were therapeutic; 154,553 were performed at teaching hospitals and 142,282 at non-teaching hospitals. ERCP infection rates (Odds Ratio [OR] = 1.041, 95% Confidence Interval [CI]:1.012-1.071, p 1.095-1.071, p<0.001) were higher in teaching hospitals when compared to non-teaching hospitals. There was no difference observed among the rates of post-ERCP pancreatitis or GI perforation between the two groups. Among diagnostic ERCPs, GI hemorrhage was higher in teaching compared to non-teaching hospitals (OR = 1.206 95% CI:1.068-1.362, p <0.001). Likewise, length of stay was increased in teaching hospitals (8.3 vs 7.2 days, p<0.05 for diagnostic and 6.5 vs 5.8 days, p<0.05 for therapeutic ERCPs). Conclusion: In conclusion, teaching hospitals were noted to have a higher mortality rate associated with inpatient ERCPs as well as higher rates for infection and GI hemorrhage. Additionally, length of stay was longer for patients that underwent inpatient ERCPs in teaching hospitals.782_A Figure 1. Differences in ratios of pancreatitis, GI hemorrhage, GI perforation, infection and mortality in inpatient diagnostic and therapeutic ERCPs among teaching versus non-teaching hospitals782_B Figure 2. Percentages of demographic variables among therapeutic and diagnostic ERCPs between teaching and non-teaching hospitals