Relationship Between Hospital ERCP Volume, Length of Stay and Technical Outcomes

Abstract

Background: The relationship between hospital procedure volume and outcome has been recognized for various specialties and procedures. Although increasingly used and in existence for 40 yrs, to date, data on the relationship between hospital volume and outcome of ERCP is scant. We sought to examine health-related outcomes following ERCP in relation to hospital procedure volume. Methods: We used the National Inpatient Sample (NIS) database to evaluate health-related outcomes among patients who underwent ERCP from 1998-2001. Logistic and multiple regression models were used to estimate the association of hospital procedure volume with length of stay (LOS), rates of procedural failure, and mortality. Procedural failure was defined by the need to perform percutaneous transhepatic biliary drainage or open common bile duct exploration following ERCP. We used fixed effect models to adjust for all time invariant hospital characteristics for each hospital within the dataset. Results: Data from 2629 hospitals that performed 199,625 ERCP's were evaluated. Sixty-three percent of patients were females; mean age 59.4 yrs (SD = 20.1). Diagnosis included choledocholithiasis in 39.7%, neoplasm (16.6%), pancreatitis (14.1%), post-cholecystectomy syndrome (14.1%), and others (15.4%). The median number of ERCP's performed in participating hospitals was 49/yr (range:1-1004) with 25% of hospitals performing ≥ 100 ERCP/yr and 5% performing ≥ 200/yr. Significant trends in the relation between volume and outcome were observed with respect to LOS and procedural failure. The median LOS was lower in high-volume (≥200 ERCP/yr) than low-volume (≤100 ERCP/yr) hospitals (6.9 vs. 7.8 days, p < 0.0001) and the mean difference in expected LOS was 1.08 days (P < 0.0001). The rates of procedural failure was lower in high-volume than low-volume hospitals (6.7% vs. 9.3%, p < 0.0001). There was no detectable association between ERCP volume and inpatient mortality (OR 1.01; 95% CI 0.95 – 1.08). Conclusions: Patients who undergo ERCP at high-volume hospitals have shorter length of stay and lower procedural failure rates than those undergoing ERCP at low-volume hospitals. These findings have important implications for health care policy decision making and resource utilization. Background: The relationship between hospital procedure volume and outcome has been recognized for various specialties and procedures. Although increasingly used and in existence for 40 yrs, to date, data on the relationship between hospital volume and outcome of ERCP is scant. We sought to examine health-related outcomes following ERCP in relation to hospital procedure volume. Methods: We used the National Inpatient Sample (NIS) database to evaluate health-related outcomes among patients who underwent ERCP from 1998-2001. Logistic and multiple regression models were used to estimate the association of hospital procedure volume with length of stay (LOS), rates of procedural failure, and mortality. Procedural failure was defined by the need to perform percutaneous transhepatic biliary drainage or open common bile duct exploration following ERCP. We used fixed effect models to adjust for all time invariant hospital characteristics for each hospital within the dataset. Results: Data from 2629 hospitals that performed 199,625 ERCP's were evaluated. Sixty-three percent of patients were females; mean age 59.4 yrs (SD = 20.1). Diagnosis included choledocholithiasis in 39.7%, neoplasm (16.6%), pancreatitis (14.1%), post-cholecystectomy syndrome (14.1%), and others (15.4%). The median number of ERCP's performed in participating hospitals was 49/yr (range:1-1004) with 25% of hospitals performing ≥ 100 ERCP/yr and 5% performing ≥ 200/yr. Significant trends in the relation between volume and outcome were observed with respect to LOS and procedural failure. The median LOS was lower in high-volume (≥200 ERCP/yr) than low-volume (≤100 ERCP/yr) hospitals (6.9 vs. 7.8 days, p < 0.0001) and the mean difference in expected LOS was 1.08 days (P < 0.0001). The rates of procedural failure was lower in high-volume than low-volume hospitals (6.7% vs. 9.3%, p < 0.0001). There was no detectable association between ERCP volume and inpatient mortality (OR 1.01; 95% CI 0.95 – 1.08). Conclusions: Patients who undergo ERCP at high-volume hospitals have shorter length of stay and lower procedural failure rates than those undergoing ERCP at low-volume hospitals. These findings have important implications for health care policy decision making and resource utilization.