HOME
Return to Table of Contents
IMPACT OF HOSPITAL VOLUME ON OUTCOMES IN PATIENTS WITH ESOPHAGEAL VARICEAL BLEEDING: AN ANALYSIS OF THE U.S. NATIONWIDE INPATIENT SAMPLE FROM 1998 TO 2004
KD Papay, GG Kaplan, AA Shaheen, RP Myers
Liver Unit, Division of Gastroenterology, University of Calgary, Calgary, Alberta
OBJECTIVE: Esophageal variceal bleeding (EVB) is a common complication of liver cirrhosis with a high mortality rate. The objective of this study was to assess whether a volume-outcome relationship exists for patients with EVB.
METHODS: We identified adult patients hospitalized for EVB who underwent upper endoscopy between 1998 and 2004 using the U.S. Nationwide Inpatient Sample database. Patients with missing data on in-hospital mortality, elective admissions, and those undergoing liver transplantation were excluded. Hospital volume was classified based on the average annual number of EVB admissions during the study interval (low volume [<13]; medium [13-24]; high [>24]). The primary outcome was in-hospital mortality; we also examined length of hospital stay (LOS) and total hospital charges. Regression analyses were used to identify independent predictors of these outcomes including hospital volume.
RESULTS: Between 1998 and 2004, there were 31,818 hospitalizations in 2,051 hospitals for EVB (low volume [n=10,688], medium [n=10,306], high [n=10,844]). The majority of the hospitals were low volume centres (72%). Overall, in-hospital mortality was 11.1%, median LOS was 4 days (IQR 3-7), and total charges were $17,506 ($10,802-30,569). Compared with low volume centres, admission to a high volume hospital was associated with an increased risk of death (11.9%; odds ratio [OR] 1.16 [95% CI 1.06-1.26), prolonged LOS and increased total charges (P<0.0005 for both). However, patients admitted to high volume hospitals were more likely to have negative prognostic characteristics including male gender, non-Caucasian race, non-private health insurance, alcoholic cirrhosis, complications of end-stage liver disease, and to have been transferred from another institution (P<0.05 for all). After adjusting for case-mix, hospital volume was not an independent predictor of in-hospital mortality (OR vs. low volume: medium, 0.97 [95% CI 0.88-1.07]; high, 0.95 [0.87-1.05]), however, high volume centres had shorter LOS and higher total charges (P<0.0005 for both).
CONCLUSIONS: This is the largest study to show that the volume-outcome relationship observed for some surgical procedures and medical conditions does not apply to patients with EVB.