Background  

Using 4-year nationwide population-based data for Taiwan, this study compared in-hospital surgical mortality rates with hospital volume for five cancer-related gastrointestinal resections.

Methods  

The study sample was drawn from the Taiwan National Health Insurance Research Database. A total of 34,715 patients, each of whom had undergone a cancer-related colectomy, gastrectomy, esophagectomy, pancreatic resection, or liver lobectomy between 2000 and 2003, were selected as the study sample. The outcome measure was in-hospital mortality. The study sample was categorized into five patient groups for each procedure, and logistic regression analyses were performed for each procedure after adjustment for hospital and patient characteristics to assess the independent association between hospital volume and in-hospital mortality.

Results  

The adjusted odds ratios showed a steady decline in mortality rates for colectomy, gastrectomy, esophagectomy, and liver lobectomy with increasing hospital volume. The adjusted mortality odds for these four procedures in very-high-volume hospitals, relative to very-low-volume hospitals, ranged from .65 to .05. As regards pancreatic resection, after adjustment for patient, clinical, and hospital factors, no statistically significant association was discernible between hospital volume and the likelihood of mortality.

Conclusions  

After adjustment for hospital and physician characteristics, in four of the five procedures, patients treated at higher-volume hospitals had lower in-hospital mortality rates than those treated at lower-volume hospitals. Our findings confirm, for the most part, the hypothesis that better outcomes are associated with higher-volume hospitals.