Impact of hospital resection volume and service capability on post-operative mortality following gastrectomy.

Abstract

Improved post-operative mortality following gastrectomy for cancer in hospitals with higher resection volumes has not been reported in Australia. Using a population-based study in Queensland, we aimed to compare post-operative mortality following gastrectomy between high- and low-volume hospitals stratified by their service capability. All patients undergoing gastrectomy for adenocarcinoma in Queensland between 2001 and 2015 were obtained from the Queensland Oncology Repository. Hospital service capability was defined using the 2015 Australian Institute of Health and Welfare hospital peer groupings. Hospitals were grouped into 'high-volume (≥5 gastrectomies annually), high service capability' (HVHS); 'low-volume (<5), high service capability'; and 'low-volume, low service capability' (LVLS). Negative binomial regression models were used to compare 30- and 90-day mortality rates between hospital groups adjusting for age, sex, socio-economic status, Charlson and American Society of Anesthesiologists scores, treatment regimen, stage and time-period. Potential mediation of mortality differences between hospital groups due to differences in the type of gastrectomy performed was also examined. LVLS hospitals have higher adjusted 30-day (incidence rate ratio (IRR) 2.97, 95% confidence interval (CI) 1.65-5.35) and 90-day (IRR 1.95, 95% CI 1.23-3.09) mortality rates compared with HVHS hospitals. There is no significant difference in adjusted 30-day (IRR 1.16, 95% CI 0.48-2.79) and 90-day (IRR 1.12, 95% CI 0.59-2.13) mortality rates comparing low-volume, high service capability hospitals with HVHS hospitals. The type of gastrectomy performed did not significantly influence differences in mortality compared between hospital groups. In the Australian environment, post-operative mortality following gastric cancer surgery may be optimized by centralizing gastrectomy away from hospitals characterized by LVLS.