Nationwide in-hospital mortality rate following rectal resection for rectal cancer according to annual hospital volume in Germany

Background The impact of hospital volume after rectal cancer surgery is seldom investigated. This study aimed to analyse the impact of annual rectal cancer surgery cases per hospital on postoperative mortality and failure to rescue. Methods All patients diagnosed with rectal cancer and who had a rectal resection procedure code from 2012 to 2015 were identified from nationwide administrative hospital data. Hospitals were grouped into five quintiles according to caseload. The absolute number of patients, postoperative deaths and failure to rescue (defined as in-hospital mortality after a documented postoperative complication) for severe postoperative complications were determined. Results Some 64 349 patients were identified. The overall in-house mortality rate was 3 center dot 9 per cent. The crude in-hospital mortality rate ranged from 5 center dot 3 per cent in very low-volume hospitals to 2 center dot 6 per cent in very high-volume centres, with a distinct trend between volume categories (P < 0 center dot 001). In multivariable logistic regression analysis using hospital volume as random effect, very high-volume hospitals (53 interventions/year) had a risk-adjusted odds ratio of 0 center dot 58 (95 per cent c.i. 0 center dot 47 to 0 center dot 73), compared with the baseline in-house mortality rate in very low-volume hospitals (6 interventions per year) (P < 0 center dot 001). The overall postoperative complication rate was comparable between different volume quintiles, but failure to rescue decreased significantly with increasing caseload (15 center dot 6 per cent after pulmonary embolism in the highest volume quintile versus 38 per cent in the lowest quintile; P = 0 center dot 010). Conclusion Patients who had rectal cancer surgery in high-volume hospitals showed better outcomes and reduced failure to rescue rates for severe complications than those treated in low-volume hospitals.