An observational study of volume-outcome effects for robot-assisted radical prostatectomy in England.

OBJECTIVES To investigate volume-outcome relationships in robot-assisted radical prostatectomy (RARP) for cancer using data from the Hospital Episodes Statistics (HES) database for England. MATERIALS AND METHODS Data for all adult, elective RPs for cancer during the period January 2013-December 2018 (inclusive) were extracted from the HES database. The HES database records data on all National Health Service (NHS) hospital admissions in England. Data were extracted for the NHS trust and surgeon undertaking the procedure, the surgical technique used (laparoscopic, open or robot-assisted), hospital length of stay (LOS), emergency readmissions, and deaths. Multilevel modelling was used to adjust for hierarchy and covariates. RESULTS Data were available for 35 629 RPs (27 945 RARPs). The proportion of procedures conducted as RARPs increased from 53.2% in 2013 to 92.6% in 2018. For RARP, there was a significant relationship between 90-day emergency hospital readmission (primary outcome) and trust volume (odds ratio [OR] for volume decrease of 10 procedures: 0.99, 95% confidence interval [CI] 0.99-1.00; P = 0.037) and surgeon volume (OR for volume decrease of 10 procedures: 0.99, 95% CI 0.99-1.00; P = 0.013) in the previous year. From lowest to highest volume category there was a decline in the adjusted proportion of patients readmitted as an emergency at 90 days from 10.6% (0-49 procedures) to 7.0% (≥300 procedures) for trusts and from 9.4% (0-9 procedures) to 8.3% (≥100 procedures) for surgeons. LOS was also significantly associated with surgeon and trust volume, although 1-year mortality was associated with neither. CONCLUSIONS There is evidence of a volume-outcome relationship for RARP in England and minimising low-volume RARP will improve patient outcomes. Nevertheless, the observed effect size was relatively modest, and stakeholders should be realistic when evaluating the likely impact of further centralisation at a population level.