Radical prostatectomy at academic versus nonacademic institutions: a population based analysis

Purpose: Radical prostatectomy outcomes may be better at academic institutions than at nonacademic centers. We examined the effect of academic status on 5 short-term radical prostatectomy outcomes.

Materials and methods: In the Health Care Utilization Project Nationwide Inpatient Sample we focused on radical prostatectomy performed within the 7 most contemporary years (2001 to 2007). We tested the rates of homologous blood transfusions and extended length of stay, as well as intraoperative and postoperative complications stratified according to institutional academic status. Multivariable logistic regression analyses further adjusted for confounding variables.

Results: Overall 89,965 radical prostatectomies were identified, yielding a weighted national estimate of 442,811. Of those procedures 58.2% were recorded at academic institutions. Patients at academic institutions had a lower Charlson comorbidity index and more frequently had private insurance (p <0.001). Radical prostatectomy at academic institutions was associated with fewer blood transfusions (5.4% vs 7.4%), fewer postoperative complications (10.1% vs 12.9%) and lower rates of hospital stay above the median (18.0% vs 28.2%). On multivariable analyses institutional academic status exerted a protective effect on postoperative complication rates (OR 0.93, p = 0.02) and on rates of hospital stay in excess of the median (OR 0.91, p <0.001). Similarly radical prostatectomy performed at hospitals with a high annual caseload were less frequently associated with intraoperative (OR 0.8, p = 0.01) and postoperative (OR 0.63, p <0.001) complications, length of stay beyond the median (OR 0.19, p <0.001) and homologous blood transfusions (OR 0.35, p <0.001).

Conclusions: Even after adjusting for annual hospital caseload, radical prostatectomy performed at academic institutions is associated with better outcomes than radical prostatectomy performed at nonacademic institutions. This relationship illustrates averages and does not imply that academic institutions invariably offer better care.