T. L. Kindel1, D. Lomelin1, J. Jolley1, C. Krause1, N. Bills1, D. Oleynikov1 1University Of Nebraska Medical Center,General Surgery,Omaha, NE, USA
Introduction: Robotic-assisted laparoscopic prostatectomy (RALP) is the most commonly performed robotic-assisted surgical procedure in the USA. Robotic-assisted surgeries have a shorter learning curve than laparoscopy due to enhanced range of motion, increased tactile feedback, and 3-D imaging. The learning curve for RALP has traditionally been described by operating time, intraoperative blood loss, conversion rate, or margin-negative specimens. The number of cases to reach optimal performance varies by metric from 25 cases for technical proficiency to 200 cases for physician confidence. This study examines the RALP learning curve by comparing the peri- and post-operative complication rate (PCR) with regard to institutional case volume over time.
Methods: A retrospective, cross-sectional study was performed of the Healthcare and Utilization Project National Inpatient Sample (HCUP NIS) database from 2009-2011. The database was queried based on the concurrent ICD-9 codes for prostate cancer (185), radical prostatectomy (60.5) and laparoscopic robotic assistance (17.42). PCR was defined by the 21 most common RALP surgical complications. Hospital volume was calculated as the number of RALPs performed per year. Demographics and procedure data collected included patient age, co-morbidities, and length of stay. ANOVA was used to compare demographics and median tests investigated patient status (complicated, uncomplicated) by hospital volume and hospital volume by complications rates of ≤10%. Linear-by-linear association and regression analysis were conducted for trends. Association, regression analysis, and hospital volume compared to complications rate of ≤10% excluded institutions with <10 cases/year and no complications.
Results: 28,438 RALP surgeries were identified from 2009-2011 (2009, n=9,384; 2010, n=8,532; 2011, n=10,522). Mean patient age was 61.48 years. Length of stay did not change over time. The mean number of pre-existing chronic conditions was lower in 2009 than in 2010 or 2011 (p<0.001). Despite this increase in patient co-morbidities, RALP PCR for all hospitals, regardless of volume, decreased over time from 14% to 9.7% (r2=0.79, p<0.001). The average PCR decreased with increasing hospital volume, plateauing at 10%. Across years, an estimated hospital volume of ≥196 cases/year predicted a ≤10% PCR. The median institutional RALP volume achieving a ≤10% complication rate was 54 (IQR: 29-97) cases/year. Hospitals with ≤10% PCR had a greater volume of surgeries than those with a PCR >10% (p<0.001).
Conclusion: ~~As hospital volume increases, PCR decreases and patient safety improves. The average estimated PCR decreased over the study period with ≥196 cases/year predicting a ≤10% PCR. The PCR plateaued at 10% with little predicted benefit of volumes ≥196 cases/year. This creates a novel, evidence-based benchmark to improve clinical outcomes for hospitals performing RALP.