Predicted versus observed 30-day perioperative outcomes using the ACS NSQIP surgical risk calculator in patients...

The purpose of the study was to evaluate the accuracy of the American College of Surgeons NSQIP Surgical Risk Calculator for predicting risk-adjusted 30-day outcomes for patients undergoing partial nephrectomy (PN) for renal cell carcinoma (RCC).

A single institution, multi-surgeon, prospectively maintained database was queried for patients undergoing PN for RCC from 1998 to 2015. 21 preoperative factors were analyzed for each patient with predicted risk for 30-day complications, mortality, and length of stay (LOS) calculated. Differences between the mean predicted risk and observed rate of surgical outcomes were determined using two-sided one-sample t test with significance at p < 0.05. Subgroup analyses of outcomes stratified by surgical approach were also performed.

470 patients undergoing PN for RCC were analyzed. Comparing NSQIP predicted to observed outcomes, clinically significant underestimations occurred with rates of overall complications (9.16 vs. 16.81%, p < 0.001), surgical site infections [SSI] (1.65 vs. 2.77%, p < 0.001), urinary tract infection [UTI] (1.41 vs. 3.40%, p < 0.001), and LOS (3.25 vs. 3.73 days, p < 0.001). On subgroup analysis, 209 open PN and 261 minimally invasive PN (MIPN) were performed. The NSQIP calculator consistently underestimated overall complications, SSI, UTI, and LOS (p < 0.001) among both surgical approaches, while overestimating MIPN severe complications (p < 0.001). Clinically important differences persisted when stratifying the MIPN group by laparoscopic (N = 111) and robotic (N = 150) approaches.

The ACS NSQIP Surgical Risk Calculator had significant discrepancies among observed and predicted outcomes. Additional analyses confirmed these differences remained significant irrespective of surgical approach. These findings emphasize the need for urologic oncology-specific calculators to better predict surgical outcomes in this complex patient population.