Predicting 90-day mortality after bariatric surgery: an independent, external validation of the OS-MRS prognostic risk score

Abstract

BackgroundThe Obesity Surgery Mortality Risk Score (OS-MRS) was developed using data from 1995 to 2004; it has yet to be validated for more recent patients in integrated delivery system settings. The objective of this study was to validate the OS-MRS using data from electronic health records in a distributed data network. MethodsWe conducted a retrospective cohort study of 3,817 adults who underwent an open (21.4%) or laparoscopic (78.6%) gastric bypass surgery between 2005 and 2007 in the Scalable Partnering Network. Our main outcome was all-cause mortality during the 90 days after surgery. We scored patients’ risk of mortality by adding characteristics according to the OS-MRS (i.e., 1 point for each predictor). ResultsSixteen of 3,817 (0.42/100; 95% CI, .24–.68) patients died within 90 days. The OS-MRS discriminated low-risk and high-risk patients effectively: low-risk (2 of 1,654 patients; .12 deaths/100 patients), intermediate-risk (10 of 2,008 patients; .50 deaths/100 patients), and high-risk (4 of 155 patients; 2.58 deaths/100 patients). High-risk patients were 21.3 times more likely to die in the first 90 days after surgery than low-risk patients (risk ratio = 21.3; 95% CI, 3.9–115.6). ConclusionIn these 10 U.S. healthcare delivery systems, the OS-MRS appears valid—albeit with the caveat that we observed a small number of deaths. The OS-MRS appears useful for identifying the small fraction of patients at high risk for 90-day mortality after open and laparoscopic RYGB.