A retrospective assessment of prognostication in 456,685 patients undergoing elective major non-cardiac surgery

The accuracy of patient prognostication varies extensively in studies of select populations. Prognostication is an important component of patient selection for elective surgery. This retrospective study assessed patient prognostication prior to elective surgery by comparing the observed with the expected number of deaths in such patients. We used population-based administrative data to identify all adults undergoing one of 13 most common elective major non-cardiac gender-neutral surgeries in Ontario, Canada from 2002-2014. Survival status within one year of surgery was determined by linking to vital statistics. Expected death risk was determined with health administrative data and a previously derived and externally validated index. We identified 456,685 patients of which 17,266 (3.8%) died within one year of surgery. Patients whose expected one-year death risk was > 25% accounted for 5.3% of the entire cohort (n = 24,178) but 51.7% of all deaths (n = 8,927). The overall observed death risk was significantly lower than expected (standardized mortality ratio [SMR], 0.72; 95% confidence interval, 0.71 to 0.73; P < 0.0001). The SMRs were significantly < 1 (values ranged from 0.54 [partial liver resection] to 0.93 [total knee replacement]) in 11/13 (85%) surgery types. Improved outcomes were especially notable in patients with a higher expected death risk. Only 35/5,539 (0.6%) surgeons had one-year patient death risks that exceeded the population average. The observed number of deaths within one year of elective surgery is significantly lower than expected, with minimal inter-surgeon variation. These results suggest that patient selection for major elective non-cardiac surgery identified individuals with better than expected survival and whose survival was not adversely influenced by their surgery.