Estimation of Mortality Risk for Vascular Operations by Trainees vs Attending Surgeons

ObjectiveThe development of the ability to estimate patient mortality preoperatively in surgical trainees has not been well studied.DesignProspective comparative study in which the expected perioperative mortality risk and the maximum tolerable mortality at which operation would still be offered were estimated by the operating surgeons immediately before planned vascular procedures.SettingUniversity vascular surgery teaching service.ParticipantsPredicted and maximum allowable mortality risks were compared between trainees and attending surgeons, with the mortality calculated using the Veterans Administration Surgical Quality Improvement Program (VASQIP) as a reference.ResultsSurveys were performed before 379 procedures over a 10-month period. The median expected mortality risk predicted by trainees (2%; interquartile range [IQR]: 1%-5%) was higher than the risk predicted by attending surgeons (1%; IQR: 0.8%-3%) (p < 0.01). The median expected mortality risk calculated by VASQIP (0.8%; IQR: 0.4%-1.7%) was less than that estimated by trainees by a median of 0.3% (IQR: 0.2%-3.2%) or and that by attending surgeons by 0.3% (IQR: 0.2-1.3%) (p < 0.01). The median maximum tolerable mortality risk predicted by trainees (10%; IQR: 5%-27.5%) was equal to the risk predicted by attending surgeons (10%; IQR: 5%-17.5%). The perioperative mortality calculated by VASQIP exceeded the maximum tolerable mortality offered by trainees or attending surgeons in 1% of cases each. Discrepancies between expected mortality and maximum tolerable mortality for trainees and attending surgeons were greater for younger (postgraduate year 1 or 2) trainees (0.8%; IQR: 0-3.0%) than for more senior (postgraduate year 4 or 5) trainees (0.4%; IQR: 0.1%-2.0%).ConclusionSurgeons in training overestimated the perioperative mortality risk of operations and were willing to tolerate a greater mortality risk compared with attending surgeons. Both trainee and attending surgeons tended to overestimate the perioperative mortality risk compared with that calculated by VASQIP.