Operative mortality rates among surgeons - Comparison of POSSUM and p-POSSUM scoring systems in gastrointestinal surgery

PURPOSE: The original Physiological and Operative Severity Score for the en Umeration of Mortality and morbidity and the more recent Portsmouth predict or equation for mortality scoring systems were developed to provide risk-ad justed mortality rates in general surgery. The aim of this study was to com pare crude and risk-adjusted operative mortality rates among four surgeons using the above scoring systems and assess their applicability for patients scored retrospectively. METHODS: A total of 505 consecutive patients under going major gastrointestinal surgery were analyzed, 65 percent underwent co lorectal, 27.5 percent underwent upper gastrointestinal, and 7.5 percent un derwent small-bowel surgery. The observed:predicted mortality ratios using the Physiological and Operative Severity Score for the enUmeration of Morta lity and morbidity and Portsmouth predictor equation for mortality scoring systems were calculated for each surgeon. RESULTS: The actual overall opera tive mortality rate was 11.1 percent (elective was 3.9 percent, and emergen cy was 25.1 percent). The Portsmouth predictor equation for mortality equat ion predicted a mortality rate of 11.3 percent (P = 0.51). However, the Phy siological and Operative Severity Score for the enUmeration of Mortality an d morbidity scoring system was found to overpredict death by a factor of tw o: 21.5 percent (P < 0.001). Mortality rates among the four surgeons varied from 7.6 to 14.7 percent but depended on the proportion of elective vs. em ergency surgery. The observed:predicted ratio for Portsmouth predictor equa tion for mortality was close to unity (0.905-1.067) for all surgeons, but i t was 0.45 to 0.56 for Physiological and Operative Severity Score for the e nUmeration of Mortality and morbidity. CONCLUSION: The Portsmouth predictor equation for mortality equation seems to be a more accurate predictor of m ortality in gastrointestinal surgery. It would seem to provide the best cho ice for analyzing operative mortality rates for individual surgeons, taking into account variation in case mix and fitness of patients even when score d retrospectively. This has important implications for the future assessmen t of surgeons' clinical standards and the assessment of quality of surgical care.