PURPOSE: Accurately measuring and benchmarking outcomes is critical to quality assessment and improvement. Although existing risk adjusted databases allow for comparisons across institutions for highly standardized procedures on a homogeneous cohort, it is unclear that such data is useful to assess the competency of an individual surgeon; especially for an operation like colectomy which is performed for such varied indications. We hypothesized that a rigorous, prospective quality mechanism would allow for meaningful comparisons between individual surgeons.
Methods: All patients (pts) undergoing colectomy at a university hospital from 7/03-6/06 were entered into a mandatory prospective database. The vast majority of elective resections were performed by one of four faculty surgeons on a GI surgery service (Group 1). The remaining seven faculty surgeons typically perform urgent colectomy for pts admitted while on-call (Group 2). Two private practitioners also perform colectomy but self-report complications (Group 3). Complications in Groups 1 and 2 were recorded daily by a nurse practitioner rounding with the surgical team and adjudicated at regular team meetings. A complication index was calculated by dividing the number of total complications by the number of colectomies performed. The anastomotic leak rate and overall complications between groups and individual surgeons were compared using league table analysis and Fisher's exact test.
Results: Five hundred fifty-six colectomies were performed during the study period, 384 by Group 1. One hundred eighteen (21.2%) underwent laparoscopic (lap) resection. There was no difference in the complication rate for lap vs open procedure (32% vs 34%). The complication rate in Group 1 was 34.1% vs 43.3% for Group 2 (p=0.04). Complications were self-reported in 3 pts in Group 3 (5.9%). One Group 1 surgeon had an anastomotic leak rate of 9.9% and a complication index of 0.93 vs 3.3% and 0.53 compared to the other Group 1 surgeons (p=0.04 and p<.01 respectively).
Conclusions: A prospective surgeon-driven quality process can facilitate meaningful outcome comparisons and identify an outlier with appropriate grouping. The nature of a surgeon's practice and the method of complication detection are critical confounding variables.