Comparison of 30-day outcomes after emergency general surgery procedures: potential for targeted improvement

Background: Patients who undergo emergency operations represent a high-risk population and have been shown to have a high risk of poor outcomes. Little is known, however, about the variability in the quality of emergency general surgical care across hospitals or within hospitals across different procedures. The objectives of this study were to identify risk factors associated with adverse events, to compare 30-day outcomes after 3 common emergency general surgery procedures within and across hospitals, and thus, to determine whether the quality of emergency surgical care is procedure-dependent or intrinsic to other aspects of the hospital environment.

Methods: Patients who underwent emergency appendectomy, cholecystectomy, or colorectal resection at 95 hospitals that submitted at least 20 of each procedure were identified in the 2005-2008 American College of Surgeons National Surgical Quality Improvement Project database. Outcomes of interest included 30-day overall morbidity and serious morbidity/mortality. Step-wise logistic regression generated patient-level predicted probabilities of an outcome. Based on the expected probabilities, observed to expected (O/E) ratios for each outcome, after each of the 3 procedures, were calculated for each hospital. Hospitals were divided into terciles based on O/E ratios. The agreement on hospital outcomes performance for overall morbidity and serious morbidity/mortality after appendectomy, cholecystectomy, and colorectal resection was assessed using weighted kappa statistics.

Results: Of the 30,788 appendectomies, 1,984 (6.44%) patients had any morbidity, and 1,140 (3.70%) patients had a serious morbidity or died. Of the 5,824 cholecystectomies, 503 (8.64%) patients had any morbidity, and 371 (6.37%) patients had a serious morbidity or died. Of the 8,990 colorectal resections, 4,202 (46.74%) patients had any morbidity, and 3,736 (41.56%) patients had a serious morbidity or died. For overall morbidity, O/E ratios for appendectomy ranged from 0.26 to 2.36; O/E ratios for cholecystectomy ranged from 0 to 3.04; O/E ratios for colorectal resection ranged from 0.45 to 1.51. For serious morbidity/mortality, O/E ratios for appendectomy ranged from 0.23 to 2.54; O/E ratios for cholecystectomy ranged from 0 to 4.28; O/E ratios for colorectal resection ranged from 0.59 to 1.75. Associations of risk-adjusted hospital outcomes based on tercile rank between procedures demonstrated slight but significant agreement for both overall morbidity (weighted kappa between 0.20 and 0.22) and serious morbidity/mortality (weighted kappa between 0.18 and 0.22). Despite this, 7 (7.4%) hospitals for overall morbidity and 9 (9.5%) hospitals for serious morbidity/mortality were rated in the highest (best) tercile for all procedures. Eight (8.4%) hospitals for overall morbidity and 8 (8.4%) hospitals for serious morbidity/mortality were rated in the lowest tercile for all procedures.

Conclusion: Emergency general surgery procedures, particularly colorectal resections, were associated with substantial 30-day overall morbidity and serious morbidity/mortality. Most hospitals did not have consistent risk-adjusted outcomes across all 3 procedures, but for a substantive minority of institutions (7-10%), good or bad performance was generalizable across procedures. Individual hospitals should examine their procedure-specific outcomes after emergency general surgery operations to focus quality improvement initiatives appropriately.