Surgical site infection following hysterectomy: adjusted rankings in a regional collaborative.

Backround: Surgical site infection after abdominal hysterectomy (defined as open and laparoscopic) will be a metric used to rank and penalize hospitals in the Hospital Acquired Condition Reduction program. Hospitals whose Hospital Acquired Condition Reduction score places them in the bottom quartile will lose 1% of reimbursement from the Centers of Medicaid and Medicare Services.

Objectives: The objectives of this analysis were to develop a risk adjustment model for surgical site infection after hysterectomy, to calculate adjusted surgical site infection rates, to rank hospitals by the predicted to expected (P/E) ratio, and to compare the number of outlier hospitals with the number in the bottom quartile.

Study Design: This was a retrospective analysis of hysterectomies from the Michigan Surgical Quality Collaborative performed between July 1, 2012, and July 1, 2014. Superficial, deep, and organ space surgical site infections were categorized according to Centers for Disease Control and Prevention criteria. Deep and organ space surgical site infections were considered 1 group for this analysis because these spaces are contiguous after hysterectomy. Hospital rankings focused on deep/organ space events because the Hospital Acquired Condition Reduction program will rank and penalize based on them, not superficial surgical site infection. Hierarchical multivariable logistic regression, which takes into account hospital effects, was used to identify risk factors for all surgical site infections and deep/organ space surgical site infections. Predicted to expected ratios for deep surgical site infection were calculated for each hospital and used to determine hospital rankings. Outliers were defined as those hospitals who predicted to expected confidence intervals crossed the reference line of 1. The number of outlier hospitals was compared with the number in the bottom quartile.

Results: The overall surgical site infection rate following hysterectomy was 2.1% (351 of 16,548). Deep/organ space surgical site infection accounted for 1.0% (n = 167 of 16,548). Deep surgical site infection was associated independently with younger age, longer surgical times, gynecological cancer, and open hysterectomy. There was a marginal association with blood transfusion. After risk adjustment of rates and ranking by the predicted to expected ratio, there was a change in quartile rank for 42.8% of hospitals (21 of 49). Two hospitals were identified as outliers. However, if the bottom quartile was identified, as called for by the Hospital Acquired Condition Reduction program, 10 additional hospitals would be targeted for a penalty. Hospitals with < 300 beds were most likely to see their quartile rank worsen, whereas those > 500 beds were most likely to see their quartile rank improve (P = .01).

Conclusion: After adjusting for patient-related factors and site variation, more than 40% of hospitals will change quartile rank with respect to deep surgical site infection. Identifying a quartile of hospitals that are statistically different from others was not feasible in our collaborative because only 2 of 12 hospitals were outliers. These findings suggest that under the Hospital Acquired Condition Reduction program, many hospitals will be unjustly penalized.