Variations in postdelivery infection and thrombosis by hospital teaching status

Abstract

Objective: Using a population-based cohort, we examined hospital-level variation overall and by teaching status in 2 maternal outcomes, postpartum infections, and thrombosis.

Study design: Linked birth certificate and hospital admission records for mother and infant were collected on all deliveries in Pennsylvania and California from 2004 through 2005. A risk adjustment model was created using maternal and fetal comorbidities identified by International Classification of Diseases-9 codes. Hospitals were classified as teaching (TH) or nonteaching hospitals (NTH) based on the presence of obstetrics and gynecology residents. Rates of infections and thrombosis were evaluated overall and by hospital teaching status.

Results: A total of 939,871 patients were evaluated from 402 hospitals (369 NTH and 33 TH). The unadjusted infection and venous thromboembolic events (VTE) rates were higher in TH vs NTH (infection: 2.04% vs 1.07%, P < .001; VTE: 1.04% vs 0.08%, P < .001). There was variation in the rates of these complications across hospitals, with the adjusted observed/expected ratio rates for infection and thrombosis for each hospital, ranging from 0-5.2 and 0-8.6, respectively.

Conclusion: There is substantial variation in infection and thrombosis rates among hospitals both overall and by teaching status, suggesting that these 2 outcomes may be useful measures of inpatient obstetric quality.

Keywords: delivery complications; infections; obstetrics; quality; thrombosis.