Improving Detection of Surgical Site Infections After Cesarean Delivery: What Are Diagnosis Codes Missing? [28Q]

Abstract

INTRODUCTION: Surveillance for surgical site infections (SSI) following cesarean can be challenging. We sought to identify cases of SSI over two years and describe our methods of case ascertainment. METHODS: We queried our hospital's data warehouse for cesareans deliveries from 2016-2017 among women >18 years of age. Three methods for case finding were used: i) ICD-10 diagnoses codes, ii) reporting complications in postoperative notes, and iii) infection control surveillance activities. For ICD-10 codes, we sought endometritis following delivery and infection of obstetrical surgical wound, infection following procedure, cellulitis, and/or peritoneal abscess. We included endometritis as an organ/space SSI by CDC definitions. Chart review was performed for all cases identified by any of the 3 methods to confirm SSI diagnosis as per CDC criteria. RESULTS: Among 3,507 cesareans, ICD-10 codes identified 76 confirmed cases of endometritis and 31 SSI. The positive predictive value for endometritis was 88.0% (95% CI 80.2-93.1) and for SSI codes, 53.5% (95% CI 43.0-63.6). Adding puerperal sepsis produced one additional SSI case, and searching the postoperative note yielded another 26 cases of endometritis and one wound infection. Six SSI were identified by infection control not identified by other methods. In total, we identified 141 SSI, including 98 endometritis-only cases. The incidence of all SSI, endometritis, and SSI excluding endometritis was 4.02%, 2.79%, and 1.23%. CONCLUSION: Improved ascertainment methods are needed to inform and assess performance improvement efforts for cesarean SSI. Reliance on diagnosis codes may miss a significant number of cases and is poorly predictive of SSI other than endometritis.