Inter-hospital comparison of rates of surgical site infection following caesarean section delivery: evaluation of a multicentre surveillance study

Background: Short postoperative stays following caesarean section delivery make it difficult to assess accurately the risk of surgical site infection (SSI). Methods of case-finding that minimize variation are required to support effective surveillance systems, especially where used for benchmarking. Aim: To evaluate the efficacy of case-finding methods for SSI following caesarean delivery and their utility in establishing benchmark rates of SSI. Methods: Hospitals conducted surveillance over one or two 13-week periods. Patients were reviewed during their inpatient stay, post partum by community midwives and via patient questionnaire at 30 days post delivery. To estimate the reliability of case-finding methods, case-note reviews were undertaken in a random sample of four hospitals. Findings: A total of 404 SSIs were detected in 4107 caesarean deliveries from 14 hospitals. The median time to SSI was 10 days, 66% were detected in-hospital or by community midwives, and an additional 34% were patient-reported. The rate of SSI was 9.8% but the proportion of patients followed up varied significantly between centres. The estimated sensitivity and specificity of case-finding was 91.4% [95% confidence interval (CI): 53.4-98.4] and 98.6% (95% CI: 98.4-98.8), the positive predictive value 91.0% (95% CI: 82.4-96.1) and negative predictive value 98.6% (95% CI: 93.9-99.5). Conclusions: Combined case ascertainment methods are a feasible way to achieve active post-discharge surveillance and had high negative and positive predictive values. Additional SSIs can be detected by patient questionnaires but rates of SSI were strongly influenced by variation in intensity of both healthcare worker- and patient-based case-finding. This factor must be taken into account when comparing or benchmarking rates of SSI. (c) 2013 The Healthcare Infection Society. Published by Elsevier Ltd. All rights reserved.