Impact of hospital participation in national surveillance in prevention of surgical site infection, England

Background: Embedding surveillance into clinical practice is consistently recognised in national and international guidelines as being key to prevention of surgical site infection (SSI). Given the resource needed to undertake high-quality SSI surveillance, assessing the effectiveness of participation in reducing rates of infection is important. Analyses were undertaken to assess effectiveness using the UK Health Security Agency (UKHSA) SSI Surveillance Service, one of the largest such surveillance databases in Europe. Methods: Surveillance data from the start of the UKHSA SSI surveillance programme in 1997 up to 2019 submitted by NHS and private hospitals were extracted for analysis. Data from 2020 onwards were not analysed to remove any influence of the COVID-19 pandemic. Hospitals followed standardised methodology with prospective patient follow-up for 30 days after surgery, or up to a year where prosthetic implants were used. SSIs were clinically defined based on CDC criteria and detected during the inpatient stay only up until 2008. For categories introduced after 2008 (cardiac, breast, cranial, spinal), SSIs detected during readmission to hospital were also included. Data for categories of surgery where 10 or more hospitals participated for more than two years were analysed to assess within-hospital trends according to the time since each hospital started surveillance. Results: A total of 2,280,166 procedures were submitted by 422 hospitals between 1997 and 2019. Of these 2,246,319 procedures were submitted by hospitals undertaking more than 2 years of surveillance in a given category of surgery. Thirteen categories of surgery under surveillance met the selection criteria (>10 hospitals submitting data for >2 years). Of these, all 13 saw substantial reductions in SSI risk since inception of surveillance. The greatest reductions in SSI risk over the first 4 years of surveillance were seen for knee replacement (risk ratio=0.34), limb amputation (0.41), hip replacement (0.43) and vascular (0.45) surgery, all of whom saw SSI risk more than halve over this period. Applying baseline SSI risks observed for each category to the number of procedures under surveillance in the subsequent 3 years (524,423 operations), 2887 fewer SSIs were observed in participating hospital across all categories. Discussion: We identified substantial reductions in SSI risk in hospitals participating in surveillance spanning orthopaedic, cardiothoracic, general, colorectal, and gynaecological surgery, with close to 3000 fewer infections arising in the 3 years subsequent to implementation of surveillance than would have been estimated from baseline infection risk. Conclusion: Our study supports the effectiveness of participation in national SSI surveillance schemes in reducing SSI risk. Through application of standardised methodology, combined with centralised alerting and support of hospitals identified as outliers, hospitals have succeeded in using surveillance to improve patient safety.