A diagnostic algorithm for the surveillance of deep surgical site infections after colorectal surgery

Abstract

Objective:Surveillance of surgical site infections (SSIs) is important for infection control and is usually performed through retrospective manual chart review. The aim of this study was to develop an algorithm for the surveillance of deep SSIs based on clinical variables to enhance efficiency of surveillance.Design:Retrospective cohort study (2012–2015).Setting:A Dutch teaching hospital.Participants:We included all consecutive patients who underwent colorectal surgery excluding those with contaminated wounds at the time of surgery. All patients were evaluated for deep SSIs through manual chart review, using the Centers for Disease Control and Prevention (CDC) criteria as the reference standard.Analysis:We used logistic regression modeling to identify predictors that contributed to the estimation of diagnostic probability. Bootstrapping was applied to increase generalizability, followed by assessment of statistical performance and clinical implications.Results:In total, 1,606 patients were included, of whom 129 (8.0%) acquired a deep SSI. The final model included postoperative length of stay, wound class, readmission, reoperation, and 30-day mortality. The model achieved 68.7% specificity and 98.5% sensitivity and an area under the receiver operator characteristic (ROC) curve (AUC) of 0.950 (95% CI, 0.932–0.969). Positive and negative predictive values were 21.5% and 99.8%, respectively. Applying the algorithm resulted in a 63.4% reduction in the number of records requiring full manual review (from 1,606 to 590).Conclusions:This 5-parameter model identified 98.5% of patients with a deep SSI. The model can be used to develop semiautomatic surveillance of deep SSIs after colorectal surgery, which may further improve efficiency and quality of SSI surveillance.

Semiautomated surveillance of deep surgical site infections after colorectal surgeries: A multicenter external validation of two surveillance algorithms

“…This 5-predictor model had a sensitivity of 98.5% and a workload reduction of 63.3%. 20 External validation or actual implementation studies of new methods for automated surveillance are scarce. 21,22 As reported by 2 systematic reviews, only 23% of the included studies used a separate validation cohort 23 and only 25% of automated surveillance were used in clinical routine.…”

Using artificial neural networks for pattern recognition of post-surgical infections

“…Previous studies has used algorithms for the surveillance of deep surgical site infections after colorectal surgery (MULDER et al, 2019) and for detecting infectious disease high-risk patients (HAYKIN 2004) For the development of this study, it was essential to understand which information is collected and held at the hospital as well as the method and time of collection. In addition, interpreting the data was essential to figure out what was actually relevant to include in the database.…”

Automated surveillance systems for healthcare-associated infections: results from a European survey and experiences from real-life utilization

“…The algorithms used in semi-automated surveillance can be classification algorithms or decision trees, comprising of a set of indicators derived from structured data from hospital information systems [18][19][20][21]. The selection of indicators that are incorporated in the algorithms are based on prior experience and clinical knowledge, statistical methods or machine learning techniques [18,19,[22][23][24]. For fully automated surveillance, algorithms perform HAI ascertainment without human interference.…”