Evaluating an evidence-based bundle for preventing surgical site infection: a randomized trial.

Objective: To determine if an evidence-based practice bundle would result in a significantly lower rate of surgical site infections (SSIs) when compared with standard practice.

Design: Single-institution, randomized controlled trial with blinded assessment of main outcome. The trial opened in April 2007 and was closed in January 2010.

Setting: Veterans Administration teaching hospital.

Patients: Patients who required elective transabdominal colorectal surgery were eligible. A total of 241 subjects were approached, 211 subjects were randomly allocated to 1 of 2 interventions, and 197 were included in an intention-to-treat analysis.

Interventions: Subjects received either a combination of 5 evidenced-based practices (extended arm) or were treated according to our current practice (standard arm). The interventions in the extended arm included (1) omission of mechanical bowel preparation; (2) preoperative and intraoperative warming; (3) supplemental oxygen during and immediately after surgery; (4) intraoperative intravenous fluid restriction; and (5) use of a surgical wound protector.

Main Outcome Measure: Overall SSI rate at 30 days assessed by blinded infection control coordinators using standardized definitions.

Results: The overall rate of SSI was 45% in the extended arm of the study and 24% in the standard arm (P = .003). Most of the increased number of infections in the extended arm were superficial incisional SSIs (36% extended arm vs 19% standard arm; P = .004). Multivariate analysis suggested that allocation to the extended arm of the trial conferred a 2.49-fold risk (95% confidence interval, 1.36-4.56; P = .003) independent of other factors traditionally associated with SSI.

Conclusions: An evidence-based intervention bundle did not reduce SSIs. The bundling of interventions, even when the constituent interventions have been individually tested, does not have a predictable effect on outcome. Formal testing of bundled approaches should occur prior to implementation.