What is the quality of reporting of studies of interventions to increase compliance with antibiotic prophylaxis?

Background Despite studies reporting successful interventions to increase antibiotic prophylaxis compliance, surgical site infections remain a significant problem. The reasons for this lack of improvement are unknown. This review evaluates the internal and external validity of quality improvement studies of interventions to increase surgical antibiotic prophylaxis compliance. Study Design Three investigators independently performed systematic literature searches and selected eligible studies that evaluated interventions to improve perioperative antibiotic prophylaxis timing, type, and/or discontinuation. Studies published before the Surgical Infection Prevention project inception in 2002 were excluded. Each study was assessed based on modified criteria for evaluating quality improvement studies (Standards for Quality Improvement Reporting Excellence) and for facilitating implementation of evidence into practice (Reach-Efficacy-Adoption-Implementation-Maintenance). Results Forty-six articles met inclusion criteria; 93% reported improvement in antibiotic prophylaxis compliance. Surgical site infections were evaluated in 50% of studies and 65% reported an improvement. Less than 5% of studies used randomization, allocation concealment, or blinding. Nine percent of studies described efforts to minimize bias in the design results and analysis and 13% described a sample size calculation. Approximately one-third of studies described participant adoption of the intervention (26%), factors affecting generalizability (33%), or implementation barriers (37%). Most studies (80%) used multiple interventions; no single intervention was associated with change in compliance. Studies with the lowest baseline compliance showed the greatest improvement, regardless of the intervention(s). Conclusions The methodology and reporting of quality improvement studies on perioperative antibiotic prophylaxis is suboptimal, and factors that would improve generalizability of successful intervention implementation are infrequently reported. Clinicians should use caution in applying the results of these studies to their general practice.