Purpose: Audit and feedback can improve physicians' practice; however, the most effective type of feedback is unknown. Inadvertent perioperative hypothermia is associated with postoperative complications and remains common despite the use of effective and safe warming devices. This study aimed to measure the impact of targeted audit and feedback on anesthesiologists' intraoperative temperature management and subsequent patient outcomes.

Methods: This study was a three-arm cluster randomized-controlled trial. Anesthesiologists' intraoperative temperature management performance was analyzed in two phases. The first was a baseline phase with audit but no feedback for eight months, followed by an intervention phase over the next seven-month period after participants had received interventions according to their randomized group allocation of no feedback (control), benchmarked feedback, or ranked feedback. Anesthesiologists' percentage of hypothermic patients at the end of surgery (primary endpoint) and use of a warming device were compared among the groups.

Results: Forty-five attending anesthesiologists who took care of 7,846 patients over 15 months were included. The odds of hypothermia (temperature < 36°C at the end of surgery) increased significantly from pre- to post-intervention in the control and ranked groups (control odds ratio [OR], 1.27; 95% confidence interval [CI], 1.03 to 1.56; P = 0.02; ranked OR, 1.26; 95% CI, 1.01 to 1.56; P = 0.04) but not in the benchmarked group (OR, 1.05; 95% CI, 0.87 to 1.28; P = 0.58). Between-arm differences in pre- to post-intervention changes were not significant (benchmark vs control OR, 0.83; 95% CI, 0.62 to 1.10; P = 0.19; ranked vs control OR, 0.99; 95% CI, 0.73 to 1.33, P = 0.94). No significant overall effect on intraoperative warmer use change was detected.

Conclusion: We found no evidence to suggest that audit and feedback, using benchmarked or ranked feedback, is more effective than no feedback at all to change anesthesiologists' intraoperative temperature management performance. Feedback may need to be included in a bundle to produce its effect.

Trials registration: www.clinicaltrials.gov (NCT02414191). Registered 19 March 2015.