Prospective audit for antimicrobial stewardship in intensive care: impact on resistance and clinical outcomes.

BACKGROUND The impact of antimicrobial audit and feedback on outcomes of critically ill adults is unclear. METHODS A prospective study was performed in the intensive care units (ICU) of a public hospital in Atlanta, GA. Critically ill adults receiving empiric imipenem or piperacillin-tazobactam were eligible. Outcomes for 3 periods were compared: baseline (B, February to May 2006), model 1 (M1, October 2006 to July 2008), and model 2 (M2, September 2008 to February 2009). No audit was performed during B. During M1, an infectious diseases physician evaluated patients, and a critical care pharmacist communicated recommendations to the treating team. During M2, an infectious diseases physician directly participated in interdisciplinary rounds with the medical ICU team. RESULTS One hundred ninety-four patients were included during B, 415 during M1, and 83 during M2. M1 and M2 were associated with appropriate antimicrobial selection (B, 70%; M1, 78%; M2, 82%; P = .042) and with lower rates of resistance (B, 31%; M1, 25%; M2, 17%; P = .033). Logistic regression analysis confirmed that audit and feedback were independently associated with appropriate antimicrobial selection and prevention of resistance. The association remained strongest for M2. CONCLUSION Audit and feedback had an influence on antimicrobial prescription patterns in the ICU with a favorable impact on the emergence of resistance.