Information on the long-term effects of non-restrictive antimicrobial stewardship (AMS) strategies is scarce. We assessed the effect of a stepwise, multimodal, non-restrictive AMS programme on broad-spectrum antibiotic use in the intensive care unit (ICU) over an 8-year period. Components of the AMS were progressively implemented. Appropriateness of antibiotic prescribing was also assessed by monthly point-prevalence surveys from 2013 onwards. A Poisson regression model was fitted to evaluate trends in the reduction of antibiotic use and in the appropriateness of their prescription. From 2011 to 2019, a total of 12,466 patients were admitted to the ICU. Antibiotic use fell from 185.4 to 141.9 DDD per 100 PD [absolute difference, -43.5 (23%), 95% CI -100.73 to 13.73; p = 0.13] and broad-spectrum antibiotic fell from 41.2 to 36.5 [absolute difference, -4.7 (11%), 95% CI -19.58 to 10.18; p = 0.5]. Appropriateness of antibiotic prescribing rose by 11% per year [IRR: 0.89, 95% CI 0.80 to 1.00; p = 0.048], while broad-spectrum antibiotic use showed a dual trend, rising by 22% until 2015 and then falling by 10% per year since 2016 [IRR: 0.90, 95% CI 0.81 to 0.99; p = 0.03]. This stepwise, multimodal, non-restrictive AMS achieved a sustained reduction in broad-spectrum antibiotic use in the ICU and significantly improved appropriateness of antibiotic prescribing.
Keywords: ICU; antimicrobial stewardship; appropriateness of antibiotic prescribing; broad-spectrum antibiotics; interrupted time series analysis.