An Offer You Can’t Refuse: Clinical Impact of Accepting or Rejecting a Recommendation from an Antibiotic Stewardship Program

Abstract Background The outcomes associated with the acceptance or refusal of a recommendation from an antimicrobial stewardship program (ASP) on an individual level have not been studied yet. Our objective was to compare the clinical characteristics and mortality of patients for whom a recommendation from an ASP, based on prospective audit and feedback and triggered by a computerized decision support system, was accepted or refused. Methods We performed a retrospective cohort study of all hospitalized adult patients who received intravenous or oral antimicrobials in two tertiary care hospitals in Canada between 2014 and 2016, and for whom a recommendation was issued by an ASP. Results We identified 1,251 recommendations throughout the study period. Among the recommendations made by the pharmacist to prescribers, 1,144 (91.5%) were accepted. The most frequent interventions were immediate scheduling end of treatment (n = 364, 29%), dosing/frequency adjustments (n = 321, 26%), streamlining (n = 251, 20%), and switching from intravenous to oral therapy (n = 247, 20%). The antimicrobials most frequently targeted by recommendations were piperacillin/tazobactam (n = 273, 22%) and fluroquinolones (n = 267, 21). Overall, the length of the antimicrobial targeted by the recommendation was significantly shorter when a recommendation was accepted (0.37 days vs. 2.11 days; P < .001). In the multiple logistic regression analysis, the independent risk factors associated with in-hospital mortality were the Charlson score, issuance of a recommendation for a patient in the intensive care unit, the duration between admission and the recommendation, issuance of a recommendation in 2016 (compared with 2014), and age of the patient. A recommendation issued on a fluoroquinolone or oral penicillin/first generation cephalosporin was associated with lower odds of mortality. After adjustment, refusal of a recommendation by the attending physician was associated with a higher, albeit nonsignificant, risk of mortality (AOR, 1.81; 95% CI, 0.89–3.68; P = .10). Conclusion The duration of the antimicrobial treatment was significantly shorter when a recommendation triggered by an ASP program was accepted. This decrease in antimicrobial duration was not associated with increased mortality. Disclosures J. Perron, Lumed Inc., the company that commercializes APSS: Shareholder, Shareholder; V. Nault, Lumed Inc., the company that commercializes APSS: Shareholder, Shareholder; M. Beaudoin, Lumed