Estimating the Per-Application Effectiveness of Chlorhexidine Gluconate and Mupirocin in Methicillin-resistant Staphylococcus aureus Decolonization in Intensive Care Units

Introduction: Chlorhexidine gluconate and mupirocin are widely used to decolonize patientswith methicillin-resistant Staphylococcus aureus (MRSA) and reduce risks of infection in hospitalized populations. The probability that a treated patient would be decolonized, which we term per-application effectiveness, is difficult to directly measure. Quantifying the efficacy of per-application effectiveness of CHG and mupirocin is important for studies evaluating alternative decolonization strategies or schedules as well as identifying whether there is room for improved decolonizing agents. Methods: Using a stochastic compartmental model of an intensive care unit (ICU), the per-application effectiveness of chlorhexidine and mupirocin were estimated using approximate Bayesian computation. Extended sensitivity analysis examined the potential impact of a latent period between MRSA colonization and detection, the timing of decolonization administration, and parameter uncertainty. Results: The estimated per-application effectiveness of chlorhexidine was 0.15 (95% Credible Interval: 0.01, 0.42), while the estimated effectiveness of mupirocin was is 0.15 (95% CI: 0.01, 0.54). A lag in colonization detection markedly reduced both estimates, which were particularly sensitive to the value to the modeled contact rate between nurses and patients. Gaps longer than 24-hours in the administration of decolonizing agents still resulted in substantial reduction of within-ICU MRSA transmission. Discussion: The per-application effectiveness estimates for chlorhexidine and mupirocin suggest there is room for substantial improvement in anti-MRSA disinfectants, either in the compounds themselves, or in their delivery mechanism. Despite these estimates, these agents are robust to delays in administration, which may help in alleviating concerns over patient comfort or toxicity.