The impact of site-specific clearance on methicillin-resistant staphylococcus aureus decolonization

Methicillin-resistant Staphylococcus aureus (MRSA) colonizes multiple body sites, and carriage is an important risk factor for MRSA infection. Successful decolonization reduces disease incidence; however, decolonization protocols vary in the number of body sites targeted, and the impact of site-specific treatments is not well understood. Here, we used data from a randomized controlled trial (RCT) of MRSA decolonization using chlorhexidine body and mouth wash and nasal mupirocin to quantify the contribution of each treatment component to the success of the protocol. We estimated mouthwash as the least effective treatment component and the combined effect of MRSA clearance at the nares, skin, and wound as 93% (90% credible interval 85%-99%) of the full decolonization. Our model can estimate the effectiveness of hypothetical treatments in silico and shows enhancing MRSA clearance at nares will achieve the largest gains. This study demonstrates the use of machine learning to go beyond what is typically achieved by RCTs, facilitating evidence-based decision-making to streamline clinical protocols.