Evaluating a Prediction-Driven Targeting Strategy for Reducing the Transmission of Multidrug-Resistant Organisms

The transmission of multidrug-resistant organisms (MDROs) in the healthcare setting is an ongoing challenge affecting at least 2 million patients in the United States each year via infection and leading to over 20,000 deaths. Many mathematical models have been developed to approximate MDRO transmission dynamics, focusing most often on evaluating the impact of various infection-control strategies. However, although the insights derived from these studies are useful, the models do not typically have the ability to support decision making for infection-control practitioners in real time. In this study, we design a detailed agent-based model of MDRO transmission-focusing on methicillin-resistant Staphylococcus aureus in the intensive care unit setting-and validate its transmission dynamics using data collected during a multisite randomized, controlled trial. We leverage this model to develop and evaluate the effectiveness of a prediction-driven approach for targeting patients for contact precautions (i.e., requiring all visiting healthcare workers to wear personal protective equipment) in a simulated intensive care unit based on their daily likelihood of becoming colonized by the organism. We show that we can predict these outcomes with moderate to high accuracy across a broad range of scenarios and that these predictions can be used to efficiently target patients for intervention and, ultimately, to reduce the overall acquisition rate in the unit.