Modeling and Simulation to Support Epidemiological Decision-Making in Healthcare Settings

点击翻译按钮获取中文摘要

Our proposal assembles an exceptional group of researchers to join the Modeling Infectious Diseases in Healthcare (MInD) network. The University of Utah will serve as the hub, with Harvard School of Public Health operating as a major node. We propose a program of research that will use models and data to make significant contributions to the body of evidence that informs prevention and control of healthcare-associated infections due to antibiotic-resistant pathogens. In the CDC’s 2019 Antibiotic Resistance Threat Report, it was estimated that the burden of resistant infections, including Clostridioides difficile, exceeded 3 million cases per year, with approximately 48,000 deaths. Two of the urgent threats listed in CDC’s Threat Report – carbapenem-resistant Enterobacteriaceae and carbapenem resistant Acinetobacter – and four of the serious threats – extended spectrum beta lactamase-producing Enterobacteriaceae, methicillin-resistant Staphylococcus aureus, vancomycin resistant enterococci, and multi-drug resistant Pseudomonas aeruginosa – are directly addressed in our proposal. Our proposal includes two tightly interwoven and complementary projects that tackle key controversies in infection control and antibiotic stewardship, highlighting the role that dynamic models can play in healthcare epidemiology. Project 1 uses the distinction between horizontal and vertical infection control strategies as a framework to systematically evaluate interventions to prevent transmission, ranging from the level of a hospital room to the region level, with multiple connected facilities. Project 1 powerfully leverages the model, data, and results generated by the CDC-funded "Granular Modeling – Simulating the Transmission of Healthcare- Associated Infections in Hospitals and Control Strategies" project. For Aim 1.1, we will use our agent based model, which has detailed representation of transmission components, to simulate trials of horizontal interventions. For Aim 1.2 we will use a blend of mathematical theory and data analysis to assess the facility- level impact of vertical control strategies across a range of surveillance and cohorting scenarios. For Aim 1.3 we will use a multi-facility framework to model the effects of combinations of horizontal and vertical interventions across multiple pathogens and perform cost-effectiveness analyses that account for indirect population-level effects due to transmission reduction. Project 2 leverages the immense data resources associated with the VA health system’s electronic health records to examine the effect of antibiotic selection pressure on antibiotic resistance profiles for a broad set of high priority pathogens. We will test hypotheses derived from evolutionary models about the impact of co-selection and about the temporal relationships between changes in antibiotic use and antibiotic resistance. We will also apply multivariate time series methods and empirical dynamic modeling to forecast and explain trends in resistance. Our goal is to improve understanding of why some forms of bacterial resistance are decreasing, others are increasing, and others are at relative equilibrium.