Inference of heterogeneous transmission of antimicrobial resistant pathogens in health care settings

PROJECT SUMMARY/ABSTRACT Antimicrobial resistance (AMR) was associated with over 4.95 million deaths globally in 2019 and is projected to cause over 10 million deaths annually by 2050. Healthcare systems are the critical settings where novel AMR organisms (AMROs) may emerge and spread to the broader community. Strategies for cost-effective AMRO containment in healthcare settings should target locations and individuals that contribute most to AMRO transmission; however, effective inference methods to accurately identify these targets are currently lacking. The objective of this project is to develop novel inference systems to quantify heterogeneous transmission rates in hospital wards and identify individuals carrying AMROs for six high-priority organisms in four major hospitals in New York City. We will use electronic health records, hospitalization information, laboratory test results, and genetic sequence data to pursue two specific aims: 1) infer location-specific transmission rates for multiple AMROs in hospital wards; 2) estimate individual-level AMRO carriage probabilities using multi-type observations. The project will leverage an innovative combination of advanced modeling techniques, new inference methodology, and unique datasets on multiple co-circulating AMROs. The specific aims build on our prior research on modeling and inference of AMROs and abundant preliminary analyses. The proposed research is significant because it addresses a pressing need in global public health – the emergence and spread of AMR pathogens in healthcare settings. The expected outcome of this project will produce novel inference methods for identifying hospital wards and individuals driving the spread of multiple AMROs in healthcare systems. As the data sources used in the studies are widely available in electronic health records, the inference system can be generalized for use in other hospital systems.