Advancing Approaches to Identify Young Children at Risk for Post-Discharge Mortality - PROJECT SUMMARY/ABSTRACT This K23 Mentored Patient-Oriented Research Career Development Award application details a rigorous plan for training and research activities that will provide critical skills and experience I need to become an independent investigator with expertise in prognostic research for post-discharge mortality (PDM) among young children in sub-Saharan Africa. I am an Assistant Professor of Pediatrics and Emergency Medicine at Emory University School of Medicine, an attending physician at Children’s Healthcare of Atlanta, and a pediatric research scientist in the Child Health and Mortality Prevention Surveillance (CHAMPS) network. My K23 training goals are to develop 1) proficiency in advanced methods in prognostic research, risk assessment tool evaluation, and large data analysis for PDM clinical prediction modeling in sub-Saharan Africa, 2) expertise in biomarkers of endothelial dysfunction for PDM, and 3) advanced experience in collaborative global health research. To achieve these goals, I have constructed an expert, multidisciplinary team of mentors, advisors, and collaborators. My co-primary mentors are Claudia R. Morris, MD and Cynthia G. Whitney, MD, MPH. Dr. Morris is a Professor of Pediatrics and Emergency Medicine at Emory University School of Medicine and an expert in the role of arginine dysregulation in endothelial dysfunction. Dr. Whitney is a Professor of Global Health at Emory Rollins School of Public Health, the PI of CHAMPS, and an expert in childhood mortality surveillance in sub-Saharan Africa. My mentor Rishikesan Kamaleswaran, PhD, MSc is an Associate Professor of Biomedical Informatics at Emory and will mentor me through the development machine learning risk assessment tools for PDM. PDM rates among young children in sub-Saharan Africa are 3-18% within months of hospital discharge, far outpacing the 0.1% rate we described in US children. Despite this immense disparity, there are no implemented risk assessment tools to identify young children at risk for PDM. The studies proposed here are designed to test the hypotheses that a machine learning approach to identifying young children at risk for PDM in Tanzania and Liberia will have greater discriminatory value than logistic regression models, that our risk assessment tools developed in Tanzania and Liberia will accurately identify young children in Kenya at risk for PDM, and that a key biomarker of endothelial dysfunction (i.e., the global arginine bioavailability ratio) will be lower in children who die after hospital discharge than those who survive. I will use the many resources at Emory University School of Medicine, Emory Rollins School of Public Health, the CHAMPS network, and the Kenya Medical Research Institute to add these studies to the Bill & Melinda Gates Foundation supported CHAMPS platform with a focus on the identification of young children at risk for PDM. These studies, the proposed coursework, and didactics will give me the training, skills, and experience needed to develop into an independent investigator equipped to conduct future research to reduce PDM.
