Dr. Zhe He (School of Information), Dr. Michael Killian (College of Social Work), and their research team recently received a $395,200 R21 grant from the National Library of Medicine of National Institutes of Health for a project entitled “Prediction of Health Outcomes and Adverse Events in Pediatric Organ Transplantation in Florida”. This is a collaboration between FSU investigators and Dr. Dipankar Gupta (University of Florida), Dr. Paolo Rusconi (University of Miami), and Dr. Jennifer Garcia (University of Miami).
The goal of the project is to improve the prediction of patient and graft survival. “Our research will further our ability to predict post-transplant health outcomes using large datasets from pediatric heart, kidney, and liver transplant recipients in the two largest centers in Florida,” said Dr. He. “We aim to identify high-risk patients and to support clinical care and decision-making within pediatric organ transplant centers. We are excited about this funding from NIH, which will move our collaboration forward and advance the field of organ transplantations.”
The project began three years ago. In 2019, the research team received a 1-year Precision Health Pilot Grant from UF-FSU Clinical and Translational Science Award. With the support of the pilot grant, the team built machine learning models with the data from a pediatric organ transplant center in Texas to predict 1, 3, and 5 year hospitalizations of children who had kidney, liver, and heart transplants. The paper was published in JAMIA Open. With the support of this grant, the research team will be able to develop machine learning models to predict health outcomes and adverse events among children who had solid organ transplantations. They will use natural language processing techniques to extract social determinants of health and psychosocial factors from clinical notes and combine them with other structured data including demographic, familial, medical, health, and other posttransplant characteristics. Then they will apply machine learning and deep learning to predict posttransplant outcomes including organ rejection, graft loss, and patient survival, while ranking the risk factors that affect posttransplant outcomes in children.
The use of advanced machine learning modeling with multiple patient-level data sources represents a significant advancement over the extant research and an increase in the translational utility of patient data sources. Results will then be applicable to pediatric organ transplant centers nationally and the methodology germane to other areas of pediatric health.
“This project represents an important advancement in pediatric organ transplantation,” added Dr. Killian. “Data on the psychological and social determinants of the health and quality of life of these children are collected in large datasets or administrative record keeping. Instead, what we know about these children and their families is contained in the text notes, assessments, and evaluations completed by transplant social workers, psychologists, and other multidisciplinary transplant team members. By working with Dr. He and with funding from NIH, we can bring this valuable information about the patient and family forward into model development and prediction of health outcomes. It’s been my dream for years to enrich our statistical work using psychosocial information which plays such a large role in clinical decision-making.”
Future research will translate the derived models into interventions to enhance clinical care and transplant team decision-making within pediatric transplant centers.