Precision Medicine for Glucocorticoids in Sepsis
This project involves the application of precision medicine approaches (unsupervised clustering, supervised learning) to datasets of four already completed trials of glucocorticoids in sepsis (a severe form of infection). Trials of glucocorticoids in sepsis have been conflicting, some showing benefit and others showing no effect. We are trying to identify the patients that may benefit from glucocorticoids and those that may be harmed. This project also includes access to observational electronic health record datasets and COVID-19 datasets, causal inference approaches, and Bayesian methodologies. This project is R01 supported.
Precision Medicine for Nutrition in EDEN
This project is similar to Project #1 but focuses on incorporation of biomarker data into precision medicine approaches. This project focuses on a single trial of nutrition in acute respiratory distress syndrome - the EDEN trial - in which 1000 patients were randomized to receive either a low or high level of nutrition. The overall trial was negative. Biomarker specimens are available for ~900 patients, both prior to and following randomization. In our initial studies, we are using unsupervised clustering to identify hypoinflammatory and hyperinflammatory subphenotype membership using clinical and protein cytokine biomarker data. In supervised analyses, we are incorporating additional data on endocrine hormone biomarkers (with a particular focus on the incretin hormones glucagon-like peptide-1 and glucose-dependent insulinotropic peptide) to determine biomarker-informed individualized treatment effects. This project is R21 supported.