INTEGRATIVE DATA APPROACHES FOR RESISTANT HYPERTENSION IDENTIFICATION AND PREDICTION

项目总结

PROJECT SUMMARY This K01 proposal will facilitate my career development and advance my goal of becoming an independent investigator focused on discovery and prediction of factors associated with cardiovascular disease risk and drug response. My research will accomplish this through investigations that include biomedical “Big Data” from multiple sources, such as electronic health record (EHR) based data, claims based data, and genomics and other `omics data. The objective for this application is to utilize large datasets to identify characteristics predictive of resistant hypertension (RHTN). RHTN describes a subset of hypertensive (HTN) individuals with elevated blood pressure (BP) despite use of multiple anti-HTN medications. Based on current estimates of the prevalence of RHTN among HTN adults, over 12 million Americans could have RHTN. While these individuals' BP remains uncontrolled, they are at a 27% increased risk for adverse cardiovascular outcomes. The central hypothesis is that variance in the prevalence of RHTN can be explained by clinical factors, biochemical factors, `omic factors, and medication adherence. To test the central hypothesis, I will complete the following Specific Aims: 1) Validate the RHTN computable phenotype within OneFlorida through manual EHR chart review, 2) Identify characteristics and predictors of RHTN in the real-world population within EHR based data, 3) Estimate the level of anti-HTN adherence within a real-world RHTN population, and 4) Quantify the variability that estimated anti-HTN medication adherence explains in predicting RHTN. In order to build on my strong expertise and background in human genetics and pharmacogenomics, I will also conduct an Exploratory Aim: Integrate `omics data with EHR based data to characterize `omic signatures of adverse HTN outcomes. I will utilize data from OneFlorida and ADVANCE, two of the Clinical Data Research Networks within the National Patient Centered Clinical Research Network or PCORnet, giving me access to longitudinal EHR-based data on up to ~14 million individuals. The proposed study is significant because it will identify clinical, biochemical, `omic, and adherence characteristics associated with RHTN, allowing HTN patients with a higher risk for RHTN or non-adherence to be identified sooner, and targeted to precision treatment regimens. To successfully conduct this work, I requires specific training in 1) the validation of computable phenotypes, 2) the refinement of prediction models using large datasets, 3) the complexities associated with integration of data from EHR and claims based sources, 4) the complexities associated with integration of data form EHR and `omics based sources, and 5) clinical decision support. This training plan was designed with my strong mentoring team (William Hogan, MD, MS; Rhonda Cooper-DeHoff, PharmD, MS, George Michailidis, PhD; Dana Crawford, PhD, and Francois Modave, PhD). Finally, the rich training environment at the University of Florida, coupled with my previous training and experience, innovative research plan, high-quality training plan, and outstanding mentoring team give me the highest likelihood of successful transition to research independence.