Personalizing Cardiovascular Care for Veterans

Veterans Health Administration, Decatur PA

Investigators

Linked publications, trials & patents

Abstract

Personalizing Cardiovascular Care for Veterans (PI: Rodney A. Hayward, MD) Objectives: An overriding goal of medical care should be tailoring care to the individual person's risks, benefits and preferences. To accomplish this, it is first necessary to accurately and reliably estimate an individual patient's risks of morbidity and mortality in the absence of treatment. For cardio- and cerebrovascular (CCV) disease, the leading cause of morbidity and mortality in the US, several CCV risk prediction tools already exist, however, these tools have substantial shortcomings, including that they require providers to manually enter risk factor information, were developed on and calibrated to patient populations quite different from the patient population served by VHA, and do not utilize the full spectrum of clinical data available in VA's electronic medical record (VA EMR). Therefore, we propose a 1-year pilot project using 5- years of VA EMR data (2003 thru 2007) at 6 VA facilities to: 1. Construct a longitudinal EMR-derived dataset for initial exploration of the feasibility of developing an automated Coronary Artery Disease (CAD) risk prediction tool;and, 2. Conduct preliminary analyses to develop a cardiovascular mortality risk prediction model using Weibull accelerated failure regression. It is hoped that this feasibility project will be the first step in a process directed at developing automated tools that can be integrated into the VA EMR or a web-based interface (such as MyHealtheVet) to aid clinicians to optimize and personalize treatment decisions in the outpatient setting. If this pilot study is successful, an IIR will be submitted in the coming year. Research Plan/Methods: For Specific Aim 1, detailed information on all subjects meeting VA's EPRP "established patient" criteria will be included from 6 facilities, including data from hospitalization, outpatient visit, laboratory, pharmacy, vitals and mortality files. Data will be obtained from files maintained at the Austin Information and Technology Center (AITC) as well as from the Corporate Data Warehouse (CDW). Specifically, we will extract laboratory and pharmacy data from the DSS National Data Extracts and information about outpatient visits, inpatient use, diagnoses (ICD-9 codes) and procedures (CPT codes) from the SAS Medical Datasets. For Specific Aim #2, we will obtain cause of death for 1000 subjects from the National Death Index (NDI). Preliminary model development will be conducted on a stratified random sample of 5000 men 40 years of age or older at the 6 facilities: a nested case-control sample that will include 4000 patients with no death during the study period and 1000 patients who died during year's 2-5 of follow-up (death in year 1 is an exclusion criterion). Preliminary CCV prediction models for Specific Aim 2 will be developed using Weibull accelerated failure regression modeling.