Abstract 16711: Predicting Progression of Heart Failure Using Administrative Data and Medical Records

Introduction: Heart failure (HF) affects nearly six million Americans. It is the leading cause of hospitalization in people older than age 65. About 50% of people who develop HF die within five years of diagnosis. Lifestyle change and proper medical treatment may help improve HF symptoms and prevent worsening of HF. Hypothesis: It is hypothesized that a predictive model can be developed to determine the risk of progression to a more severe HF stage in the next year for individuals enrolled in a Medicare Advantage plan. Methods: A claims-based approach was developed to approximate the HF stages (B/C/D) using AHA/ACA HF staging guidelines. The future HF stage in the next 12 months was used as the target of the predictive model. During the model development, a broad set of data sources were used, including administrative, medical, and pharmacy claims information, lab data, demographics, and health risk survey data. More than 3,000 potential risk factors from over 200,000 individuals with HF were extracted, i...