OBJECTIVES: The aim of this research was to develop a model that consistently predicts diffusion patterns for oral prescription drugs used to treat cardiovascular conditions. METHODS:  This study used 2007-2014 the MarketScan Commercial Claims and Encounters Database, which includes claims from privately insured individuals. Additional information related to characteristics of eight drugs with indications for either hyperlipidemia, diabetes mellitus, type 2, hypertension or DVT/pulmonary embolism treatment was collected from the U.S. Food and Drug Administration (FDA) and the Pharmaprojects Citeline Database. The analysis was restricted to new molecular entities that were launched during the study period of 2008-2013 and patients over the age of 18 who were prescribed at least one of these medications. Limited dependent variable regressions were used to model the diffusion of selected drugs with cardiovascular indications as measured by patient utilization over time. Drug-related covariates and patient-level demographic characteristics were used to estimate the number of patients who receive a drug. Covariates included measures related to the strength of clinical evidence and risk/benefit recommendation for a treatment, the number of competitors at launch, and how innovative a drug was at launch. RESULTS: Findings suggest that the number of patients with an indicated diagnosis is the most significant overall predictor of diffusion in a commercially insured population regardless of therapy. However, models predicting diffusion at different points in time show that some factors are more significant than others depending on whether diffusion is forecasted from launch or at six, 12, or 24 months. CONCLUSIONS: Demand-based forecasting of the expected market share for new products is a common internal practice within the pharmaceutical industry. However, little information is publically available about the models used to estimate diffusion. Understanding the rate at which drugs diffuse could help to inform healthcare budgets.