OBJECTIVES: Previous studies have used Markov models to predict future CHF hospitalizations based on a patient's multiple prior admissions. Our objective is to develop a model that predicts future admissions based on left ventricular ejection fraction (LVEF) and the occurrence of a CHF admission in the preceding quarter. Another objective is to use this model to examine, in a managed care setting, the cost-benefit of the following CHF management strategies: ACE inhibitor therapy, beta blocker therapy, CHF disease management, and intensive case management. METHODS: We constructed a Markov model to project the medical and non-medical costs for a cohort of CHF patients in a commercially insured population. The model defines four disease states based on LVEF status (normal vs. low) and a prior admission (yes vs. no) in the preceding quarter. The probability of a future admission depends on the patient's disease state as well as age and gender. Plan-specific CHF admission probabilities were calibrated using claims analyses. Each intervention is assumed to change health service utilization patterns and improve survival. Cost-benefit is expressed as net savings in medical and non-medical costs. RESULTS: For a population of 100,000 commercially insured individuals, CHF case rate and medical costs are predicted to increase steadily over the next 5 years. Intensive case management appears to be the most cost-beneficial intervention, with an estimated per member per month (PMPM) net savings of $0.11 in the first year. Beta blocker therapy is expected to save $0.08 in the first year, while ACE inhibitors and CHF disease management show net positive savings starting year 2. All interventions remain cost-beneficial when non-medical savings are eliminated from the analyses. CONCLUSIONS: The four CHF management strategies modeled are not only effective in improving clinical outcomes; they are also predicted to be cost-beneficial in a managed care setting.