OBJECTIVES: Clinical trials indicate that the development of heart failure (HF) impacts post-myocardial infarction (MI) prognosis. There is little information available on the economic burden of HF in post-MI patients, outside of a clinical trial setting. This study examines and compares the impact of developing HF on cost of care and length of stay (LOS) in patients hospitalized for MI using alternative models for estimation. METHODS: Claims data for patients hospitalized with a principal diagnosis of MI between 1998 and 2000 were used to compare cost of care and LOS between those with and without HF at initial hospitalization. Patients with a diagnosis of HF in the 6-months preceding the MI were excluded. Total hospitalization costs and LOS were analyzed for 15,160 patients using OLS, log-OLS and generalized linear models (GLM). Bootstrapping was used to obtain variances for cost estimates. Age, gender, type of MI, death, Medicare/non-Medicare, and comorbidities were included as covariates in the models. RESULTS: Comparing post-MI patients who develop HF during initial hospitalization (N = 2968) to those without HF at initial hospitalization (N = 12192) the mean incremental cost and incremental LOS were $7098 (SD = $434), and 3 days (SD = 0.13) respectively using the GLM. Based on several diagnostic tests, GLM was found to be the best estimator as long as the link function used is appropriate. Log-OLS estimates were biased both due to heteroscedasticity and the inappropriateness of log-transformation. OLS estimates were both biased and inefficient especially for costs. CONCLUSIONS: Patients who develop post-MI HF during initial hospitalization for MI have significantly higher hospitalization costs and increased LOS as compared to patients without post-MI HF. Treatments aimed at preventing or delaying the onset of post-MI HF could result in significant cost savings.