Although heart failure (HF) afflicts nearly 5 million Americans, the long-term cost of HF care has not been described previously. In a prospective, longitudinal cohort of community-dwelling elderly from 4 regions, we examined the long-term costs and resource use of elderly patients with HF.

We linked 4860 elderly participants in the National Heart, Lung, and Blood Institute Cardiovascular Health Study to Medicare part A and part B claims from 1992 to 2003. Costs were calculated from Medicare payments and discounted at 3% annually. We applied nonparametric estimators to calculate mean costs and resource use per patient for a 10-year period. To describe the relationship between patient characteristics and long-term costs, we constructed censoring-adjusted regression models.

There were 343 participants (84.8% white; 50.1% men; mean age, 78.2 years) with prevalent HF and 4517 participants without HF at study entry. Mean follow-up was 6.7 years (median, 6.4 years). The 10-year survival rates were 33% and 63% for the prevalent HF and nonprevalent HF groups (P < .001), respectively. The mean 10-year medical costs were significantly higher for the prevalent HF cohort ($54,704 vs $41,780, P < .001). The higher costs associated with HF were also reflected in greater resource use with more hospitalizations (P < .05) and more intensive care unit days (P < .05). Participants with HF had more physician visits (P < .05), with most of these encounters involving noncardiology physicians. However, in multivariate models, prevalent HF was not an independent predictor of higher costs.

Patients with HF consume substantially more health care resources than their elderly peers, and these higher costs persist through 10 years of follow-up. Many of these costs may be related to other comorbid conditions.