Risk factors for outcomes in heart failure (HF) were derived from populations in clinical trials, at hospital discharge, or in localized geographic or socioeconomic strata before the widespread use of beta blockers. This study observed 4,280 patients in a community-based HF registry for 1 year after completing carvedilol titration. Independent risk factors for death, hospitalization for HF, or hospitalization for cardiovascular reasons other than HF were first identified by age-, gender-, and race-adjusted analyses, then by multivariate analysis adjusted simultaneously for all factors. Over this period, 7% of patients died, 11% were hospitalized for HF, 12% were hospitalized for other cardiovascular reasons, and 27% had > or =1 of these events. The most significant outcome predictors were New York Heart Association class III or IV, history of hospitalization for HF or other cardiovascular reasons, and angina pectoris, all associated with increased odds of having an adverse outcome (all p < or =0.001). The left ventricular ejection fraction was not a significant outcome predictor by multivariate analysis. The odds ratio for an adverse outcome was significantly reduced for patients with hypertensive or idiopathic causes of HF and for those whose physicians had graduated from medical school > or =24 years earlier compared with <14 years earlier (all p <0.005). In conclusion, easily obtained historical information predicts clinical outcomes in patients with HF in the year after initiating carvedilol. In this unselected community population, these historical factors were better predictors of risk than the left ventricular ejection fraction.