Placebo-controlled randomized trials have demonstrated the efficacy of selected β blockers on outcomes in chronic heart failure (HF), but the relative effectiveness of different β blockers in usual clinical care is poorly understood. We compared 12-month risk of rehospitalization for HF associated with receipt of different β blockers in 7,883 adults hospitalized for HF within 2 large health plans between January 1, 2001 and December 31, 2002. Beta-blocker use was ascertained from electronic pharmacy databases and readmissions within 12 months were identified from hospital discharge databases. Extended Cox regression was used to examine the association between receipt of different β blockers and risk of readmission for HF after adjustment for potential confounders. During follow-up, there were 3,234 person-years of exposure to β blockers (39.3% atenolol, 42.0% metoprolol tartrate, 12.3% carvedilol, and 6.4% other). Crude 12-month rates of readmissions for HF were high overall (42.6 per 100 person-years). After adjustment for potential confounders, cumulative exposure to each β blocker, and propensity to receive carvedilol compared with atenolol, adjusted risks of readmission were not significantly different for metoprolol tartrate (adjusted hazard ratio 0.95, 95% confidence interval 0.85 to 1.05) or for carvedilol (adjusted hazard ratio 0.92, 95% confidence interval 0.74 to 1.14). In conclusion, in a contemporary cohort of high-risk patients hospitalized with HF, we found that adjusted risks of rehospitalization for HF within 12 months were not significantly different in patients receiving atenolol, shorter-acting metoprolol tartrate, or carvedilol.