Aims: Internationally, research indicates that pharmacotherapy for chronic heart failure (CHF) is sub-optimal. Traditionally, assessment of drug use in heart failure has focused on the use of individual agents irrespective of CHF severity. This study investigates drug use for CHF patients in general practice with respect to the available evidence, incorporating both disease severity and the use of combination drug regimes. Methods and results: A cross-sectional survey of 769 Dutch CHF patients was performed as part of IMPROVEMENT of HF study. For each New York Heart Association severity classification the minimum treatment appropriate for the heart failure severity according to the scientific evidence available at the time of the study (1999) was defined. The proportion of patients treated with each drug increased with increasing severity, with the exception of the β-blockers. Patients with less severe heart failure were approximately four to eight times more likely to receive evidence-based treatment than those with more severe heart failure. Discussion: To assess pharmacological treatment of heart failure, in relation to the available evidence, it is important to take severity into account. While the number of drugs prescribed increased with increasing severity, the use of evidence-based regimes was lower in patients with more severe heart failure.