Since recent insights indicate that most patients with heart failure and a normal cardiac rhythm (HFNR) should not be treated with digitalis, we investigated the treatment policy of general practitioners in this respect. Our descriptive study consisted of a questionnaire followed by a computer based review of patient records of 14 family practices cooperating in the Registration Network of Family Practice (RNFP) and 10 'external' practices. All GPs (n = 51) received the questionnaire on the prescription policy of the GP and cooperating specialists, and on their experience with adverse effects of digitalis therapy. After 2-4 months the GPs of the RNFP (n = 41) received a floppy disc with an algorithm selecting from their problem lists patients with a registered diagnosis of heart failure. The external GPs (n = 10) received an algorithm selecting patients treated with digitalis, since their problem lists were not completely up to date. In all, data of 63,500 patients were examined. On anonymised standard forms, generated by the practice computer, the GPs provided data on the exact diagnosis, medication and adverse effects. Outcome measures of the study were: prescription attitudes (questionnaire) and actual digitalis use (in the selected cases). 33% of the GPs stated that they used digitalis as the preferred therapy of heart failure with normal cardiac rhythm; 30% of the 82 selected HFNR patients had digitalis, while in 42% of the 149 selected patients treated with digitalis a HFNR was registered. Adverse effects had been observed in 3 cases, but could not be excluded retrospectively in 70 cases (49%). Many GPs have not yet adapted their prescription behaviour to recently published insights. Development of a 'digitalis standard in general practice' is recommended.