24 hour blood pressure monitoring is a well established method in the field of antihypertensive research. Patients self recorded blood pressure values are an additional option to overcome the disadvantages of casual office readings--however they are not frequently used within intervention trials. To prove the usefulness of selfrecordings in clinical trials we investigated both selfrecordings taken twice a day and casual readings within intervals of 1 to 3 weeks, in this study on the efficacy and tolerability of the ACE-inhibitor Accupro. 108 hypertensive patients (grade WHO I to II) were included in this trial for ten weeks. Although blood pressures were measured by the patients using sphygmomanometers of the same type and the physicians, decisions to treat or to increase dosage were based on the patients' recordings only. Accupro was dispensed according to the package leaflet at a daily dosage of 5 mg up to 40 mg. In case of failing response to monotherapy, Accupro was combined with Diltiazem or with a diuretic. 7 patients discontinued the treatment due to mild adverse events, one did not cooperate. 82 of the remaining patients were treated effectively with Accupro monotherapy--60 (73%) got one dose daily, 22 (27%) 2 doses per day,--and in 18 patients a drug combination was required. Therapeutic response (RRd < or = 90 mm Hg) was gained within 86 of the 100 evaluable patients according to the doctors' and 83 according to the patients' records. In this respect the two methods used gave comparable overall results. This somewhat surprising fact is due to the design of the study, because treatment decisions were based on the selfrecordings only. Clinical trials based on selfrecordings are in some points preferable to casual office readings: As patients being normotensive at home should not be included into an interventional study, a change of dosage within this group is avoided. Additionally the compliance of a cooperative patient taking his blood pressure twice daily is at a high level. Measurements of each single patient may be evaluated statistically by time series-analysis regarding longterm distribution of blood pressure-values. Taking the means of selfrecordings over adequate time-intervals eliminates the influence of "outliers" (occasionally extremely high or low values) and also reduces the standard deviation compared to that of the casual readings. Research work based on self recordings provides more information and therefore more security for treatment decisions.