OBJECTIVE: We evaluated three devices for self-measurement of blood pressure - the Omron HEM-705CP, the Philips HP5332 and the Nissei DS-175 - according to the revised protocol of the British Hypertension Society (BHS). The results were also analysed according to the criteria for accuracy of the revised standard of the Association for the Advancement of Medical Instrumentation (AAMI). DESIGN: The revised BHS protocol is divided into two parts. Part I, the part applicable to this study, comprises the main validation procedure and has five phases: Before-use device calibration; in-use (field) phase; after-use device calibration; static device validation; report of evaluation. METHODS: Three models of each device passed the before-use device calibration test, after which they entered the in-use phase, which involved use of the three recorders for a month; inter-device calibration was assessed again at the end of the month. There was no difference in calibration testing between the three models of each device, and therefore one of each was selected randomly; the main validation test was carried out in 85 subjects with a wide range of pressures, and the results were analysed according to the BHS grading system from A to D. RESULTS: The Omron HEM-705CP achieved an overall B/A grading and fulfilled the AAMI accuracy criteria; the Philips HP5332 achieved an overall C/A grading and failed the AAMI accuracy criteria for measuring systolic pressure; the Nissei DS-175 achieved an overall D/A grading and failed the AAMI accuracy criteria for measuring systolic pressure. When the BHS and AAMI criteria were applied to tertiles of pressure (low-pressure range < 130/80 mmHg; medium-pressure range 130-160/80-100 mmHg; high-pressure range > 160/100 mmHg) all three devices were less accurate in the high-pressure range: the Omron HEM-705CP achieved C/B grading while continuing to fulfil the AAMI criteria; the Philips HP5332 dropped to D grading for systolic pressure and the Nissei DS-175 achieved a lower D grading for systolic pressure. The mean and standard deviation of the first mercury sphygmomanometer measurements were 148+/-35/88+/-22 mmHg. Acceptability by the users was good and the manufacturer's manual was satisfactory for all three devices. CONCLUSIONS: On the basis of these results, the Omron HEM-705CP was the most accurate of the three devices tested, achieving Grade B for systolic and Grade A for diastolic pressure, as well as fulfilling the AAMI criteria for accuracy for both systolic and diastolic pressure. It can therefore be recommended for the clinical measurement of blood pressure and is the first inexpensive device to satisfy the accuracy criteria of these protocols.