The aim of this study was to determine concordance between physician and patient blood pressure (BP) measurements in an ambulatory setting. A diagnostic intervention cross-sectional study using a convenience sample was employed. A total of 106 hypertensive patients were included in the study. Patients who were unable to perform their self-measurement or those with cardiac arrhythmia were excluded. BP was determined nine times in each subject in the medical office in a randomised order: BP was taken three times by the physician using a mercury sphygmomanometer (SPH-Hg), three times by the physician using a validated, automated oscillometer (Omron HEM 705 CP), and three times by the patient himself with the same device. The intraclass correlation coefficient was calculated. In all, 59 women and 47 men aged 65.7 (10) years were analysed. Mean BP measurements for the physician using the mercury sphygmomanometer, the physician using the Omron, and the patient using the same device were: 136 (15.8)/80 (11), 137 (17.9)/80 (10), and 139* (17.6)/80 (10) mmHg, respectively. BP control was 48.1, 48.1, and 36.8*% (*P < 0.05), respectively. Intraclass correlation coefficients for systolic/diastolic pressures were: 0.77/0.65 (physician-sphygmomanometer Hg, physician-Omron; P < 0.001), 0.75/0.64 (physician-sphygmomanometer Hg, patient-Omron, P < 0.001), and 0.83/0.83 (physician-Omron, patient-Omron; P < 0.001). In conclusion, the three types of measurement in the medical office were significantly concordant. Patient office self-measurement showed a tendency to increase systolic BP and worsen BP control.