Comparison of nurse- and physician-determined clinic blood pressure levels in patients referred to a hypertension clinic

BACKGROUND: When measuring BP, the physician induces a transient pressor response triggered by an alarm reaction. This 'white-coat effect' can influence therapeutic decisions. Whether it depends on the characteristics of the physician has not been evaluated. OBJECTIVE: To assess the 'white-coat effect' induced by several physicians in a large sample of patients, using the blood pressure measured by trained nurses as a reference. SETTING: Referral hypertension clinic. METHODS: Patients were selected for the study if they had been referred for the first time to the clinic and if they had had their supine systolic/diastolic blood pressure measured by a trained nurse (mean of the last two of three measurements taken every 1 min by an oscillometric device) and a physician (auscultatory method using a standard mercury sphygmomanometer). Physicians were included in the study provided they had seen at least 25 patients during the study period. The between-physician difference was assessed using linear regression analysis. Physician blood pressure was the dependent and nurse blood pressure was the independent variable. RESULTS: From 1 January 1997 to 15 September 1997, 1062 patients (50% male, aged 52 +/- 14 years), seen by 10 physicians (26-187 patients per physician) and one nurse were included for analysis. The mean systolic/diastolic blood pressure for physicians was 162 +/- 27/ 97 +/- 15 mmHg and that for the nurse was 155 +/- 24/ 88 +/- 14 mmHg. The nurse-physician differences were -6 mmHg (range -67 to +66) for systolic and -8 mmHg (-44 to +31) for diastolic blood pressures. Major differences were observed between individual physicians. Intercepts of the physician blood pressure versus nurse blood pressure relationship ranged from 0.1 -60.7 mmHg for systolic and from 13.3-55.3 mmHg for diastolic pressures. The slopes of this relationship differed less between physicians for systolic (0.72-1) than for diastolic pressures (0.56-0.97). There was no difference between the patients seen by physicians in patients' age, sex, tobacco consumption, anti-hypertensive treatment or target-organ damage. CONCLUSION: Large between-physician differences exist in the magnitude of the white-coat effect that cannot be explained by patient characteristics. Physicians should therefore not make any decisions based on blood pressure measured manually during a first encounter.