Random variability of blood pressure complicates the diagnosis and subsequent treatment of hypertension. To evaluate the importance of the number of blood pressure measurements in the correct diagnosis and control of hypertension, the authors used a Bayesian model to estimate the true average blood pressure of a group of newly diagnosed hypertensives, then calculated the diagnostic error that would result from monitoring methods using 24 daytime measurements or from using only three random monitoring measurements. The study population consisted of 129 individuals with newly diagnosed mild hypertension according to standard criteria, who were also evaluated with an ambulatory blood pressure monitor. In true normotensives (daytime diastolic blood pressure <90 mm Hg), the negative predictive value with three measurements was 0.92, and it rose to 0.96 with monitoring methods. In mild hypertensives (90-104 mm Hg), the positive predictive value was 0.64 with three measurements and 0.84 with monitoring methods, thus reducing the rate of false mild hypertensives from 35% to 15%. Finally, in patients with moderate or severe hypertension (>104 mm Hg), the positive predictive value improved from 0.26 with three readings to 0.61 with monitoring methods. Similar results were observed with daytime systolic pressure measurements. As the number of measurements increased, the diagnostic error due to the random variability of blood pressure became progressively smaller. In cases of hypertension, the large improvement in predictive values may justify using monitoring methods to confirm standard diagnosis.