We compared the reproducibility over time of blood pressure measured at the health examinations (screening blood pressure) and blood pressure measured at home (home blood pressure). Both screening and home blood pressure were measured in subjects of a rural community. Subjects measured their own blood pressure at home once in the morning using a semiautomatic oscillometric blood pressure measuring device at least three times (on at least 3 days) in each of two 4-week periods separated by one year. Similarly, two screening blood pressure measurements were obtained from the subjects at each of two health examinations also taken 1 year apart. A total of 136 untreated subjects without cardiovascular complications (40 men and 96 women, 56 +/- 11.7 years, mean +/- SD) were analyzed in the study. The correlations between the first and second blood pressure measurements of the subjects were significantly higher for the home blood pressure measurements (systolic: r = 0.844 and diastolic: r = 0.830) than for the screening blood pressure measurements (systolic: r = 0.692 and diastolic: r = 0.570). The mean differences between the first and second home blood pressure (0.8 +/- 7.7 mm Hg for systolic BP and 0.9 +/- 5.5 mm Hg for diastolic BP) were significantly smaller than those for the screening blood pressure (-3.9 +/- 13.8 for systolic BP and -3.1 +/- 10.2 for diastolic BP) (P < .001 for both comparisons), suggesting that the reproducibility of home blood pressure over time is superior to that of screening blood pressure. Such reliable blood pressure measurements obtained at home have a clinical significance for the diagnosis and treatment of hypertension and as a tool for evaluating the efficacy of antihypertensive drugs. Home blood pressure measurements also may be more useful than screening blood pressure measurements in predicting future cardiovascular events.