Forty-five American and English doctors were surveyed to determine whether differences in their beliefs about the incidence and morbidity of complications and the role of testing for patients with chronic hypertension could be contributing to the large cross-national differences previously demonstrated in ambulatory test use for such patients. For each of nine tests, the number the English doctors thought they "should order" was significantly less than that of the American group. English doctors also estimated a lower incidence for each of seven possible complications of hypertension, but there was no difference in perceived morbidity of the complications. Both groups thought results from testing would alter therapy for only a small proportion of patients; however, the English estimate was significantly smaller than the American (16 percent versus 27 percent: p less than 0.05). The reasons given for testing were very similar in the two countries except that American doctors ranked as more important the reason that patients used the quantity of test ordering as an indicator of quality care. These results suggest that differences in ambulatory test use are consistent with physicians' beliefs about the number of tests they should be ordering, disease incidence, the likelihood that testing will affect patient care, and patient expectations. Further study should be directed toward understanding the contributions that differences in physician beliefs about the natural histories of diseases and patient expectations make to variations in test ordering.