BACKGROUND Blood pressure (BP) control rates in the United States have not improved significantly during the past decade. There has been limited study of improvement efforts focusing on guideline implementation and changes in the model of care to address hypertension. METHODS Five physician (MD)/registered nurse (RN)/licensed practical nurse (LPN) teams in a large community practice modified their care model in 1997 to manage hypertensive patients as part of guideline implementation efforts. The other 25 MD teams in the same setting practiced in the usual model, but were exposed to the guideline recommendations. BP control rates of patients in each group were assessed monthly. After nine months of testing the new care model, 10 additional teams adopted the model. RESULTS In the pilot group, hypertension control rates showed statistically significant improvement from pre- (33.1%) to postimplementation (49.7%). After adjusting for age, this was significantly greater than the improvement in the control group (p = 0.033). Medication changes were more frequent in the pilot group (32.3%) than in the control group (27.6%); however, the differences were not statistically significant. A longitudinal examination of the hypertension patients in the study showed that improved BP control was sustained for at least 12 months. CONCLUSIONS A change in the model of care for hypertensive patients within a primary care practice resulted in significant, sustainable improvement in BP control rates. These changes are consistent with the chronic care model developed by Wagner; practice redesign appeared to be the most important change.