To assess the extent of three year follow up of blood pressure, weight, and smoking habit in patients with raised blood pressure identified at health checks. Retrospective audit of medical and nursing records. Three general practices in Oxfordshire. 386 of 448 patients with raised blood pressure (diastolic greater than or equal to 90 or systolic greater than or equal to 160 mm Hg) identified from 2935 patients aged 35-64 attending health checks in 1982-4. All records of blood pressure, weight, and smoking habit in the medical record were abstracted for three years after the initial health check. All 42 patients with an initial diastolic blood pressure greater than or equal to 105 mm Hg and 316 of 344 patients with an initial pressure of 90-104 mm Hg had at least one further measurement of their blood pressure. Follow up of smoking habit and of weight was less complete with only half of the 100 smokers and 67 of the 87 obese patients (body mass index greater than or equal to 30) having any documented follow up of these risk factors. Annual follow up in the second and third years occurred in 228/297 (76.8%) and 232/320 (72.5%) in patients with blood pressure greater than 95 mm Hg at the beginning of each year. For patients who smoked annual follow up in these years occurred in fewer than a third and for those who were obese in just over half. On the assumption that those not followed up had not changed, at the end of three years the proportion of patients with diastolic blood pressure greater than or equal to 100 mm Hg had fallen from 61 patients (15.8%) to 31 (8.1%); the proportion of smokers had fallen from 103 (26.7%) to 94 (24.4%); and the proportion of obese patients had fallen from 87 (22.5%) to 79 (20.5%). These changes were modest and in the absence of a control group cannot be attributed necessarily to health checks. Although the standard of follow up was better than in previously reported studies of the management of hypertension, the results emphasise the need to develop formal protocols for dietary and antismoking interventions and to evaluate formally the effectiveness (and cost effectiveness) of health checks.