The authors assessed the long-term efficacy of a weight-loss program combining the use of a very low-calorie diet and behavior modification. This program involved medical supervision as well as group behavior modification classes, which emphasized long-term lifestyle changes, along with nutrition education. A study was conducted in 1984 and again in 1991, of unselected consecutive patients who had completed this hospital-based, weight-control program 12 to 18 months earlier. Patients were asked to return to the medical center to be weighed and complete a questionnaire regarding biobehavioral factors. The results of both the 1984 and 1991 follow-up studies were very similar. The patients' initial body mass index was 34.6 kg/m(2) +/- 0.57 kg/m(2) (mean +/- standard error of the mean, n = 109) and average weight loss was 23.2 kg +/- 1.1 kg, resulting in a body mass index of 26.2 kg/m(2) +/- 0.42 kg/m(2) upon completion of the program. At 12 to 18 months of follow-up, mean body mass index was 29.6 kg/m(2) +/- 0.54 kg/m(2). There was a wide spectrum of long-term compliance and recidivism. On average, 61% of patients kept off at least 50% of the weight they had originally lost. In addition, medication usage for hypertension, hyperlipidemia, and diabetes was reduced among the patients with reduced obesity. A retrospective analysis of patient characteristics and habits did not accurately predict which patients would be successful long-term. In conclusion, the combined use of a very low-calorie diet plus behavior modification was successful in the majority of patients who completed the program. Because there was a significant improvement in a number of medical problems, this approach continues to warrant consideration for the treatment of obesity.