This prospective study was initiated 3 years ago to evaluate the outcome and to identify predictors of success or failure in patients admitted to a rehabilitation program for chronic low-back pain. Multiple parameters were evaluated, including psychologic data (MMPI, personal interview, pain drawing, etc.), physical measurements (flexibility, strength and endurance), and demographic data concerning the patient's home and working environment. Information was available on each patient admitted to the program prior to his admission, at completion of the program, 6 weeks following completion of the program and 3 months following completion of the program. A telephone interview was carried out 2 1/2 years following the patient's discharge from the program. Linear regression analysis was used to identify the important independent variables with regard to the dependent variables of relief of back pain, return to work and increased activities at home. Demographic data were of no value as a predictor with the exception of age and returning to work. The patients over the age of 50 returned to work with much less frequency than those less than 50. Psychologic information from the MMPI and similar tests were of no value. The personal preadmission interview of a trained psychologist, however, was a good predictor of an individual's eventual return to work and overall improvement. Worker's Compensation and other litigation was a negative factor in a patient's prognosis. The treatment team's prognosis at the time of discharge from the program was the best overall predictor of a patient's chance of success or failure in the longterm.