Three years ago, an 18-month randomized controlled trial of chronic, nonspecific low back pain investigated the effectiveness of instructing subjects to restrict bending activities in the early morning, when the fluid content in the disc is increased. Pain days (as recorded by daily diaries) were reduced 23% in the treatment group, compared to a 2% reduction in the control (sham treatment) group. The purpose of the current follow-up study was to determine whether the results of that trial were maintained during the 3 years following completion of the trial. A questionnaire was mailed to the 60 subjects who completed the original trial. Fifty subjects completed the questionnaire for a response rate of 83%. Thirty-one subjects (62%) continued to restrict bending activities in the early morning, and experienced a further reduction of 10.1 pain days per month (51%) since the completion of the experiment. However, some of the subjects who did not continue to comply also improved. Several possible explanations for the improvement of noncompliant subjects are offered.