The aim of this study was to assess the long-term effect of a supervised fitness programme on patients with chronic low back pain. The design of the study was a single blind randomised controlled trial with follow-up, by postal questionnaire, 2 years after intervention. The Oswestry Low Back Pain Disability Index was used as the outcome measure to assess daily activity affected by back pain. Eighty-one patients with chronic low back pain, who were referred to the physiotherapy department of a National Health Service orthopaedic hospital, were randomised to either a supervised fitness programme or a control group. Patients in the intervention group and control group were taught specific exercises to be continued at home and referred to a backschool for back care education. In addition, the intervention group attended eight sessions of a supervised fitness programme. Sixty-two patients (76%) with a mean age of 37 years, returned the Oswestry Low Back Pain Disability Index questionnaire. Of these, 29 were in the intervention group and 31 in the control group. Patients in the intervention group demonstrated a mean reduction of 7.7% in the Oswestry Low Back Pain Disability Index score (95% confidence interval of mean paired difference 3.9, 11.6 P < 0.001), compared with only 2.4% in the control group (95% confidence interval of mean paired difference -2.0, 6.9 P > 0.05). Between group comparisons demonstrated a statistically significant difference in disability scores between the treatment and control group (mean difference 5.8, 95% confidence interval 0.3, 11.4 P < 0.04). This study supports the current trend towards a more active treatment approach to low back pain. We have demonstrated clinical effectiveness of a fitness programme 2 years after treatment but this needs to be replicated in a larger study which should include a cost effectiveness analysis, further analysis of objective functional status and a placebo intervention group.