An outpatient multimodal cognitive-behavioural treatment program (MMCBT) for chronic spinal pain was evaluated during an 18-month follow-up period. The treatment included a 1-day course for the patients' work supervisors. The aim of the study was to evaluate the long-term effect of the treatment program as well as the effect of a work supervisor-training program on the patients' return to work. The design was a matched cohort study with four repeated measures. Two groups of subjects suffering from non-specific spinal pain (treatment group n=67, no treatment control group n=29) were assessed with regard to pain intensity, sick-leave, pain coping ability and behavioural changes at work and in personal life. Recordings on supervisors attending the course and changes of supervisory behaviour were also obtained. The results show a significant between group difference in pain intensity and perceived pain coping ability at work, favouring the MMCBT group. Further, the MMCBT intervention enhances self-reported behavioural changes in areas within the subjects' own control. There is not sufficient statistical support to accept the assumption of MMCBT being superior in reducing sick-leave, either with or without the education of supervisors. Even when supervisors changed their behaviour as reported by the patient, no significant effect was found on patients' return to work. In conclusion, the MMCBT do not seem to be effective in reducing sick-leave compared to no treatment, but the MMCBT program is superior in decreasing pain intensity, enhancing self-reported behavioural changes in personal life and improving pain coping ability at work.