Recently, Christensen et al. reported the clinical effects of a low-cost rehabilitation program equally efficient to a relatively intensive program of individual, physiotherapist-guided exercise therapy. Yet, the low-cost approach is not fully supported as an optimal strategy until a full-scale economic evaluation, including extra-hospital effects such as service utilization in the primary health care sector and return-to-work, is conducted. The objective of this study was to conduct such evaluation i.e. investigate the cost-effectiveness of (1) a low-cost rehabilitation regimen with a behavioural element and (2) a regimen of individual exercise therapy, both in comparison with usual practice, from a health economic, societal perspective. Study design was a cost-effectiveness evaluation of an RCT with a 2-year follow-up. Ninety patients having had posterolateral or circumferential fusion (indicated by chronic low back pain and localized pathology) were randomized 3 months after their spinal fusion. Validated pain- and disability index scales were applied at baseline and at 2 years postoperative. Costs were measured in a full-scale societal perspective. The probability of the behavioural approach being cost-effective was close to 1 given pain as the prioritized effect measure, and 0.8 to 0.6 (dependent on willingness to pay per effect unit) given disability as the prioritized effect measure. The probability of the exercise therapy approach being cost-effective was modest due to inferior effectiveness. Results proved robust to relevant sensitivity analysis although a differentiated cost-effectiveness ratio between males and females was suspected. In conclusion, a simple behavioural extension, of setting up group meetings for patients, to a regimen with a strict physiotherapeutic focus was found cost-effective, whereas the cost-effectiveness of increasing frequency and guidance of a traditional physiotherapeutic regimen was unlikely in present trial setting.