The aim of this cost-effectiveness study was to compare a combined operant programme plus cognitive/relaxation programme with an operant programme plus attention-control and to compare both programmes with a waiting-list control group and with operant rehabilitation provided, as usual, by the same rehabilitation centre. One hundred and forty eight patients with chronic low back pain were randomly assigned to the different conditions. The economic endpoints were the costs of the programme and other health care utilisation, costs for the patient, and indirect costs associated with production losses due to low back pain. The effects were measured in terms of global assessment of change and utilities, using rating scale and standard gamble methods. The 3-year study determined that adding a cognitive component to an operant treatment did not lead to significant differences in costs and improvement in quality of life when compared with the operant treatment alone. Compared with the common individual rehabilitation therapy it can be concluded that the same effects can be reached at the same or lower costs with a shorter, more intense standardised group programme. The operant treatment alone is more effective than providing no treatment in the waiting-list control group.