The effect of intensive dynamic back extensor exercises for patients with chronic low back was investigated in a controlled clinical trial where the patients underwent a three months intensive training program a total 30 sessions. The 105 patients were divided into three groups: a treatment group, an alternative group which underwent 1/5 of the treatment group's exercise program per session, and an alternative group in which the treatment consisted of thermotherapy, massage and mild exercise. Regardless of whether the outcome of treatment is assessed qualitatively or quantitatively, a statistically significant, favourable difference was found between the results for the treatment group and those for the alternative groups. The therapy was found to be without risk, but patients with clinical signs of current lumbar nerve root compression or radiological signs of spondylolysis or halisteresis of the column were excluded from the study. The authors recommend the introduction of this form of treatment as a general offer in the primary and secondary health sector. Back exercises can advantageously be conducted in a group; thus, the resources required for this treatment do not differ significantly from those required for many of the other forms of treatment generally employed, which often require individual attention.