We compared the results of two multimodal back treatment programs (program A: n = 134; program B: n = 175) using pain, functional capacity, sickness absence, subjective state of health, depression, and work status as outcome variables. Whereas in program A exercise was guided by pain, in program B a "no pain, no gain" rationale was used as a basis for intensive physical training. Neither of the programs included direct attempts to influence the patient's environment (i.e., to find employment or to modify working conditions). In both programs, a clear decrease in pain and an increase in functional capacity was found from the pretreatment phase to the 12-month follow-up. These changes were, however, stronger in program B. Days of absence decreased significantly in group B, but the change was not significantly different from that in group A. On the whole, the more intensive training showed somewhat better results, even though the difference was not substantial. In contrast to some earlier results, there was no statistically significant increase in the proportion of those employed after treatment in either group. The results indicate that intensive physical exercise does not, as such, solve the problem of back disability in a country that has a highly developed social security system. More active interventions in work and work-life are needed.