Systematic reviews have shown that as non-operative treatments exercise, behavioural and multimodal treatment programs are effective for chronic low back pain. There is, however, a lack of knowledge concerning the association between changes in treatment process variables and changes in outcome for the three treatment forms. The objective of this systematic review was to evaluate which changes in treatment process variables predict outcome of exercise, behavioural and multimodal treatment of chronic low back pain. Medline, Embase and PsychInfo were systematically searched. A descriptive analysis was used to summarize the results regarding the outcomes pain, disability and return to work (RTW). 13 studies were identified. The results showed that functional coping mechanisms and pain reduction were associated with a decrease in disability and increase in RTW, and physical performance factors were not. Related to pain reduction decreases in disability, functional coping mechanisms as well as physical performance factors were associated. Strong conclusions cannot be drawn from this review, because of the heterogeneity and the limited number of studies. The results of this review raise the question if changes in behavioural variables and reductions of disability which facilitate an improvement in function, may be more important than physical performance factors for successful treatment of chronic low back pain. This is relevant for the refinement of future treatment programs.