Although cognitive behavioural treatments (CBT) have proven efficacy in improving symptom management, pain-related distress, physical performance and return to work. few studies have examined the relationship between changes in behavioural process variables during treatment and improvement in outcome variables following treatment. We designed a multimethod assessment strategy to test the relative contribution of changes in physical capacity and pain-related anxiety to treatment outcome variables. Low back pain patients (n = 59) were treated with an intensive programme of physical exercise and CBT. Comparisons from pre- to post-treatment showed significant improvement in pain severity, interference, affective distress, activity level, and depression. Improvements in pain-related anxiety were associated with improvements in all outcome variables except interference. Of three physical capacity composite scores, improvement in only one (lumbar extension and flexion capacity) was associated with improvements in all outcome variables except interference. Further analyses demonstrated that the relationship between changes in pain-related anxiety and treatment outcome were independent of changes in physical capacity performance. Changes during treatment in pain-related anxiety may be more important than changes during treatment in physical capacity when predicting the effect of treatment on behavioural outcome measures. These results are discussed in the context of how to improve assessment of the chronic pain patient and improve the effectiveness of multidisciplinary CBT.