Twenty-eight studies were included in the review (3,669 participants). Sample size ranged from 20 to 570. Lack of reliable sham treatment of controls and inadequate blinding were reported. Follow-up was insufficient in most studies to determine maintenance of any intervention effect. The lack of reporting of variables that might influence the effect of an intervention was identified as a problem.
Twenty-one of 28 studies demonstrated statistically significant differences between non-pharmacological treatment groups and no intervention or sham intervention; those who received non-pharmacological interventions reported lower pain intensity.
Eighteen studies focused on physical interventions whereas ten considered psychosocial interventions. Seven of 10 psychosocial interventions showed efficacy of the intervention for pain and/or disability.
Results were also reported by intervention.