Twenty-three studies were identified (n=1,396 participants, range six to 133). The mean validity score for studies was 11.22 (SD 2.96, range 5 to 16); seven studies scored less than 10. There were 30 intervention groups. ITT data were available for 11 studies. Where available, drop-out rates were 20.85% for intervention groups and 20.06% for control groups. Where available, follow-up ranged from two to 48 months.
Psychological interventions had a significant effect on short-term (Hedges’s g=0.37, 95% CI 0.27 to 0.48; number of interventions in analysis (nt) = 26) and long-term pain reduction (Hedges’s g=0.47, 95% CI 0.30 to 0.65; nt=15) for an average follow-up of 7.4 months. Psychological interventions were significantly effective in reducing sleep problems in the short term (Hedges’s g=0.46, 95% CI 0.28 to 0.64; nt=17) and long term (Hedges’s g=0.41, 95% CI 0.14 to 0.68; nt=12) and depression in the short term (Hedges’s g=0.33, 95% CI 0.20 to 0.45; nt=20) and long term (Hedges’s g=0.34, 95% CI 0.22 to 0.46; nt=16). Psychological interventions significantly affected functional status in the short term (Hedges’s g=0.42, 95% CI 0.25 to 0.58; nt=14) and long term (Hedges’s g=0.52, 95% CI 0.29 to 0.75; nt=11) and catastrophizing in the short term (Hedges’s g=0.33, 95% CI 0.17 to 0.49; nt=8) and long term (Hedges’s g=0.40, 95% CI 0.22 to 0.59; nt=7).
Analyses for controlled studies alone in the short term gave significant effects for pain reduction (Hedges’s g=0.34, 95% CI 0.05 to 0.64; nt=5) and depression (Hedges’s g=0.44, 95% CI 0.21 to 0.66; nt=12), but no significant effects for sleep problems (nt=3), functional status (nt=2) and catastrophizing (nt=4). Analyses for controlled studies plus studies with active comparisons found psychological interventions significantly reduced pain intensity (Hedges’s g=0.50, 95% CI 0.14 to 0.86; nt=10), depression (Hedges’s g=0.56, 95% CI 0.19 to 0.93; nt=8) and catastrophizing (Hedges’s g=0.47, 95% CI 0.11 to 0.82; nt=2) but no significant effect on sleep problems (nt=6) and functional status (nt=6).
Detailed results were reported for the moderator analyses. CBT was significantly better than other treatments for short-term pain reduction (Hedges’s g=0.60, 95% CI 0.46 to 0.76; nt=6 versus Hedges’s g=0.27, 95% CI 0.17 to 0.37; nt=18 for psychological treatments excluding CBT). CBT and relaxation/biofeedback were significantly better than other treatments for treating sleep problems. Higher treatment dose was associated with better outcomes.
There was little evidence of publication bias.