Table 2.
Overview of included studies.
| n of studies | % | |
|---|---|---|
| Therapy types | ||
| Manual therapy with spinal manipulation | 48 | 24.2 |
| Craniosacral therapy and gentle myofascial release | 22 | 11.1 |
| Other manual therapy | 64 | 32.3 |
| Rehabilitation or physiotherapy | 22 | 11.1 |
| Self-management | 5 | 2.5 |
| Cognitive-behavioural and other psychotherapy | 27 | 13.6 |
| Spiritual or energetic or esoteric healing | 8 | 4.0 |
| Other | 2 | 1.0 |
| Intervention complexity† | ||
| Simple | 112 | 56.6 |
| Complex | 86 | 43.4 |
| Pain descriptor | ||
| Musculoskeletal pain | 121 | 61.1 |
| Diffuse chronic pain | 18 | 9.1 |
| Cancer-related pain | 6 | 3.0 |
| Visceral pain | 5 | 2.5 |
| Neuropathic pain | 5 | 2.5 |
| Pregnancy-related pain | 1 | 0.5 |
| Not specified | 1 | 0.5 |
| Median | Q1/Q3 | |
|---|---|---|
| Sample size at randomization | ||
| Overall sample size (all trial arms combined) | 64 | 40/101 |
| Sample size per trial arm (only groups included in review) | 27 | 19.5/46 |
| n | % | |
|---|---|---|
| Registered trial protocol available | ||
| Registered | 114 | 57.9 |
| Group design | ||
| Parallel group | 163 | 92.9 |
| Cross-over | 14 | 7.1 |
| No. of study conditions per trial* | ||
| 2 | 139 | 71.6 |
| 3 | 46 | 23.7 |
| 4 | 8 | 4.1 |
| 5 | 1 | 0.5 |
| 6 | 1 | 0.5 |
| Additional nonsham comparators included | ||
| Active comparator (comparative effectiveness) ± | 29 | 14.9 |
| No treatment or waitlist | 20 | 10.3 |
| Usual care/treatment as usual | 13 | 6.7 |
The types of therapies, intervention complexity, and pain population are provided for the entire sample. Special cases: In 1 trial, data from the active intervention group were used twice to compare it with 2 different sham controls: Bialosky et al. (2014) used a “standard” and an “enhanced” sham control. Three publications reported more than 1 trial: D'Souza et al. (2008) studied 2 groups with different types of headaches, and the publication of Assefi et al. (2008) included 2 active interventions and a matching sham control each. Finally, Sharpe et al. (2012) reported 2 trials in a single publication, which were treated entirely independently here. In general, only patients who informed the present analyses are counted in this table; patients were not counted twice, and analyses of reporting refer to individual trials.
Each intervention or sham intervention was counted, irrespective of whether the trial was a single-arm cross-over trial. ± So-called attention controls were not counted as active comparator, only experimental conditions that were clearly assessed because they were deemed potentially effective alternatives (comparative effectiveness intention).
Intervention complexity: Single-step or single-technique interventions were judged as “simple,” irrespective of how often these were applied, and others as complex. N = 194 publications with 198 comparisons between treatment and a sham control.