. 2016 Jul 29;6:30675. doi: 10.1038/srep30675

Table 5. Subgroup analysis of possible sources of heterogeneity explained by the design of sham acupuncture for individual conditions.

Covariate and classification (condition)	No. of studies or subsets	SMD (95% CI)^*	P for SMD	P for heterogeneity^†	I² (%)	P for interaction^‡	Adjusted R²^$
Sham location (SP)
Same	3	−0.84 (−1.14 to −0.54)	<0.001	0.633	0	0.46	100
Lateral	1	−0.47 (−1.46 to 0.53)	0.359	n.a.	n.a.
Irrelevant	1	−0.39 (−0.63 to −0.16)	0.001	n.a.	n.a.
Sham depth (LBP)
Non-penetration	5	−1.23 (−1.98 to −0.48)	0.001	0.001	78.7	0.30	62.69
Superficial	4	−0.20 (−0.31 to −0.08)	0.001	0.404	0
Normal	1	−0.50 (−0.85 to −0.14)	0.006	n.a.	n.a.
Sham location (MP)
Same	8	−1.37 (−1.88 to −0.86)	<0.001	0.001	70.4	0.46	46.46
Lateral	4	−0.19 (−0.57 to −0.19)	0.318	0.464	0
Irrelevant	1	−0.74 (−1.47 to −0.01)	0.047	n.a.	n.a.
Sham depth (MP)
Non-penetration	7	−1.48 (−2.11 to −0.84)	<0.001	0.001	74.9	0.11	47.20
Superficial	6	−0.47 (−0.79 to −0.15)	0.004	0.320	14.7
Normal	0
Sham depth (FM)
Non-penetration	3	−0.10 (−0.83 to 0.64)	0.790	0.124	52.1	0.51	100
Superficial	0
Normal	2	0.15 (−0.17 to 0.47)	0.365	0.601	0

FM, fibromyalgia; LBP, low back pain; MP, myofascial pain; SP, shoulder pain.

^*Standardized mean difference (SMD) obtained using conventional random-effects model.

^$Percentage of heterogeneity explained by covariate (%).

^†P values for heterogeneity across studies computed using Cochrane’s Q test.

^‡P values for comparisons between subgroups computed using χ² test.