Table 5

Analysis of effect size, heterogeneity, and ANOVA testing of difference by study type.

Outcome Measure	Study Design	Number of Studies Included in This Analysis (Multiple Study Groups Possible)	Effect Size	Lower Limit	Upper Limit	Higgins I ² Statistic of Heterogeneity	Variance	SE	N
Oswestry Disability Index	Prospective study	8	0.890	0.821	0.960	0%	0.0013	0.0354	1774
	Randomized prospective study	11	0.843	0.729	0.957	0%	0.0034	0.0581	1374
	Retrospective study	32	0.884	0.846	0.922	0%	0.0004	0.0196	3314
	ANOVA Q random effects test with separate estimates of T ²				P = 0.742		Total N		6462
VAS-back	Prospective study	3	0.425	0.047	0.803	3%	0.0371	0.1926	284
	Randomized prospective study	11	0.806	0.653	0.959	0%	0.0061	0.0780	1542
	Retrospective study	15	0.715	0.587	0.843	0%	0.0043	0.0653	2560
	ANOVA Q random effects test with separate estimates of T ²				P = 0.112		Total N		4386
VAS-leg	Prospective study	6	0.912	0.885	0.938	0%	0.0002	0.0136	1632
	Randomized prospective study	11	0.860	0.733	0.987	0%	0.0042	0.0650	1542
	Retrospective study	35	0.866	0.819	0.912	0%	0.0006	0.0238	4716
	ANOVA Q random effects test with separate estimates of T ²				P = 0.175		Total N		7890