Table 4.

Pain and disability outcome measures - post-intervention

Outcome measure	Study	Sample size			Statistic used	Heterogeneity	Effect size		Test for overall effect		Notes
		Intervention	Control	Total		I 2 (%)		[95 % CI]	Z-value	p-value
PAIN INTENSITY
Education versus usual care
“average pain”	Linton 1997; Soares 2002; Sparkes 2012; Ruehlman 2013	248	213	461	SMD random	0	−0.01	[−0.19, 0.17]	0.12	0.90	Figure 2
PPQ - pain in the last week	Ferrell 1997	10	10	20	MD random	n/a	−2.80	[−21.09, 15.49]	0.30	0.76	Sample >65 years
Comparison of different types of education
SF36 - bodily pain	van Oosterwijck 2013	15	15	30	MD random	n/a	−3.40	[19.98, 13.18]	0.40	0.69
DISABILITY
Education versus usual care
Disability or interference	Linton 1997; Ruehlman 2012; Sparkes 2012	230	196	426	SMD random	49	0.02	[−0.31, 0.34]	0.11	0.91	Figure 4
Comparison of different types of education
SF36 - physical function	van Oosterwijck 2013	15	15	30	MD random	n/a	5.30	[−8.64, 19.24]	0.75	0.46
Roland Morris Disability Questionnaire	Moseley 2004	31	27	58	MD random	n/a	−2.00	[−3.55, −0.45]	2.53	0.01	Favours education
Function and Disability (pooled data using negative RMDQ score for direct comparison)	van Oosterwijck 2013; Moseley 2004	46	42	88	SMD random	0	0.52	[0.09, 0.95]	2.38	0.02	Figure 6; favours education
SF36 - physical function	Ferrell 1997	10	10	20	MD random	n/a	6.70	[−9.11, 22.51]	0.83	0.41	Sample >65 years

PPQ patient pain questionnaire, SF-36 RAND 36-item health survey, 95 % CI 95 % confidence interval, effect size represented as standardised mean difference (SMD) or mean difference (MD) depending on statistic used; Random = random effects model; heterogeneity is not applicable (n/a) when reported as single study