Table 4.

Summary of Results: Minimal Important Difference (MID)

Measure	Range	Number of studies	N per study (ref)	Minimal important difference (MID)
				Estimated using a clinical anchor				Estimated using statistical approaches
				ROC/optimal cutoff	95% limit cutoff	Average change among responders	Change difference, responders vs. non-responders	Minimal detectable change	Smallest detectable difference	SEM*
Oswestry Disability Index	0–100	3	4759	4.0		8.2	8.3	2.0†
			6355	7.5				16.7‡
			10745					10.7§
Visual Analog Scale	0–10 mm	1	4759	3.0		3.2	2.0	2.2†
Visual Analog Scale	0–100 mm	1	11869						49#
Bodily Pain Index	0–10	2	20550							0.6
			22250							0.7
PEG	0–10	2	20550							1.8
			22250							1.9
Chronic Pain Grade-intensity	0–100	2	20550							9.0
			22250							9.9
Chronic Pain Grade-disability	0–100	2	20550							8.7
			22250							10.3
Roland Morris Disability Questionnaire	0–24	4	6355	3.5				4.9‡
			20550							1.0
			22250							1.2
			14334					7.5‡
SF-36 Bodily Pain Scale	0–100	2	20550							9.8
			22250							11.8
Numeric Rating Scale	0–10	3	6355	4.0				2.4‡
			13537	3.5	4.7
			13868	2.5		3.7		4.5‡

*Standard error of measurement

†Estimated by the upper value of the 95% confidence interval for average change score seen in the cohort defined by anchor to be non-responders

‡Estimated by 1.96 × square root of 2 × SEM test-retest

^§SEM determined from participants classified as stable. The SEM was then used to calculate the 90% CI and then multiplied by the square root of 2

^#Estimated by the standard deviation of the difference values × 1.96