Construct validity of the OCTOPuS stratification algorithm for allocating patients with knee osteoarthritis into subgroups

Table 5 Comparison of observed effects of usual exercise therapy for each subgroup with hypothesized effects (research question 3; findings resulting in accepted hypotheses in bold)

	Low muscle strength subgroup		Obesity subgroup		High muscle strength subgroup
	Observed	Hypothesized	Observed	Hypothesized	Observed	Hypothesized
Knee pain
Effect size¹	1.05	0.8 ± 0.2	1.10	0.5 ± 0.2	0.82	0.2 ± 0.2
% persons with MIC²	70%	> 67%	72%	33–67%	66%	< 33%
Physical function
Effect size¹	0.79	0.8 ± 0.2	0.78	0.5 ± 0.2	0.49	0.2 ± 0.2
% persons with MIC³	79%	> 67%	76%	33–67%	76%	< 33%
Quad. strength / 30s-CST
Effect size¹	0.74/ 0.73	0.8 ± 0.2	0.27/ 0.60	0.5 ± 0.2	0.19/ 0.32	0.2 ± 0.2
% persons with MIC^4,5	49% / 31%	> 67%	32% / 28%	33–67%	7% / 9%	< 33%

MIC = minimal important change. ¹ Effect size (within-group) = change score within group / standard deviation at baseline; ² MIC defined as improvement on NRS/VAS pain (0–100) ≥ 15% and/or ≥ 1 point [25]; ³MIC defined as improvement on WOMAC physical function (0–100) ≥ 12% [26]; ⁴ MIC defined as improvement on quadriceps strength ≥30% [27]; ⁵ MIC defined as improvement on 30s-CST ≥ 2 repetitions [28]

ISSN: 1471-2474