Table 3.

Proportion of variation explained by grouping structure and covariates

			Variance explained in % (PEV) a
					Funding regulations
	Variance b	ICC c	Context	Health factors	Deductibles	% New treatments	% Nine sessions series	Responsiveness	Collinearity d	Total
By grouping level e
Physician	.021	.049	.000	.079	.000	.048	.251	.029	.323	.730
Physio.	.028	.063	.022	.033	.000	.055	.174	.032	.274	.590
Patient	.393	.888	.000	.040	.001	.002	.000	.000	.002	.045
Overall	.442	1								.112

^aThe proportion of explained variation (PEV, i.e. squared semi-partial correlation coefficient) represents the amount of variance that is explained by the regressors included in the model.

^bTotal variance potentially explained at all levels.

^cThe intra-class correlation coefficient (ICC) allows the partitioning of the total variability in the outcome into its three variance components: physicians, physiotherapists and patients.

^dUnless the regressors are all orthogonal, the prognostic factors’ specific PEVs do not add up to the total PEV, the difference representing the collinearity effect due to the inclusion of all regressors into the model.

^eThe third level (canton) was treated as a fixed effect and therefore no variance component appears in the disaggregation of the total variance.