TABLE 3.

The relative chance of being medication adherend dependent on the level of shared decision making, the role of three suggested mediators and controlled for the covariates.

	TAI binary (0–49 versus 50)						TAI binary (0–45 versus 46–50)
	Total study population (N = 396)		COPD (+/-asthma)		Asthma (N = 202)		Total study population (N = 396)
			(N = 194)
	OR (95% CI)	% Attenuation	OR (95% CI)	% Attenuation	OR (95% CI)	% Attenuation	OR (95% CI)	% Attenuation
Baseline model 1	1.005 (0.988–1.022)	NA	1.010 (0.989–1.032)	NA	0.994 (0.967–1.021)	NA	1.001 (0.981–1.020)	NA
Plus covariates a	1.004 (0.987–1.021)	−0.1% c	1.008 (0.985–1.031)	−0.2% c	0.993 (0.966–1.021)	−0.1% c	1.000 (9.979–1.021)	−0.1% c
Plus autonomy	1.004 (0.984–1.024)	0.0% d	1.009 (0.979–1.039)	0.1% d	0.995 (0.967–1.025)	0.2% d	0.998 (0.974–1.023)	−0.2% d
Plus competence	1.001 (0.983–1.019)	−0.3% e	1.007 (0.983–1.031)	−0.2% e	0.990 (0.962–1.018)	−0.5% e	0.993 (0.971–1.016)	−0.5% e
Plus relatedness	1.003 (0.985–1.021)	−0.1% f	1.003 (0.980–1.027)	−0.4% f	1.011 (0.968–1.029)	2.1% f	0.998 (975–1.020)	0.2% d
Fully adjusted model b	1.002 (0.982–1.023)	−0.2% g	1.008 (0.978–1.038)	−0.5% g	0.997 (0.966–1.029)	−1.4% g	0.994 (0.969–1.020)	−0.6% g

^a age, sex, illness perception, social support, socio-economic status.

^bbaseline model with all covariates, autonomy, competence and relatedness.

^c Percent attenuation= (β_{model 1 + covariates}–β_{model 1})/(β_{model 1}) × 100.

^d Percent attenuation= (β_{model 1 + covariates + autonomy}–β_{model 1+ covariates})/(β_{model 1+ covariates}) × 100.

^e Percent attenuation= (β_{model 1 + covariates + competence}–β_{model 1+ covariates})/(β_{model 1+ covariates}) × 100.

^f Percent attenuation= (β_{model 1 + covariates + relatedness}–β_{model 1+ covariates})/(β_{model 1+ covariates}) × 100.

^g Percent attenuation= (β_{fully adjusted}–β_{model 1+ covariates})/(β_{model 1+ covariates}) × 100.