Table 10. Preference for advance refusal of all medication: univariate and corrected associations.

Univariate regression of age and years since diagnosis, and all multivariate regression was conducted using multiple imputation.

		Univariate regression Odds ratio (95% CIs)	Model controlling for age, gender, ethnicity, education and comorbidity Odds ratio (95% CIs)
Age		0.97 (0.95 – 0.98)          p < 0.0005 **	0.97 (0.96 – 0.98)          p < 0.0005 **
		n = 924	n = 897
Years since diagnosis		0.97 (0.95 – 0.99)          p < 0.0005 **	0.99 (0.97 – 1.01)              p = 0.322
		n = 928	n = 897
Hospitalisation	Never	1.00	1.00
	Ever	0.71 (0.44 – 0.98)                 p = 0.038	0.75 (0.54 – 1.06)              p = 0.103
		n = 921	n = 891
Gender	Male	1.00	1.00
	Female	1.75 (1.18 – 2.61)                p = 0.005	1.58 (1.05 – 2.38)                p = 0.029
	Other	2.56 (0.63 – 10.36)              p = 0.188	2.70 (0.64 – 11.40)              p = 0.177
		n = 926	n = 897
Trust in healthcare professionals	Agree	1.00	1.00
	Neutral	1.27 (0.84 – 1.94)              p = 0.260	1.22 (0.79 – 1.89)              p = 0.366
	Disagree	2.35 (1.61 – 3.41)          p < 0.0005 **	2.08 (1.41 – 3.08)          p < 0.0005 **
		n = 923	n = 897

**indicates statistical significance after correction for multiple comparisons.