Table 3.

Parameters estimates for régression models (GEE), assessing for the relation between outcome and (a) psychotrope use, (b) benzodiazepine, and (c) any psychiatric disorders.

Parameter	Estimate	SE (robust)	Wald z	p-Value	OR (95% CI)
A
Intercept	0.258	0.560	0.21	0.645	1.294 (0.432; 3.874)
Time	−0.013	0.013	0.97	0.325	0.987 (0.961; 1.013)
Use of psychotrope (yes)	0.725	0.233	9.71	0.002	2.064 (1.308; 3.255)
Gender (female)	0.152	0.240	0.40	0.526	1.165 (0.727; 1.864)
Age	−0.016	0.001	2.60	0.107	0.984 (0.965; 1.003)
B
Intercept	0.411	0.559	0.54	0.462	1.508 (0.504; 4.513)
Time	−0.013	0.013	0.91	0.340	0.987 (0.962; 1.013)
Use of benzodiazepine (yes)	0.493	0.225	4.80	0.028	1.637 (1.053; 2.543)
Gender (female)	0.209	0.238	0.77	0.380	1.233 (0.773; 1.966)
Age	−0.015	0.001	2.43	0.119	0.985 (0.966; 1.004)
C
Intercept	−0.115	0.581	0.04	0.843	0.891 (0.285; 2.781)
Time	−0.014	0.014	1.00	0.318	0.987 (0.961; 1.013)
Psychiatric disorders (yes)	0.875	0.235	13.91	0.000	2.400 (1.515; 3.801)
Gender (female)	0.128	0.238	0.29	0.592	1.136 (0.712; 1.812)
Age	−0.009	0.010	0.76	0.383	0.991 (0.971; 1.011)

Reference categories are reported in parenthesis for categorical predictors.