Table 2.

Odds ratios of MDE2010 associated with separation based on nested logistic models without (Model 1) and with (Model 2) the introduction of loss of income and loss of support in men and women in a union in 2006

	Women		Men
	Model 1	Model 2	Model 1	Model 2
Total and direct effects of separation	Total 1.8**	Direct 1.6	Total 3.9***	Direct 3.7***
Changes in income and support (2006–2010):
Income decline or stagnation versus increase	–	1.6**	–	1.2
Loss of emotional support (Yes vs. No)	–	2.2**	–	2.3**
Covariates:
Time since separation (continuous)	1.1	1.1	0.6*	0.6*
MDE2006 (Yes vs. No)	6.3***	5.8***	6.4***	5.9***
Income in 2006 (cont.)	1.0	1.0	1.0	1.0
Duration of the broken union (cont.)	1.0	1.0	1.0	1.0
Number of unions prior to 2006 union (cont.)	1.3*	1.3*	0.9	0.9
Number of cohabiting children in 2006 (cont.)	1.0	1.0	1.3***	1.3***
Unemployed versus Employed in 2006	0.9	0.9	2.2	2,2
Retired versus Employed in 2006	1.1	1.1	0.9	0,9
Inactive versus Employed in 2006	0.9	0.9	2.8***	2.8***
Middle level of education versus low level	1.0	1.0	0.9	0.9
High level of education versus low level	0.5***	0.5**	0.7	0.7
Age in 2006 (cont.)	1.0	1.1	1.1	1.1
Age squared	1.0	1.0	1.0	1.0
Number of current health problems in 2006 (cont.)	1.3***	1.2***	1.3***	1..3***
Number of health problems before 2006 (cont.)	1.0	1.0	1.4**	1.4**
Adverse health event/s as child (Yes vs. No)	1.9***	1.7***	2.1**	2.2**
N	3743		3578

Statistically significant at the level ***p < 1%; **p < 5%; *p < 10% (two tailed test)

Model 1: Nested logistic regression of the risk of MDE2010 providing the total effect of separation, i.e., without the variables of changes in income and support between 2006 and 2010 (but including the residuals) (see Box 1). Model 2: Nested logistic regression of the risk of MDE2010 providing the direct effect of separation, i.e., considering the indirect effect of variables of changes in income and support between 2006 and 2010