Forced residential mobility and social support: impacts on psychiatric disorders among Somali migrants

Table 4 Associations with psychiatric disorder from stepwise logistic regression models: only retained variables shown

Model 1		OR	95%CI	p value	N	R²
Moved in last 5 years	Yes	1
Gender	No	0.21	0.07-0.62	0.01	128	0.21
	Men	1
	Women	0.35	0.14-0.86	0.02
Friendship network	Low	1
	High	0.35	0.14-0.84	0.02
Relative network	Low	1
	High	0.46	0.2-1.07	0.07
Confiding relationships	Low	1
	High	0.42	0.18-1.0	0.05
Physical health sf12	Poor/fair	1
	Good/excellent	0.48	0.18-1.32	0.15
Employment in UK	No	1
	Yes	0.21	0.04-1.23	0.08
Model 2
Choice over move	Own choice or not moved	1			137	0.2
Gender	Other's choice	2.64	1.16-5.98	0.02
	Men	1
	Women	0.5	0.22-1.16	0.11
Friendship network	Low	1
	High	0.34	0.15-0.79	0.01
Relative network	Low	1
	High	0.48	0.21-1.09	0.08
Confiding relationships	Low	1
	High	0.43	0.19-0.99	0.05
Physical health sf12	Poor/fair	1
	Good/excellent	0.41	0.16-1.05	0.06
On job seekers allowance	No	1
	Yes	1.97	0.84-4.61	0.12

Stepwise logistic regression models; psychiatric disorder as an outcome. Modelled by forward entry if significant to a p < 0.2 level. Variables entered including age, gender, discrimination score, networks as binary variables using median split (relatives, friendship and confiding and emotional relationships), physical health using median split on the SF12, employment in the UK (yes/no), on benefits (job seekers' allowance, or income support, marital state (single, married, sep/widowed/divorced), legal status (asylum seeker/refugee status), tenure (temporary/permanent), period living in the UK, index of multiple deprivation of the current address and trauma history (yes/no).

ISSN: 1472-698X