Socioeconomic predictors and consequences of depression among primary care attenders with non-communicable diseases in the Western Cape, South Africa: cohort study within a randomised trial

Table 3 Patient characteristics independently associated with changes^a in unemployment and welfare grant income: logistic and linear regression models

Outcome	Unemployed at follow-up^b				Monthly welfare grant income at follow up (Rand)^c
Explanatory variable	OR^d	95 % CI		p	Coefficient	95 % CI^d		p
CESD-10 score ≥10 at baseline	1.25	1.04	1.51	0.016	55	18	91	0.004
Age (per year)	1.05	1.04	1.06	<0.001	9	7	11	<0.001
Men vs. women	0.70	0.57	0.86	0.001
Chronic respiratory disease					66	25	108	0.003
Diabetes					27	0	55	0.048
Highest education				<0.001^e				0.017 ^e
•None (reference)	1.00				0
•Primary	0.78	0.54	1.13	0.186	−38	−85	8	0.105
•Secondary	0.59	0.38	0.90	0.014	−67	−115	−19	0.007
•Tertiary	0.19	0.10	0.35	<0.001	3	−212	218	0.975
Language				0.052^e				<0.001^e
•Afrikaans (reference)	1.00				0
•Xhosa	0.68	0.46	1.00	0.047	−158	−228	−88	<0.001
•English	1.21	0.85	1.73	0.288	−99	−162	−37	0.003
Unemployed at baseline^a	13.9	10.7	18.2	<0.001
Grant income at baseline (per 1000 Rand per month)^a					602	522	682	<0.001

^a Change modelled with analysis of covariance, that is, with baseline value as covariate
^b Logistic regression model
^c Linear regression model
^d OR odds ratio, CI confidence interval
^e Wald test for all categories of variable

ISSN: 1471-2458