Maternal mortality in South Africa in 2001: From demographic census to epidemiological investigation

Table 3 Risk factors of maternal mortality at the household level, South Africa, 2001 (from linear logistic regression)

Household characteristics	Beta	Standard error	Net effect	Relative risk	P-value	Signif.
Level of education	-0.0592	0.0131	287	0.83	0.0000	*
Wealth distance	-0.0651	0.0231	298	0.86	0.0048	*
Urban areas	0.2477	0.1076	444	1.28	0.0213	*
Black/African (ref)			347	1.00
Coloured	-0.3624	0.2520	242	0.70	0.1504
Indian	-0.9890	0.5889	129	0.37	0.0931
White	-1.6661	0.5859	66	0.19	0.0045	*
Western Cape (Ref)			347	1.00
Eastern Cape	0.8304	0.2709	792	2.28	0.0022	*
Northern Cape	0.3743	0.4357	504	1.45	0.3902
Free State	0.6608	0.3017	670	1.93	0.0285	*
Kwazulu Natal	0.9320	0.2664	876	2.53	0.0005	*
North-West	0.8292	0.2857	791	2.28	0.0037	*
Gauteng	0.2955	0.2745	466	1.34	0.2818
Mpumalanga	0.4790	0.3023	559	1.61	0.1131
Limpopo	0.0882	0.3050	379	1.09	0.7724
Constant	-5.0586	0.2911	347

N.B. Reference categories: Race = Black/African, Province = Western Cape. Net effects are calculated for dummy variables, and for one standard deviation for quantitative variables (education and wealth). Education is measured in years of schooling. Wealth distance is counted as the distance in the wealth index from the average (6 items).

ISSN: 1478-7954