Table 4

Logistic regression analysis of variables related to respondents believing themselves at risk of getting HIV

Variables
Crude OR
Adjusted OR
95% CI of adjusted OR
Mantel-Haenszel chi-square

Rural location
1.56
1.31
1.05–1.63
5.93
Household income sufficient for needs
0.43
0.46
0.36–0.58
40.78
Less than 25 years old
0.52
0.53
0.41–0.67
27.79
Beaten by partner in last 12 months
2.63
2.69
1.88–3.85
29.51

The variables included in the original saturated model were: rural/urban location, type of roof construction, household food sufficiency, income sufficiency, sex of respondent, age group, marital status, education level, and whether beaten by partner.

Cockcroft et al. BMC International Health and Human Rights 2007 7:5   doi:10.1186/1472-698X-7-5