Table 2

Univariate and Multivariate Logistic Regression to Characterize Associations with HIV Status among MSM in Peri-Urban Cape Town

Characteristic

Total

HIV Positive

Univariate

Multivariate

n (%)

Yes n (%)

No n (%)

OR (95% CI)

aOR (95% CI)

Not always wearing condoms with men

88 (47.6%)

18 (20.5%)

70 (79.5%)

1.9 (1.0-3.7), *

2.3 (1.0-5.4),**

Received money for Transactional Sex

24 (12.0%)

11 (45.8%)

13 (54.2%)

2.9 (1.2-6.9),*

NS

Having been blackmailed

21 (10.5%)

9 (42.9%)

12 (57.1%)

2.4 (1.0-6.2) (p = 0.06)

4,4 (1.3-15.4), *

Having been raped

22 (11.0%)

11 (50.0%)

11 (50.0%)

3.5 (1.4-8.5),**

5.7 (1.6-20.2) **

Higher number of male partners (>5)

35(17.7%)

15 (42.9%)

20 (57.1%)

2.8.(1.3-6.1), **

NS

More than 26 years old

77 (38.5%)

26 (33.8%)

51 (66.2%)

2.0 (1.1 - 3.8), *

4.2 (1.6-10.6),**

Unemployment

79 (39.7%)

27 (34.2%)

52 (65.8%)

2.2 (1.1-4.2), *

3.7 (1.5-9.3),**

Less Educated

20 (10.0%)

10 (50.0%)

10 (50.0%)

3.4 (1.3.-8.8), *

NS

Rural Origin

102 (53.6%)

64 (62.7%)

38 (37.3)%)

4.1 (2.0-8.8),**

6.0 ( 2.2-16.7), **

± Not all columns add up due to missing values

*p < 0.05

**p < 0.01

NS - Not Significant

Baral et al. BMC Public Health 2011 11:766   doi:10.1186/1471-2458-11-766