Table 2.

HIV prevalence by demographic characteristics and province

Variable	WCP, n/N (%; 95% CI)	FSP, n/N (%; 95% CI)	P value	WCP and FSP, n/N (%; 95% CI)
N	108/802 (13.5%; 0% to 37.3%)	270/1104 (24.5%; 23.3% to 25.7%)	0.270	378/1906 (17.2%; 1.9% to 32.6%)
Age category (years)
15–19	7/101 (6.9%; 0% to 24.2%)	17/186 (9.1%; 3.3% to 15.0%)	0.753	24/287 (7.8%; 0% to 18.2%)
20–24	25/226 (11.1%; 0% to 34.4%)	68/363 (18.7%; 10.5% to 27.0%)	0.438	93/589 (14.0%; 0% to 28.4%)
25–29	33/178 (18.5%; 0% to 59.1%)	77/244 (31.6%; 28.2% to 34.9%)	0.425	110/422 (23.0%; 0% to 50.0%)
30–34	20/124 (16.1%; 0% to 34.8%)	60/141 (42.6%; 30.9% to 54.2%)	0.029	80/265 (24.0%; 8.6% to 39.5%)
35–39	12/88 (13.6%; 0% to 36.3%)	35/102 (34.3%; 27.5% to 41.1%)	0.073	47/190 (20.0%; 1.2% to 38.6%)
40–44	3/21 (14.3%; 0% to 47.5%)	12/30 (40.0%; 31.5% to 48.5%)	0.106	15/51 (23.3%; 1.1% to 45.5%)
45–49	–	0/3 (0; –)	–	–
≥50	0/1 (0; –)	–	–	–
Marital status
Never married	57/395 (14.4%; 0% to 37.7%)	105/547 (19.2%; 12.5% to 25.8%)	0.614	162/942 (16.1%; 1.2% to 30.9%)
Currently married	36/305 (11.8%; 0% to 37.7%)	78/318 (24.5%; 16.4% to 32.7%)	0.263	114/623 (15.4%; 0% to 34.9%)
Divorced/separated/  widowed	4/31 (12.9%; 0% to 44.8%)	50/102 (49.0%; 35.2% to 62.9%)	0.045	54/133 (32.9%; 6.8% to 59.0%)
Married/cohabited before but current status unrecorded	11/71 (15.5%; 3.0% to 28.0%)	37/137 (27.0%; 20.9% to 33.1%)	0.083	48/208 (20.3%; 13.0% to 27.7%)
Highest educational level
Primary	13/92 (14.1%; 1.1% to 27.2%)	52/171 (30.4%; 24.5% to 36.3%)	0.034	65/263 (20.8%; 11.9% to 29.8%)
Secondary (any grade)	86/593 (14.5%; 0% to 38.5%)	208/866 (24.0%; 22.7% to 25.3%)	0.333	294/1459 (17.9%; 3.2% to 32.5%)
Finished secondary school	13/145 (9.0%; 0.0% to 25.1%)	37/208 (17.8%; 14.3% to 21.3%)	0.214	50/353 (12.1%; 2.3% to 21.8%)
Tertiary	9/114 (7.9%; 0% to 34.3%)	8/54 (14.8%; 0% to 29.8%)	0.561	17/168 (8.9%; 0% to 32.1%)
Employed
Yes	28/245 (11.4%; 0.0% to 38.3%)	37/131 (28.2%; 25.1% to 31.4%)	0.159	65/376 (14.2%; 0.0% to 37.8%)
No	80/556 (14.4%; 0.0% to 36.1%)	233/973 (23.9%; 22.4% to 25.5%)	0.290	313/1529 (18.2%; 6.0% to 30.4%)
Wealth quintile*
Lowest quintile  (poorest)	29/157 (18.5%; 10.7% to 26.2%)	66/218 (30.3%; 24.4% to 36.1%)	0.028	95/375 (22.5%; 18.2% to 26.8%)
Second lowest quintile	38/157 (24.2%; 6.1% to 42.3%)	55/217 (25.3%; 23.2% to 27.5%)	0.871	93/374 (24.6%; 13.0% to 36.2%)
Middle quintile	20/167 (12.0%; 0.0% to 35.4%)	50/218 (22.9%; 17.1% to 28.8%)	0.276	70/385 (15.6%; 0.0% to 31.5%)
Second highest quintile	12/159 (7.5%; 0% to 26.2%)	53/219 (24.2%; 21.5% to 26.9%)	0.070	65/378 (13.2%; 0.0% to 28.5%)
Highest quintile  (wealthiest)	5/144 (3.5%; 0% to 14.8%)	44/215 (20.5%; 6.5% to 34.4%)	0.059	49/359 (13.6%; 0% to 24.0%)
Setting
Urban formal	18/455 (4.0%; 0% to 9.6%)	NA	NA	NA
Urban informal	90/347 (25.9%; 21.4% to 30.9%)	124/520 (23.8%; 20.2% to 27.7%)	<0.001	214/867 (25.2%; 23.3% to 27.1%)
Rural formal	NA	146/584 (25.0%; 24.6% to 25.4%)	NA	NA

*Household wealth quintiles were generated separately for each province using data for the specified province only, without weighting. The two-sample test of equality of proportions test statistic for survey data was used to generate P values. This test statistic involves use of the ‘lincom’ command in STATA and takes consideration of study design, that is, the two-stage stratified cluster random sampling and design weights, where appropriate, to generate point estimates, their CIs and P values. The wider CI for WCP than FSP is due to clustering, as described in text. Furthermore, please note that WCP had larger probability weights, used in weighting of the survey-derived data than FSP, as described in text.

–, insufficient data for generation of estimates; differences in totals is due to missing data; %, prevalence of HIV adjusted for design effect; FSP, Free State Province; NA, not applicable; WCP, Western Cape Province.