Skip to main content

Table 4 Comparing population-based and ANC sentinel-surveillance HIV prevalence estimates

From: Population and antenatal-based HIV prevalence estimates in a high contracepting female population in rural South Africa

   

ANC based HIV prevalence estimates1

Population based HIV prevalence estimates

Ratio of ANC and population based estimates2

C hi2 (p)3

All women

Crude

37.9

25.2

1.50

< 0.001

  

15–19

19.5

6.9

2.83

< 0.001

  

20–24

37.0

28.7

1.29

0.002

  

25–29

54.3

49.6

1.09

0.254

  

30–34

51.7

47.0

1.10

0.362

  

35+

3 6.5

27.7

1.32

0.053

  

Age standardised

36.2

27.5

1.32

< 0.001

  

Age & location 4

31.7

27.5

1.15

0.005

  

Age & catchment 5

35.6

27.7

1.29

< 0.001

Nulliparous

Crude

24.8

11.9

2.08

< 0.001

  

Age standardised

21.3

15.5

1.37

0.003

  

Age & location 4

20.1

15.9

1.26

0.037

  

Age & catchment 5

20.6

16.0

1.29

0.022

Parity 1+

Crude

46.4

33.0

1.41

< 0.001

  

Age standardised

43.0

34.4

1.25

< 0.001

  

Age & location 4

37.0

34.1

1.09

0.153

  

Age & catchment 5

41.6

34.5

1.21

< 0.001

Pregnant women

Crude

-

23.8

1.59

< 0.001

  

Age standardised

-

25.2

1.44

< 0.001

  

Age & location 4

-

25.3

1.25

0.006

  

Age & catchment 5

-

25.0

1.42

< 0.001

Clinic location (ANC); residency location (population)

Urban

Crude

40.7

26.0

1.56

0.006

  

Age standardised

39.2

27.1

1.47

0.023

 

Peri-urban

Crude

36.5

33.5

1.09

0.277

  

Age standardised

36.8

36.5

1.01

0.930

 

Rural

Crude

29.2

21.3

1.37

0.028

  

Age standardised

28.5

23.2

1.22

0.151

Projected clinic catchment (ANC); residence within projected clinic catchment (population)

Peri-urban/urban

Crude

40.7

30.5

1.33

< 0.001

  

Age standardised

38.2

33.5

1.14

0.039

 

Rural/peri-Urban

Crude

39.4

27.2

1.45

0.001

  

Age standardised

38.2

29.3

1.30

0.018

 

Rural

Crude

32.0

20.8

1.54

< 0.001

  

Age standardised

32.3

22.7

1.42

< 0.001

  1. 1 18 women excluded from the analyses with age not known
  2. 2 Ratio based on ANC totals when no direct comparable ANC group
  3. 3 Difference between two proportions (ANC-based HIV prevalence estimate compared to the population-based estimate)
  4. 4 Standardisation based on age and residency location (urban, rural and peri-urban) distribution within the ACDIS population
  5. 5 Standardisation based on age and predicted ANC clinic catchment (peri-urban/urban, peri-urban/rural and rural) distribution within the ACDIS population
Back to article page

BMC Public Health

ISSN: 1471-2458

Contact us