University of California, San Francisco Logo

University of California, San Francisco | About UCSF | Search UCSF | UCSF Medical Center

Use of antenatal clinic surveillance to assess the effect of sexual behavior on HIV prevalence in young women in Karonga District, Malawi
Global Health Sciences Literature Digest
Published October 21, 2008
Journal Article

Crampin AC, Jahn A, Kondowe M, Ngwira BM, Hemmings J, Glynn JR, et al. Use of antenatal clinic surveillance to assess the effect of sexual behavior on HIV prevalence in young women in Karonga District, Malawi. J Acquir Immune Defic Syndr 2008;48:196-202.


To measure the association between sexual behavior and HIV prevalence using antenatal clinic (ANC) HIV sero-surveillance data

Study Design

Cross-sectional observational study


Five health care facilities (one urban hospital, two rural hospitals, and two rural health centers) in the Karonga District, northern Malawi


Women aged 24 years and younger attending an ANC between 1999 and 2004

Primary Outcomes

Association between duration of sexual activity (as estimated by sex before, during, between, and after marriage) and HIV prevalence.


Women <25 years old attending an ANC between 1999 and 2002 were screened for HIV using unlinked anonymous surveys and between 2003 and 2004 were screened as part of prevention of mother-to-child transmission (PMTCT) services. Women were administered questionnaires that collected the following information:

  1. Age at start of current marriage
  2. Age at start of first marriage (AAFM)
  3. Age at first sex (AAFS, collected from 2002 through 2004)
  4. Dissolution of marriage (yes or no, regardless of reason)
  5. Age at birth of first child.

Duration of "sexual exposure" was calculated using the following categories:

  1. Duration of premarital exposure (current age-AAFS) for never-married women
  2. Duration of first marriage for ever married women (AAFM-AAFS)
  3. Duration of first marriage for women whose marriage ended (age at end of first marriage-AAFM). Age at end of first marriage estimated as 2 years for each child, or two thirds of the interval between the start of the first marriage and the start of the second marriage for women without children, or two thirds of the interval between the start of the first marriage and the start of the second marriage and one third allocated to an intermarriage period of four fifths of the interval between the start of the first marriage and the current age for those who did not remarry.
  4. Duration after first marriage included all other time since onset of sexual activity and was divided into remarried and ex-married periods.

For women screened prior to 2002, AAFS was imputed as: minimum of [current age -0.5 years, age at first birth-0.75 years, age at first marriage - median premarital sexual exposure].

Estimated background HIV prevalence in each residence area was based upon the HIV prevalence from ANC surveys in women aged 25 years and older.

Logistic regression models were used to calculate unadjusted and adjusted odds ratios (ORs) and 95% confidence intervals (CIs) of HIV prevalence for variables of interest.


Included were 2,874 women under 25 years of age. Participation was over 99%. The distribution of premarital sexual exposure was consistent across all age groups, and time spent in and after first marriage was plausible. HIV prevalence was higher with increased premarital and post-marital exposure but was relatively constant (and lower) for women in their first marriage. Independent predictors of HIV infection included residence with higher background HIV prevalence (OR 1.04, 95% CI: 1.01-1.07), premarital exposure (OR 1.19, 95% CI: 1.01-1.40), exposure after marriage (OR 1.28, 95% CL: 1.11, 1.48), age of child's father (OR 1.04, 95% CI: 1.01-1.06), previous residence in Malawi but not in Karanga (OR 1.84, 95% CI: 1.21-2.81), previous residence outside of Malawi (OR 2.72, 95% CI: 1.42-5.20), and completion of some secondary school (OR 1.57, 95% CI: 1.05-2.34). The odds of HIV infection were reduced for women's birth cohort [1975-1979, 1980-1984, 1985-1989] (OR 0.91, 95% CI: 0.82-0.99).


Based upon the methods used to estimate duration of potential sexual exposure to HIV, sex outside of a first marriage increases the risk of HIV while time in a first marriage does not. ANC surveillance should include proxy measures of sexual behaviors.

Quality Rating

Based upon the Newcastle-Ottawa rating system for case-control studies, this is a moderately good study. Cases and controls were defined by HIV testing using standard methods. Subjects were women attending an ANC, and participation was high. The weakness of the study comes from the definition of exposure. The authors developed proxies for sexual exposure based upon marital histories. Such proxies have not been validated. Age at first sex (one of the most important variables needed for these proxies) was missing for a large portion of the study population and had to be imputed. The methods used for imputing this variable have not been validated. However, there is internal validity and comparisons between the groups are likely to be reasonable. The variables that significantly increased risk for HIV were of a small magnitude, limiting the usefulness of the findings.

In Context

HIV prevalence, as measured through ANC sero-surveillance, has been stable or declining among young women in some sub-Saharan African countries.(1,2) Although these declines have been noted, behavioral data from women in these surveys has not been collected. Examining the association between sexual behaviors and HIV is important in order to understand trends and to implement and evaluate prevention programs.(3,4,5) A key finding from this study is that women in a first marriage were not at continued risk for HIV. This finding is consistent with findings from Kenya and Zambia in which less than half of HIV infections were acquired in marriage.(6) It should be noted though that these studies documented that early marriage increased the risk of HIV infection. The findings from this study should be examined within the context of its limitations. This is a cross-sectional study of ANC-attending women that excludes HIV-infected women who may have experienced difficulty conceiving or who died from their infection. As well, the validity of the responses to interviews is not known and socially desirable responses may have been provided given the methods used for data collection.

Programmatic Implications

The study demonstrates the feasibility of collecting interview data as part of sero-surveillance and/or PMTCT programs. Given the high response rate, such programs should consider routinely collecting behavioral data but with greater specificity and through methods designed for greater validity.


  1. Asamoah-Odei E, Garcia Calleja JM, Boerma JT. HIV prevalence and trends in sub-Saharan Africa: no decline and large subregional differences. Lancet 2004;364:35-40.
  2. UNAIDS/WHO. AIDS Epidemic update. December 2007. UNAIDS, Geneva. In: AIDS Epidemic Updates. 2007:11.
  3. UNAIDS/WHO Working Group of Global HIV/AIDS and STI Surveillance. Guidelines for Second Generation HIV Surveillance: The Next Decade. Geneva, Switzerland: UNAIDS; 2000.
  4. Zaba B, Slaymaker E, Urassa M, et al. The role of behavioral data in HIV surveillance. AIDS 2005;19 (Suppl 2):S39-S52.
  5. Garnett GP, Garcia-Calleja JM, Rehle T, et al. Behavioural data as an adjunct to HIV surveillance data. Sex Transm Infect 2006;82(Suppl 1):i57-i62.
  6. Glynn JR, Carael M, Buve' A, et al. HIV risk in relation to marriage in areas with high prevalence of HIV infection. J Acquir Immune Defic Syndr 2003;33:526-35.