HIV Prevalence Comparison Between Antenatal Sentinel Surveillance and Demographic and Health Survey in Rwanda§

OBJECTIVE To compare HIV prevalence from antenatal surveillance to that of the demographic and health survey (DHS), and to identify factors determining the difference of HIV prevalence between women recruited in these two surveys in Rwanda in 2005. METHODS Comparative cross-sectional study of HIV prevalence and socio-demographic factors collected by the antenatal survey in 13,745 pregnant women, seen in 30 health centres located throughout the country and those collected by the DHS among 5641 women, aged 15-49 years living in households located throughout the country. Log-binomial regression and direct standardization were used to estimate and compare HIV prevalence between the two surveys. RESULTS HIV prevalence in the antenatal survey was slightly higher than that in DHS women (4.1% versus 3.6% p=0.103). Socio-demographic characteristics were differently distributed between the two populations. Whereas, 59%, 93%, 53% of pregnant women were aged 20-29 years, married or cohabiting and living in rural areas respectively, the corresponding proportions among DHS women were 35%, 48% and 83% (p<0.001). Simultaneous standardization of antenatal prevalence according to the distribution of socio-demographic characteristics in the DHS gave an overall HIV prevalence estimate of 3.6%, similar to the prevalence measured among DHS women. CONCLUSIONS HIV prevalence in the antenatal survey overestimated that among women of the general population in Rwanda in 2005. This overestimation could be corrected by standardization of antenatal prevalence according to the distribution of age, geographical area, marital status, parity, and education, in the general population.