The lifetime risk of acquiring HIV infection in many rural as well as urban areas of southern Africa is currently as high as two-in-three. For women, much of this risk still accrues rapidly at young ages despite high levels of knowledge about HIV/AIDS. Thus, programmes that are more participatory and address underlying structural and community-level factors appear to be essential. We use cross-sectional data from a large-scale, population-based survey in rural eastern Zimbabwe to describe the relationships between membership of different forms of community group and young women's chances of avoiding HIV. Our results show that participation in local community groups is often positively associated with successful avoidance of HIV, which, in turn, is positively associated with psychosocial determinants of safer behaviour. However, whether or not these relationships hold depends on a range of factors that include how well the group functions, the purpose of the group, and the education level of the individual participant. We identify factors that may influence the social capital value of community groups in relation to HIV prevention at the individual, group, and community levels. Young women with secondary education participate disproportionately in well-functioning community groups and are more likely to avoid HIV when they do participate. Longitudinal studies are needed: (i) to establish whether community group membership supports the development of safer lifestyles or merely has greater appeal to individuals already predisposed towards such lifestyles, and (ii) to pinpoint directions of causality between hypothesised mediating factors. In-depth research is needed on the specific qualities of community groups that enhance their contribution to HIV control. However, our findings suggest that promotion of and organisational development and training among community groups could well be an effective HIV control strategy.

HIV-related sub-fertility has been reported for those populations in sub-Saharan Africa in which contraceptive use is low. We use data from a retrospective survey in rural Zimbabwe and multivariate logistic regression models to show that recent birth rates and current pregnancy rates are also lower among HIV-positive women than among HIV-negative women in those African populations where contraceptive use is high. The fertility reduction is smaller than where contraceptive use is low because age at first sexual intercourse is later and birth rates at older ages are already low. Nevertheless, total fertility is approximately 8.5 per cent lower and HIV-associated sub-fertility may account for as much as one-quarter of fertility decline in Zimbabwe since the late 1980s. Mechanisms for HIV-associated sub-fertility in rural Zimbabwe include more frequent widowhood and divorce, reduced coital frequency, increased amenorrhoea, and possibly, pelvic inflammatory disease. Miscarriage appears to be a less important factor than elsewhere possibly because syphilis is rare.

ANC estimates understate female HIV prevalence in this low fertility population but, here, the primary cause is not selection of pregnant women. ANC estimate adjustment procedures that control for contraceptive use and age at first sex are needed.