With epidemiological evidence indicating that breastfeeding transmits HIV to more than one-­‐third of infants of HIV-­‐positive mothers in Africa, WHO recommendations have advised that women make a choice between preventive options, i.e. early weaning after exclusive breastfeeding, or formula feeding as in developed countries. Based on field data collected in Burkina Faso and Senegal, this paper will examine how the notions of choice/incapacity have been interpretated by women. Case studies will be presented about women's experiences and the feeding patterns they applied that show how the social and medical context shape decisions and practices. These results reveal how some women build lay preventive strategies, unexpected by health professionals that permit them to face dilemma when neither breastfeeding nor " safe " formula feeding is possible. Introduction When the first data about cases of HIV transmission through breastfeeding were published in 1985 (Ziegler, 1985 ; OMS/WHO, 1987; Dunn, 1992), the eviction of breastfeeding was rapidly adopted as a medical recommendation for women living with HIV in developed countries. They were summoned to formula feed their infants, and formula would be provided to them through social insurance systems or social programs. The number of infants born from HIV-­‐positive women was known as limited and formula feeding appeared as an easy strategy, already available, known to be acceptable, quite efficient since it would fully eliminate the risk of HIV transmission. In developing countries, the history of prevention of HIV transmission through breastfeeding was more complex. There, a number of factors hindered the scaling-­‐up of the eviction of breastfeeding as a preventive strategy regarding HIV risk. The main ones were the risks related to formula feeding in settings where the sanitation level is low and drinking or potable water is scarce, where access to fuel or electricity for heating and refrigerating formula based milk is not general, where literacy rates amongst women are limited, and where the epidemiological environment makes diarrhea common and malnutrition a leading cause of infant mortality. Also, a number of scientists under-­‐estimated the impact of the AIDS epidemic at a time when it had not reached its peak in all African sub-­‐regions, and doubted about the necessity to set up specific policies (Jelliffe and Jelliffe, 1988 :142).