The Good Start Study was a prospective cohort study of 665 HIV positive women who attended routine PMTCT services. The women and infants were followed for 36 weeks after birth with data collection during home visits every 2 weeks until 12 weeks and then monthly until 9 months. At each scheduled visit infant feeding practices were recorded through 24 hour and previous 3 day recall. The study was undertaken in three sites; Paarl (Western Cape), Umzimkulu (Eastern Cape) and Umlazi (KwaZulu-Natal). Sites were selected to highlight differences in socio-economic regions, health infra-structure, rural-urban locations, and HIV prevalence.

Key findings:

- The quality of infant feeding counselling is poor with inadequate information provided for women to make appropriate choices; - Infant feeding counselling, as currently implemented, does not prepare women for the challenges of adhering to their infant feeding choice; - Women face new challenges in the postpartum period and ongoing community-based support and resources are important to sustain infant feeding practices; - Adherence to national feeding recommendations is low, regardless of infant feeding choice (exclusive breastfeeding or exclusive formula feeding); - Poor infant feeding practices, underlying health systems problems and high maternal viral load contribute to lower overall infant HIV-free survival; - Inappropriate infant feeding choices have important implications for infant HIV free survival.