Nutrition support is essential to improve health outcomes among HIV-exposed children. However, there are few evidence-based models for effective integration of nutrition support within clinical Prevention-of-Mother-to-Child-Transmission (PMTCT) services. To address this gap, we aimed: 1) to describe the problem of growth faltering among HIV-exposed uninfected infants at GHESKIO clinic in Port-au-Prince, Haiti, 2) to design an age- and context-appropriate infant feeding support intervention for this population and 3) to evaluate the effectiveness of the intervention in improving growth outcomes. Medical record data from 360 HIV-exposed uninfected children seen at GHESKIO between July 2007 and October 2009 were used to describe the growth pattern from birth to 23 mo (Aim 1). They were compared to the WHO 2006 Growth Standard and a nationally-representative sample of same-age Haitian children (n=1190). GHESKIO children were born proportionally stunted and improved in weight-for-age and length-for-age across the first 6 mo of life. Faltering in weight and length began at 7 mo. Length faltering continued until 24 mo. Outcomes in GHESKIO children were significantly better than the nationally-representative sample at age 24 mo. As part of the intervention-design process (Aim 2), we identified contextual factors that influenced intervention implementation. The support of higher-level stakeholders was a key facilitator. Factors related to the chaotic environment of urban poor HIV-infected mothers were the primary barriers. The caregivers club model provides a promising foundation for integration of nutrition support and PMTCT services at GHESKIO. To assess the impact of the intervention on growth outcomes among a cohort of 6-12 month old HIV-exposed, uninfected children (Aim 3), outcomes in intervention participations (n=72) were compared to a historical control group of children in the GHESKIO PMTCT program in the previous year (n= 294). Participation in the 24-week intervention was associated with a 67.3% decrease in underweight (p=.007) and 54.7% decrease in stunting (p=.029) around age 12 mo. The magnitude of the benefit on prevalence of growth faltering was large compared to other infant feeding interventions. Further implementation research is needed to make the intervention sustainable in the GHESKIO context and to apply the model to similar contexts outside GHESKIO.