There is limited information on whether integrating childhood interventions with health and nutrition services interventions is effective and feasible. In this trial we used group delivery at five routine visits from age 3-18 months, and comprised: short films of child development messages, shown in the waiting area; discussion and demonstration led by community health workers; and mothers' practice of activities. Nurses gave out and reviewed message cards with mothers, together with a few play materials. A cluster randomized trial was conducted in the Caribbean (Jamaica, Antigua and St Lucia) in 29 health centers. Centers were randomized to control (n=15) or health center intervention (n=14). We also adapted the Jamaica home visit intervention to increase feasibility at scale. Primary outcomes were child cognition, language and hand-eye coordination, and secondary outcomes caregiver knowledge, practices, maternal depression, and child growth, measured after the 18 month visit. Multilevel analyses comparing health center only with control in all 3 countries showed significant benefits for cognitive development from the health center intervention with effect size of 0.3 SD and benefits to parenting knowledge with effect size 0·4. In analyses of the two interventions in Jamaica, both benefited cognitive development with effect sizes of 0.34 SD (home visit) and 0.38 SD (health center). The most conservative analyses found benefit cost ratios of 5.3 for the health center intervention and 3.8 for home visits. Integrating parenting interventions into health services has the potential to reach a large number of children with benefits substantially higher than required investments.