This cluster-randomised proof-of-concept trial was conducted in Dagoretti, Nairobi. 50 community health volunteers (CHVs) were selected and assigned, randomly, to intervention or control arm (1:1). 101 households (one caregiver/household) under the CHV’s catchment (2-3 per CHV), with at least one child aged 6-24 months, were selected for participation. Caregivers in the intervention group (n=50) received a CHV-delivered food hygiene intervention and those in the control group (n-51) received no intervention. Blinding was not possible. Follow-up was four weeks post intervention delivery. At baseline and endline food hygiene behaviours were recorded. The primary outcome was the proportion of caregivers observed to practice five pre-specified food hygiene behaviours (handwashing with soap before preparing the child's food, handwashing with soap before feeding the child, washing the child hands with soap before they eat, heating food to boiling, and storing feeding utensils in a clean sealed container or cleaning feeding utensils before use) . Secondary outcomes assessed these five observed behaviours individually as well as a sixth behaviour - the proportion of caregivers who report always boiling the child’s drinking water.