Background.The Community Mothers Programme aims at using experienced volunteer mothers in disadvantaged areas to give support to first-time parents in rearing their children up to 1 year of age. The programme was evaluated by randomized controlled trial in 1990. Method.Seven years later, trial participants were interviewed about child health, nutrition, cognitive stimulation, parenting skills, and maternal self-esteem. The aim of this study was to see whether the demonstrated benefits at 1 year of age of this programme could be sustained at age 8. Results.One-third of the original group (38 intervention, 38 control), were contacted and interviewed. The risk for having an accident requiring a hospital visit was lower in the intervention group: relative risk (RR) 0.59, 95 per cent confidence interval (CI) 0.31-1.11. Intervention children were more likely to visit the library weekly: RR 1.58, 95 per cent CI 1.10-2.26. Intervention mothers were more likely to check homework every night: RR 1.23, 95 per cent CI 1.05-1.43 (p=0.006); and to disagree with the statement 'children should be smacked for persistently bad behaviour': RR 2.11, 95 per cent CI 1.10-4.06. They were more likely to disagree with the statement 'I do not have much to be proud of': RR 1.24, 95 per cent CI 1.04-1.40; and to make a positive statement about motherhood than controls: RR 1.53, 95 per cent CI 1.06-2.20. Subsequent children of intervention mothers were more likely to have completed Haemophilus influenzae b: RR 1.26, 95 per cent CI 1.06-1.51; and polio immunization: RR 1.19, 95 per cent CI 1.02-1.40. Conclusion.The Community Mothers programme had sustained beneficial effects on parenting skills and maternal self-esteem 7 years later with benefit extending to subsequent children.