A randomised community intervention trial was undertaken in 12 subcentres in Karur health unit district, Tamil Nadu, to compare the efficacy of two antenatal care packages. A newly recommended 'high-risk' strategy package and a uniform package recommended by the Tamil Nadu Government were each implemented in 4 randomly selected subcentres by the study team, and the latter was also delivered by routine health services in the 4 remaining subcentres. Analyses were based on 294 pregnant women on the high-risk package (HR), 242 on the Tamil Nadu Government (TNG) package and 335 women in the control series. The HR package reduced the differences between the high-risk women and the others in mean haemoglobin and the percentages with preventable neonatal morbidity and low birth weight, and consequently the overall outcome was better in the HR series than in the TNG series. Finally, the results with the TNG package were better when it was implemented by the study team than by the routine health services, in terms of preventable maternal morbidity and preventable perinatal morbidity, but there was no impact on birth weight.