Introduction This study aimed to assess the effectiveness of a child-focused early obesity prevention intervention for first-time parents in existing social networks.

Methods The Melbourne InFANT Program is a cluster-randomised controlled trial involving 542 families from 62 first-time parent groups in Melbourne, Australia (87% recruitment; 90% retention). It focuses on positive diet, physical activity and reduced sedentary behaviours from 3 to 18 months of age.

Results At intervention conclusion, children's objectively measured physical activity was similar across groups. Maternal beliefs about television and diet and corresponding child behaviours were more favourable in intervention than control participants. For example, fewer intervention group mothers believed television is helpful for development (67% vs 43%), and parents should offer other foods if their child doesn't eat their meal (44% vs 20%; both p<0.001). More intervention mothers believed parents should include fruit or vegetables in all children's meals and snacks (96% vs 91%, p=0.04), and that television should be turned off when children are eating (90% vs 81%, p=0.02). Compared to control group infants, those in the intervention group were significantly more likely to consume two or more serves of fruit (81% vs 70%, p=0.03), vegetables (56% vs 42%, p=0.04) and both fruit and vegetables daily (39% vs 26%, p<0.001) and watched an average of 92 min less television per week (β=−13.8, 95% CI −26.4 to −1.2).

Conclusions The Melbourne InFANT Program, a low-dose, low-cost obesity prevention intervention, shows promising impact on parent attitudes and beliefs and on children's diet and television viewing behaviours but not physical activity.