[Performance of the Breastfeeding Assessment Score for the prediction of early weaning in France].

Background: The Breastfeeding Assessment Score (BAS) was derived to help identify mothers at increased risk of early weaning in United States. Data are currently lacking on the accuracy of the BAS for French mother-infant pairs.

Objective: To assess the accuracy of the BAS in a French validation cohort.

Methods: We used the original data from a prospective cohort study of 488 mothers who were breastfeeding at discharge in 9 maternity wards in 2005. The outcome measures were assessed using structured follow-up telephone interviews at 4 and 26 weeks.

Results: The weaning rate was 3% at 14 days of infant age. The corresponding area under ROC curve was 0.73 [0.60-0.85] and was comparable to that observed in the derivation cohort (0.75). For a cut point of 8 recommended by the authors of the BAS, 43% of mother-infant pairs were categorized at high risk and the weaning rate in this subgroup was 5%. The mother-infant pairs with a score lower than 8 had a shorter median breastfeeding duration (18 versus 20 weeks, P=0.02), were more likely to report breastfeeding difficulties after discharge (63% versus 53%, P=0.03), and were less likely to be "very satisfied" with breastfeeding experience (66% versus 77%, P=0.007).

Conclusion: The intrinsic properties of the BAS are robust. However, its use would be of limited interest in France because of the relatively low rate of early weaning. Randomized trials are needed before recommending routine use of BAS-based breastfeeding support intervention.