Preparation and validation of a predictive model of breastfeeding initiation in the first hour of life

Abstract

Objective: The objective was to develop and validate a predictive model of initiation of breastfeeding in the first hour after delivery. Methods: Retrospective cohort study on women who gave birth between 2013 and 2018 in Spain. For data collection, an ad hoc questionnaire was designed to be filled in by the mothers, which was distributed to the different Spanish breastfeeding associations which, in turn, shared it with their associate partners. The development of the predictive model was made on a cohort of 3218 women (2/3) and was validated on a cohort of 1609 women (1/3). Mothers whose children were admitted to hospital at the time of birth were excluded. A multivariate analysis was performed by means of logistic regression, and predictive ability was determined by areas under the ROC curve (AUC). Results: 81.0 % (2608) women started breastfeeding in the first hour in the derivation cohort, and 80.1 % (1289) in the validation cohort. The predictive factors in the final model were: the highest number of children and skinto-skin contact at birth as flattering factors, while dystocic delivery reduced the likelihood of the onset of breastfeeding. The predictive ability (ROC AUC) in the derivation cohort was 0.89 (CI 95 %: 0.87–0.90), while in the validation cohort it was 0.89 (CI 95 %: 0.87–0.92). Conclusions: This three-variable predictive model has excellent predictive ability in both the derivation cohort and the validation cohort. This model can identify women who are at high risk of non-initiating breastfeeding within the first hour after delivery.