Estimating overweight risk in childhood from predictors during infancy

Area of Science:

• Pediatrics
• Public Health
• Epidemiology

Background:

• Childhood overweight and obesity are significant public health concerns.
• Early identification of at-risk infants is crucial for effective prevention strategies.

Purpose of the Study:

• To develop and validate a risk score algorithm for predicting childhood overweight in infants.
• To identify key predictors associated with overweight at age 3 years.

Main Methods:

• Utilized the UK Millennium Cohort Study, randomly dividing the sample for algorithm derivation (80%) and validation (20%).
• Employed stepwise logistic regression to build a prediction model for childhood overweight using International Obesity Task Force criteria.
• Calculated predictive metrics including R(2), area under the receiver operating curve (AUROC), sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV).

Main Results:

• Seven significant predictors for childhood overweight were identified: gender, birth weight, weight gain, maternal pre-pregnancy BMI, paternal BMI, maternal smoking in pregnancy, and breastfeeding status.
• The risk score ranged from 0 to 59, corresponding to predicted risks from 4.1% to 73.8%.
• The model demonstrated moderately good predictive ability in both derivation (AUROC = 0.721) and validation (AUROC = 0.755) cohorts.

Conclusions:

• A prediction algorithm can effectively identify infants at risk of developing overweight.
• Targeted prevention efforts by health professionals can be enhanced by using this risk score.
• Further research is needed to assess the clinical validity, feasibility, and acceptability of implementing this risk communication strategy.