This study aimed to construct a childhood obesity risk index based on predictors identified in pregnant women and 1-yr-old infants. The primary outcome was an identified obesity index of > 20% at 6-8 yr of age. Of a total sample size of 6,846 mother-child pairs, 80% and 20% were randomly allocated to the derivation and validation cohorts, respectively. For the derivation cohort, univariate and multivariate logistic regression analyses of data were conducted to identify the final predictors to determine the childhood obesity risk score algorithm. These included pre-pregnancy body mass index (BMI), child's gender, smoking during pregnancy, education, and obesity index at one yr of age. The β coefficients for categories of predictor variables were each divided by the smallest value among them. The quotient was rounded off to the integer and assigned to the risk score, and a value of zero was assigned to reference categories. A total risk score was calculated for each individual. A cutoff point ≥ 16 had 22.2% and 21.8% positive predictive values in the derivation and validation cohorts, respectively. In conclusion, the childhood obesity risk score algorithm was constructed based on generic predictors that can be easily obtained from maternal and child health handbooks.
Keywords: body mass index; childhood; obesity; pregnancy; risk score.