Parental overweight as an indicator of childhood overweight: how sensitive?

Abstract

This study determined the sensitivity and specificity of parental overweight from self-reported height and weight to identify families with overweight school age children. A cross sectional study was conducted among 3059 parents and their children (1558 boys and 1501 girls) aged 7-12 years in five primary schools of Busan, Korea. BMI was calculated from parental reported height and weight and from children's measured height and weight. Parents were considered overweight when their BMI was >25 kg/m2 (WHO, 2000). Children were considered overweight when their BMI was >95th percentile (CDC, 2000). Prevalence of overweight was calculated and logistic regressions were performed. The sensitivity and specificity of parental overweight were calculated. A total of 26% (805/3059) parents were overweight. Of the families with one overweight parent, 15% (N = 109) had an overweight child. When both parents were overweight, 17% (N = 9) had an overweight child. After adjusting for child's age and gender, parental education, family income, and spouse's BMI as required, the odds of having an overweight child were 2.5 [1.8, 3.3] for one overweight parent, and 3.2 [1.4, 7.1] for both overweight parents. While the sensitivity of one overweight parent to identify families with overweight school age children was 44%, specificity was 75%. The presence of both overweight parents provided a 3% sensitivity and 98% specificity for the identification of an overweight school age child. Although parental overweight was obtained from self-reported weight and height in Busan (Korea), it is a practical indicator to identify families with an overweight school age child, it has poor sensitivity.