How accurate is the prevalence of overweight and obesity in children and adolescents derived from self-reported data? A meta-analysis

Abstract

Objective: The aim of the current meta-analysis was to evaluate the accuracy of using BMI based on self-reported height and weight (BMIsr) to estimate the prevalence of overweight and obesity among children and adolescents.

Design: A systematic literature search was conducted to select studies that compared the prevalence rates of overweight and obesity based on BMIsr and BMIm (BMI based on measured height and weight). A random-effect model was assumed to estimate summary prevalence rates and prevalence ratio (PR).

Results: Thirty-seven studies were included. The aggregated prevalence of overweight and obesity based on BMIsr (0·190, 95 % CI 0·163, 0·221) was significantly lower than that based on BMIm (0·233, 95 % CI 0·203, 0·265). The pooled mean PR was 0·823 (95 % CI 0·775, 0·875). Moderator analyses showed that the underestimation was related to gender, age, weight status screened (overweight v. obesity) and weight status screening tool.

Conclusions: BMIsr may produce less biased results under some conditions than others. Future researchers using BMIsr may consider these findings and avoid the conditions that could lead to more severe underestimation of the prevalence of overweight and obesity among children and adolescents.

Keywords: Children and adolescents; Meta-analysis; Overweight and obesity; Prevalence; Self-reported BMI.

Fig. 1

Flowchart showing selection of studies for inclusion in the current meta-analysis (BMI_sr, BMI derived from self-reported height and weight; BMI_m, BMI derived from measured height and weight)

Fig. 2

Forest plots for the prevalence rates of overweight and obesity in children and adolescents based on BMI_sr (left) and BMI_m (right). The study-specific prevalence rate and 95 % CI are represented by the black square and horizontal line, respectively; the centre of the diamond represents the pooled prevalence rate and its width represents the pooled 95 % CI (BMI_sr, BMI derived from self-reported height and weight; BMI_m, BMI derived from measured height and weight; P _sr, prevalence rate based on BMI_sr; P _m, prevalence rate based on BMI_m)

Fig. 3

Quality assessment (, low; , unclear; , high) in the four main domains of QUADAS-2 for studies (n 37) included in the current meta-analysis (QUADAS-2, Revised Tool for the Quality Assessment of Diagnostic Accuracy^{( 53 )})