Accuracy of using self-reported data to screen children and adolescents for overweight and obesity status: A diagnostic meta-analysis

Objective To estimate the accuracy of using the self-reported body mass index (BMIsr) for screening children and adolescents for overweight and obesity status by quantitatively synthesizing individual studies in the research literature. Method Three databases, namely PubMed, Web of Science, and EBSCOhost were searched up to September 2016. Studies were included that collected both the self-reported data (BMIsr) and direct measurement data (i.e., BMI based on measured height and weight, BMIm) to screen children and adolescents for overweight and obesity status, and provided sufficient data on sensitivity and specificity. Sensitivity, specificity, likelihood ratios, and diagnostic odds ratios from each of the included studies were pooled by using a random-effects meta-analytic model, and summary receiver operating characteristic curve (ROC) was also constructed. Results Twenty-three studies were identified. For screening children and adolescents with overweight and obesity, the use of BMIsr presented a pooled sensitivity of 0.76 (95% CI, 0.76–0.77), a pooled specificity of 0.96 (95% CI, 0.96–0.97) and a pooled DOR of 92.4 (95% CI: 74.3–114.8). Moderator analyses showed that the sample regions (America vs. Europe vs. Asia), weight status screening references (IOTF vs. CDC vs. Nation-specific standard) and weight status screened (overweight vs. obesity) had contributed to the inconsistent findings concerning the screening accuracy across the studies. Conclusions Based on the results of current meta-analysis, the use of BMIsr has good overall performance with moderate sensitivity and high specificity, and it is a viable alternative when direct measurement of BMI is not available.