Abstract
Background: Despite the evidence for validity of body image pictogram (BIP) to discriminate overweight, obese, and normal individuals, there is little evidence on the probable effect of socio-demographic variables on its validity. To investigate the effects of socioeconomic status (SES), age, ethnicity, and educational level on the validity of BIP to discriminate normal weight, overweight, and obese people.
Methods: We used the Pars Cohort Study (PCS) data. Stunkard’s BIP score was used as test measure. Participants were classified as normal (body mass index [BMI]<25), overweight (BMI=25 to 29.9), and obese (BMI≥29.9) based on their BMI (kg/m2 ). Area under curve (AUC) and its 95% CI were estimated and compared. Optimal cutoff points and their sensitivity, specificity, and likelihood ratio were reported.
Results: A total of 9232 participants with a female/male ratio of 1.03 were included. The prevalence of overweight and obesity was 37.4% and 18.2%, respectively. Regardless of socio-demographic levels, the optimal cut-points to discriminate normal BMI from overweight, and overweight from obese participants were BIP score of four and five, respectively. Estimated AUC correlated with ethnicity (P<0.001) for both genders, and with SES for females (P<0.05).
Conclusion: Although BIP may be a valid measure to categorize the general adult population into normal, overweight and obese, its validity depends on SES and ethnicity. BIP may be available as a proxy measure for BMI categories in socio-demographically homogeneous populations but not in heterogeneous populations.