PT  - JOURNAL ARTICLE
AU  - Julia Dratva
AU  - Cecilie Svanes
AU  - Ane Johannessen
AU  - Bryndis Benediktsdóttir
AU  - Randi Bertelsen
AU  - Lennart Bråbäck
AU  - Bertil Forsberg
AU  - Thorarinn Gislason
AU  - Christer Janson
AU  - Rain Jogi
AU  - Eva Lindberg
AU  - Dan Norback
AU  - Vivi Schlünssen
AU  - Trude D. Skorge
AU  - Torben Sigsgaard
AU  - Marie Waatevik
AU  - Shyamali Dharmage
AU  - Francisco Gómez Real
TI  - Validation of figural stimuli to assess metabolic status in epidemiological studies
DP  - 2014 Sep 01
TA  - European Respiratory Journal
PG  - P4085
VI  - 44
IP  - Suppl 58
4099  - http://erj.ersjournals.com/content/44/Suppl_58/P4085.short
4100  - http://erj.ersjournals.com/content/44/Suppl_58/P4085.full
SO  - Eur Respir J2014 Sep 01; 44
AB  - Background: Metabolic status plays an important role in cardio-respiratory disease development and measuring it in large surveys is a challenge. The Respiratory Health In Northern Europe III (RHINE) is the second postal follow-up study of the European Community Respiratory Health Study I (stage I). RHINE III introduced modernized figure stimuli, 9 body shapes depicting very thin (1) to obese (9) men and women, to obtain self-reported data as proxies for anthropometric and metabolic status. We validate the figural stimuli using objectively measured BMI in a subgroup of subjects aged 38-66 yrs.Methods: Self-reported information on body shapes were available for N=10828 and anthropometric data for N= 1580 (weight, height, waist and hip circumference). We performed correlation analyses, ROC curves estimating the sensitivity and specificity of the figural stimuli, and ran a validation model based on the reported body shapes and additional subject data.Results: Body silhouettes and BMI correlate highly (women 0.81, men 0.74). The mean BMI and respective IQR increased exponentially with increasing body shapes. The ROC analyses yielded very good discriminatory power for overweight (BMI >30; AUC women: 0.861/ men: 0.879). Thinness (BMI<20) and normal weight (BMI 20-25) showed poorer sensitivity and specificity for identifying thin and overweight subjects. The body image model explained 70% of the BMI variability.Conclusion: The body shapes used in RHINE are a good proxy of measured BMI. They show a high discriminatory quality for overweight, a main risk factor for many chronic diseases. They can be used with good confidence, instead of measured anthropometric data, in epidemiological questionnaire studies.