Accuracy of on-line self-reported weights and heights by young adults.

Background: e-epidemiology, a convenient and low-cost research method, is becoming increasingly popular. This study seeks to validate on-line self-reported heights and weights against objectively measured data in young adults. Methods: Young adults self-reported heights and weights in an on-line lifestyle survey. These were validated using two methods: (i) measurements by staff at the primary-care clinic and (ii) measurements by a researcher within two weeks of distribution of the survey. Analyses were conducted to determine differences between the self-reported and measured heights and weights and to identify characteristics associated with under- or over-reporting of these. Results: From a total of 23 010 young adults invited to the survey, 24% provided on-line data, mean age = 19.2 (SD 3.2) years, 43% male, 91% EU citizens. Both self-reported and measured data were available for 1446 individuals (547 men, 896 women and mean age 19.2 (SD2.6) years); 1278 validated using medical records, 168 by researcher measurements. Intra-class correlations between self-reported and measured parameters were weight (r = 0.99), height (r = 0.98), with acceptable levels of agreement between measured and self-reported weight, height and BMI using Bland & Altman analyses. Self-reported weight was underestimated uniformly across BMI categories, gender and ethnicity, by a mean -0.4 (SD 0.4) kg (P < 0.001). Height was accurately reported overall across BMI and gender: both self-reported and measured heights =1.72 (SD 0.01) m, P = 0.783. Discrepancies between methods caused misclassification of BMI category for 17 (1.8%) of participants. Conclusions: Engagement of young adults with on-line research is encouraging. On-line self-reporting provides acceptably reliable anthropometric data for young adults, with under-reporting of weight by just 0.4 kg.