Revista Brasileira de Epidemiologia
Print version ISSN 1415-790X
ENES, Carla Cristina et al. Validity and reliability of self-reported weight and height measures for the diagnoses of adolescent's nutritional status. Rev. bras. epidemiol. [online]. 2009, vol.12, n.4, pp. 627-635. ISSN 1415-790X. http://dx.doi.org/10.1590/S1415-790X2009000400012.
PURPOSE: To assess the validity and reliability of self-reported height, weight, and Body Mass Index (BMI) to diagnose the nutritional status of adolescents. METHODS: The study included 360 public school students of both genders, with ages ranging from 10 to 15 years. Adolescents self-reported their weight and height, and these values were later obtained directly by interviewers. The validity of BMI based on self-reported measures was calculated using sensitivity and specificity indexes, and positive predictive value (PPV). Agreement between self-reported and measured BMI was evaluated using Kappa's weight coefficient, the Lin correlation coefficient, and Bland-Altman and Lin's plots. RESULTS: Both girls and boys underestimated their weight (-1.0 girls and boys) and height (girls - 1.2 and boys - 0.8) (p < 0.001). BMI presented moderate agreement between measured and self-reported values. Sensitivity of estimated BMI based on reported measures to classify obese subjects was higher for boys (87.5%) than it was for girls (60.9%), whereas specificity was higher for girls (92.7%) than it was for boys (80.6%). PPV was high only for classification of normal-weight adolescents. CONCLUSIONS: Self-reported measures of weight and height in adolescents do not present valid measures; therefore, they should not be used to replace measured values. Additionally, we observed that 10% of obese boys and 40% of obese girls could have remained unidentified if we had used only self-reported measures, emphasizing the effects of the low reliability of self-reporting.
Keywords : Adolescent; Body mass index; Nutritional status; Self-reported; Height; Validation studies.