[Validation of self-reported weight and height university population and factors associated with differences between self reported and measured antropometrics]

Nutr Hosp. 2013 Sep-Oct;28(5):1633-8. doi: 10.3305/nh.2013.28.5.6671.
[Article in Spanish]

Abstract
in English, Spanish

Introduction: Obesity is an important public health problem related to a higher risk of death from chronic degenerative diseases such as diabetes, cardiovascular diseases and several types of cancer. In epidemiological studies of big sample size, only self-reported weight and height can be collected for feasibility reasons and body mass index (BMI) estimates may be questioned.

Objectives: In this study we compare self-reported and measured weight, height and BMI in a mostly young population of university students, and explore factors associated with discrepancies between self-reported and measured data.

Methods: In the period 2006-2012, 628 University students (476 women) from health sciences subject gave consent to participate in this study. Self-reported weight and height were collected by questionnaire and compared with weight and height measured afterward in health exams wearing light clothes and using standardized protocols. The validity of self-reported anthropometric estimates was explored by correlation coefficients and sensitivity, specificity, predictive values and kappa to detect measured overweight/obesity (BMI ≥ 25 kg/m²). Multiple linear regression was used to explore the factors related to the discrepancies between self-reported and measured data.

Results: The mean of self-reported weight, height and BMI was 62.5 kg, 167.6 cm and 22.1 kg/m² and the mean of measured data was 62.6 kg, 167.4 cm y 22.2 kg/m² respectively. Correlations between self-reported and measured data were r = 0.97 for weight, 0.96 for height and 0.95 for BMI. The sensitivity to detect overweight (IMC ≥ 25 kg/m²) using self-reported data was 81.0%, the specificity was 98.5%, the predictive value was 90.6% and the kappa index was 0.75. The discrepancy between measured and self-reported weight, height and BMI was associated with a higher age, and a higher sleeping time was also associated to discrepancies in self-reported and measured height.

Conclusions: Despite the self-reported weight and BMI may underestimate the true weight and BMI, and self-reported height overestimate, the validity of self-reported anthropometric measures is adequate to use be used in epidemiological studies among young people with a high educational level.

Introducción: La obesidad es un problema de salud pública importante que se asocia a un mayor riesgo muerte por enfermedades crónico-degenerativas como diabetes, enfermedades cardiovasculares y varios cánceres. En estudios epidemiológicos de amplio tamaño donde a veces solo es posible obtener datos auto-referido de peso y talla, pueden surgir dudas sobre las estimaciones del índice de masa corporal (IMC). Objetivo: En este estudio se comparan datos auto-referido de peso, talla e IMC frente a datos obtenidos por medición directa en población adulta joven, y se analizan los factores asociados a la discrepancia entre datos referidos y medidos. Metodología: Entre 2006-2012 un total de 628 universitarios de ciencias de la salud (476 mujeres) aceptaron cumplimentar un cuestionario con preguntas sobre peso y talla y realizar posteriormente un examen físico con la toma del peso y talla siguiendo protocolos estandarizados en ropa ligera. El peso de la ropa ligera se sustrajo del peso medido en 1 kg para los hombres y 0,9 kg para las mujeres. Se analizó la validez de las estimaciones antropométricas auto-referidas frente a las medidas para peso, talla y obesidad mediante índices de sensibilidad, especificidad, valores predictivos y índice kappa y se usó regresión lineal múltiple para analizar los factores asociados a las discrepancias entre datos referidos y medidos. Resultado: La media del peso, talla e IMC auto-referido fueron 62,5 kg, 167,6 cm y 22,1 kg/m2, y de los medidos, 62,6 kg, 167,4 cm y 22,2 kg/m2 respectivamente. Las correlaciones entre datos declararos y medidos fueron de r = 0,97, 0,96 y 0,95 respectivamente. La sensibilidad para detectar exceso de peso (IMC ≥?25 kg/m2) mediante datos declarados fue del 81,0%, la especificidad del 98,5%, el valor predictivo positivo 90,6% y el índice kappa de 0,75. La discrepancia entre peso, talla e IMC medido y declarado se asoció significativamente con una mayor edad, y para la talla también con un mayor número de horas de sueño. Conclusiones: A pesar de una ligera infraestimación observada para el peso y el IMC y una sobreestimación para la talla auto-referidos, la validez de las medidas auto-referidas es adecuada para usar en estudios epidemiológicos en población joven.

Publication types

  • Comparative Study
  • Research Support, Non-U.S. Gov't
  • Validation Study

MeSH terms

  • Adolescent
  • Adult
  • Body Height*
  • Body Mass Index*
  • Body Weight*
  • Diagnostic Self Evaluation*
  • Female
  • Humans
  • Male
  • Physical Examination*
  • Prospective Studies
  • Self Report*
  • Students
  • Universities
  • Young Adult