Self-reported weight and height for determining nutritional status of adults and elderly: validity and implications for data analysis

摘要: This study evaluated the validity of self-reported weight and height for determining nutritional status implications their use analyzing associations with health outcomes. A population-based cross-sectional in Pelotas, Rio Grande do Sul State, Brazil, 2007 (n = 2,986) drew a sub-sample 276 individuals aged > 20 years. Mean was similar to measured weight; overestimated men (1.4cm) women (2.5cm); real body mass index (BMI) underestimated by about 1kg/m2. Even small mean differences, data variability great. The results were influenced gender, age, schooling. measures prevalence overweight obesity, unpredictable errors found analysis association outcomes (underestimation, overestimation, reversal effect measures). Correction equations reduced differences but did not resolve classification errors, or biases associations.