Body mass index and body image perception in a Mexican adult population: the accuracy of self-reporting

To evaluate the accuracy of the self-reported body mass index and body image perception in a population of Mexican adults. In 1998, in the state of Morelos, Mexico, participants in the longitudinal study of Mexican Social Security Institute workers "IMSS Cohort Study" responded to a baseline, self-administered questionnaire designed to collect a large variety of lifestyle risk factors for chronic diseases. Participants self-reported their body image by selecting the silhouette that best portrayed them, from a set of silhouettes ranked from 1 to 9. Participants also self-reported their current weight and height, unaware that direct measurements of weight and height were to follow. Four to eight months later participants were weighed and their heights were measured using standardized procedures. Spearman correlations were computed to analyze the correlation between self-reported and measured data. The Kruskal-Wallis test was used to evaluate the magnitude of difference between measured and self-reported height, weight and body mass index (BMI) (measured minus self-reported), by educational level and categories of age. Robust regression was used to evaluate the potential effect of specific individual characteristics on differences between measured and self-reported weight and height. Logistic regression analysis was used to evaluate the sensitivity, specificity, and predictive values for BMI. The study included 934 subjects, of which 62.6% were female. Females had a mean measured height of 1.55 m (SD 0.06) and weight of 65.4 kg (SD 10.9), while males had a mean height of 1.67 m (SD 0.06) and weight of 77.7 kg (SD 12.5). The mean BMI was 27.4 kg/m2 (SD 4.3) (females: 27.2 kg/m2 +/- 4.45; males: 27.8+/-3.87). The median of body image perception (BIP) was 5 (25th percentile=4; 75th percentile=6). Correlations between measured and self-reported height, weight, and BMI for all subjects were 0.94, 0.96, and 0.90, respectively. The correlation between BMI and BIP was 0.64 (0.67 for females and 0.59 for males). Self-reported mean varied no more than 1.3 cm from measured height and no more than 3.17 kg from measured weight. Error estimations of height, weight, and BMI decreased with educational level. Sensitivity and specificity before adjusting self-reported BMI with overweight and obesity categories that were collapsed into one were 94.8 and 83.0%, respectively; for BIP, those values were 87.6 and 48.9%. The results suggest that self-reported BMI and BIP can be usef