Validity of self-reported anthropometrics in Arabic-speaking adults in Australia.

Abstract

Body mass index (BMI) is generally accepted as a useful measurement for monitoring risk factors in adults. Although self-reported anthropometric measurements are deemed to be more cost-effective, its accuracy has been debated. While BMI based on self-reported measures may have to be relied on, accuracy of reporting such measures among culturally and linguistically diverse groups is unknown. Face-to-face surveys were conducted among 272 adults of Arabic-speaking backgrounds living in south western Sydney using non-probability sampling to collect data for directly measured and self-reported BMI. Agreement between both measures was determined by the Cohen's kappa coefficient analyses. The Wilcoxon matched-pairs signed-ranks and Mann-Whitney U tests were used to compare the differences in median values between both measures. The Bland-Altman analysis was conducted to identify the limits of agreement between both measurements. There was substantial agreement between both self-reported and directly measured data (kappa = 0.70). Significant small median differences were found between both direct and self-reported overall BMI measure (27.58 vs 27.34; P<.0001) with a significantly greater median difference for females compared to males (0.76 vs 0.38kg/m2 ; P=.05). However, the 95% limits of agreement were moderately large for BMI (-5.1 and 6.4). Self-reported data for height and weight are generally appropriate for calculating BMI for health promotion interventions among adults from Arabic-speaking backgrounds but should be used cautiously when assessing BMI status at the individual level. SO WHAT?: When limited resources are available for accomplishing health promotion interventions, self-reporting measures may be used as a proxy for assessing BMI.