Abstract
Body mass index (BMI) is frequently used as assessment of nutritional and health risk. Yet, there is no consensus regarding assessment of height and cutpoints for weight classification in older adults. We first investigated differences in height by self-report (SR-height) and height calculated from measured knee-height (KH-height) and derived BMI in a cross-sectional assessment of 145 older adults residing in a Midwestern United States city. We further compared the proportion of older adults classified as at risk by seven different BMI cutpoints for weight classification as reported in the literature. We found no substantive difference in SR-height and KH-height or derived BMI, but we did find significant differences in the proportions of participants classified as At Risk across different BMI cutpoints. In community and clinical practice, SR-height and KH-height, and BMI derived from these, can be used interchangeably, but the lack of consensus regarding BMI cutpoints could lead to misclassification of older adults at risk.
Notes
Note. We classified participants by BMI derived from self-reported height (SR-height) and height calculated from measured knee-height (KH-height), and by the seven different cutpoints with references to the literature. Variable names used in our analyses for these seven are shown in the third column.
Note. Variables with the same superscript differ significantly at p < 0.001.
Note. Means in the same row that do not share superscripts differ at p < .05 in the Scheffé post hoc test.
∗∗p ≤ .01.
Note. Above the diagonal are probability values for Knee Height (KH-height) Tests of Differences; below the diagonal Self Report (SR-height) Tests of Differences. References for BMI1-7 cutpoints are in Table .
∗p ≤ .05; ∗∗p ≤ .01; ∗∗∗p ≤ .001.