Weight-for-height in two national cohorts with particular reference to 10-year-old children

The idea of representing obesity or degree of malnutrition using a weight-for-height power index has existed for many years and several authors believe that such an index should be uncorrelated with height. Data from the 1958 National Child Development Study and the 1970 Child Health and Education Study have therefore been used to determine the values of the constant k which lead to the weight-for-height power index weight/[height]k being uncorrelated with height for specific age groups. Different values of k were needed both for the various age groups, and for the two sexes. For boys and girls respectively, the values of k needed at age 7 years were 2.02 and 2.12, at age 10 the values were 2.53 and 2.58, at age 11, 2.53 and 2.50 and at age 16, 2.42 and 1.71. Different values were also needed for West Indians and Asians and pubertal and pre-pubertal children. The relationships between this power index and other measurements of weight-for-height (including weight/height; weight/[height]2--the Quetelet index; weight/[height]3--the Ponderal index; relative weight for height, and standardized weight for height), the examining doctor's assessment of obesity and weight and height themselves were investigated for 10-year-old children born in 1970 to determine which of them could be thought of as best at estimating obesity. We found that there was little to choose between the index which was uncorrelated with height (using derived values of the power), and the Quetelet index.