Nutritional status and mortality: a prospective validation of the QUAC stick

In December 1970, 8,292 rural Bengali children between the ages of 1 and 9 had their height and arm circumference measured. Eighteen months later the fate of 98.8% of these children was ascertained. Overall, 2.3% of the children had died. Those below the 9th and between the 10th and 50th percentiles of arm circumference for height were at 3.4 and 1.5 times greater risk of dying, respectively, than those above the 50th percentiles. A gradient was present at every age, although it was greatest for the vulnerable 1 to 4 yr age group, for whom the relative risks were 4.5, 1.6, and 1.0, respectively. The discriminant efficiency of these categories was greatest immediately following measurement and decreased with time. During the first postmeasurement month the risk of dying in the poorest nutritional category was 19.8 times that of the best, and for the first 3 mth, 12.2 times. By the last 3 mth of followup it was only twice that of the best. Females in all three categories fared slightly worse than males, being at 1.1 times the risk of dying. This same vulnerable group of 1 to 4 yr olds could be identified without knowing their age. Limiting the analysis to children whose heights were between 65 and 89 cm resulted in relative risks, for the three categories, of 4.1, 1.6, and 1.0, respectively. These arm circumference to height categories and the Quaker Arm Circumference (QUAC) stick survey technique for which they were devised appear to be valid tools for identifying nutritionally disadvantaged individuals and populations at high risk of death.