VISUAL ESTIMATION OF ENERGY INTAKE DURING CAFETERIA FEEDING

Abstract

The accuracy of self-reported energy intake (EI) can be problematic, so alternative methods of assessing EI are employed. Visual estimation (VE), in which foods are visually quantified on volunteer trays has been used, but accuracy is not well documented. Total energy expenditure (TEE) and EI are regarded as equal during periods of weight stability. During U.S. Marine Corps recruit training, only foods consumed in the cafeteria are permitted, creating a unique opportunity to validate the accuracy of VE. PURPOSE To compare EI of female recruits determined using VE with TEE measurements. METHODS Energy intake of 18 female recruits (mean±SD, 19.8 ± 2.3 y, 21.5 ± 1.9 kg/m2, 23.1 ± 5.1 % body fat) was measured using VE over the first 3days of a 14-day period during which TEE was measured using doubly labeled water (2H218O). Percent body fat was measured by DEXA. Agreement, or bias, between EI measured using VE and TEE was determined by the Bland-Altman procedure, and the correlation between individual EI values derived by the two methods was calculated. RESULTS Weight remained stable throughout the TEE period (−0.1 ± 1.2 kg, range −2.3 to 1.8, p = NS for difference from 0). VE and TEE differed (p = 0.004) by 285 kcal (range −359 to 938 kcal). This difference accounted for 9.9% of TEE (range −20.6% to 32.2%, p = 0.01 for difference from 0). However, the relationship between EI values derived from TEE and VE was poor, r = 0.33, p = NS. The agreement between TEE and VE varied with energy needs: at lower energy requirements, VE may underestimate TEE by 832 kcal or overestimate by 270 kcal, whereas at high energy requirements, VE may underestimate TEE by 240 kcal or overestimate by 1342 kcal. CONCLUSION VE appears to be a valid method to measure EI in a group of individuals eating solely in a cafeteria. However, the lack of agreement at the individual level suggests that VE needs further refinement to accurately measure EI.