BACKGROUND

There are currently no validated methods for energy intake assessment in adolescents with intellectual and developmental disabilities (IDD). The purpose of this study was to determine the feasibility of collecting 3-day image-assisted food records (IARs) and doubly labelled water (TDEE ) data in adolescents with IDD and to obtain preliminary estimates of validity and reliability for energy intake estimated by IAR.

METHODS

Adolescents with IDD completed a 14-day assessment of mean daily energy expenditure using doubly labelled water. Participants were asked to complete 3-day IARs twice during the 14-day period. To complete the IAR, participants were asked to fill out a hard copy food record over three consecutive days (two weekdays/one weekend day) and to take before and after digital images of all foods and beverages consumed using an iPad tablet provided by the study. Energy intake from the IAR was calculated using Nutrition Data System for Research. Mean differences, intraclass correlations and Bland-Altman limits of agreement were performed.

RESULTS

Nineteen adolescents with IDD, mean age 15.1 years, n = 6 (31.6%) female and n = 6 (31.6%) ethnic/racial minorities, enrolled in the trial. Participants successfully completed their 3-day food records and self-collected doubly labelled water urine samples for 100% of required days. Images were captured for 67.4 ± 30.1% of all meals recorded at assessment 1 and 72.3 ± 29.5% at assessment 2. The energy intake measured by IAR demonstrated acceptable test-retest reliability (intraclass correlation = 0.70). On average, IAR underestimated total energy intake by -299 ± 633 kcal/day (mean per cent error = -9.6 ± 22.2%); however, there was a large amount of individual variability in differences between the IAR and TDEE (range = -1703 to 430).

CONCLUSIONS

The collection of IAR and TDEE is feasible in adolescents with IDD. While future validation studies are needed, the preliminary estimates obtained by this study suggest that in adolescents with IDD, the IAR method has acceptable reliability and may underestimate energy intake by ~9%.