Convergent Validity of a Digital Image-Based Food Record to Assess Food Group Intake in Youth

Current methods to assess the dietary behavior of youth have many limitations that reduce accuracy. Previous research has examined use of food images to assess nutrient intake. The objectives for this study are to validate the use of a novel digital image-based food record (DIFR) method to assess food group intake in youth and examine inter-analyst reliability. In 2009, a convenience sample of youth aged 9 to 12 years were recruited in Davis, CA, and asked to take images of the food they ate between 5 pm and bedtime for 7 days. To examine convergent validity, 1-day and average weekly food group intakes assessed by DIFR were compared to food group estimates derived from 24-hour dietary recalls. To examine interanalyst reliability, estimates of food group intakes made by two independent nutrition students were compared, using Spearman correlation coefficients. Data from 26 youth showed that each participant's 1-day food group intakes assessed by the DIFR and recall methods were significantly correlated (P<0.001) for both analysts for all food groups. Estimated average daily intake amounts determined by the DIFR method and recall methods were also significantly correlated for all food groups except grains (n=28). Interanalyst reliability was very good; estimates of food group intakes, provided by the two students, were significantly correlated (n=28, P<0.001). These results show great potential for use of DIFR to assess 1 day's intake of food groups, but more research is needed to determine how well the method performs in capturing usual intake and changes in intake.