O19 Measuring the Reliability of a Frequency Method for Assessing Vegetable Intake Using Photos: A Smart Phone Approach

Traditional dietary assessment tools can present several challenges, including participant and researcher burden, recall bias, cost, and time. A frequency method using mobile photos and descriptions of meals through smartphones can be a promising strategy to address these challenges. To measure the interrater reliability of assessing the frequency of vegetable intake using photos and descriptions from smartphones. This study used a pre-posttest experimental design. Participants were undergraduate students living in dormitories at a large Midwestern university. Participants (n = 85) were asked to count the number of times they ate red/orange vegetables and set a goal to eat one more time. Participants used their smartphones to upload photos and descriptions of their meals on an online platform for three days. Based on the study's objective, two raters independently coded meals using uploaded photos and descriptions of meals from smartphones. The first rater, who has several months of experience in public health research, was trained in qualitative coding of mobile photos by a senior researcher. The second rater has over three years of experience in coding mobile photos and oversaw the coding process. The primary outcome measure was the interrater reliability in assessing the number of times each of these vegetable subgroups was consumed daily: dark green vegetables, beans and peas, starchy vegetables, and other vegetables. Cohen κ was calculated to determine interrater reliability. A value of κ = 0.956 (P < 0.001) was obtained, indicating an almost perfect agreement between the two raters who independently coded each mobile photo using the same training protocols. The trained raters reliably coded the frequency of vegetable consumption using mobile photos. Therefore, a frequency method using mobile photos and descriptions of meals through smartphones is a reliable strategy to assess vegetable consumption by nutrition researchers. This mobile-phone-assisted method can be used in nutrition programs to improve data quality, reduce participant burden, and minimize recall bias. Women and Hi-Tech organization.