Method of Assessing Daily Intake‐Frequencies for Major Food Groups Using Dietary Screener Data

Abstract

Objectiveto develop procedures for scoring participants’ daily intake‐frequencies of food and beverage groups, and assess differences in intake‐frequencies by general health and body mass index (BMI)‐status.MethodsIn 2014, the National Cancer Institute (NCI) conducted the Family Life, Activity, Sun, Health, and Eating (FLASHE) cross‐sectional study of cancer preventive health behaviors. Parents and their adolescent children (ages 12–17 years) completed web‐based surveys which included a dietary screener, and self‐reported height and weight and general health. The 27‐item dietary screener asked about the frequency of foods eaten in the past 7 days. Items were grouped by similarities in healthfulness and type. The groups created were “junk food (JF),” “sugar sweetened beverages (SSB),” “sugary foods (SF),” “fast/convenience foods (F/C),” “fatty meats (FM),” “fruits and vegetables (FV).” To create daily intake‐frequency scores, all responses were converted to daily intake frequencies, and then summed to create an aggregate daily intake‐frequency score for each group. To adjust for potential participant overestimation, food groups were top‐coded so that no participant had a food group score that exceeded a z‐score of 3.29. Kruskal‐Wallis one‐way analysis of variance was used to describe differences in intake‐frequencies by general health and weight status.ResultsMean (SD) daily intake‐frequencies for parents (n=1,506) were: JF=1.41 (0.99), SSB=0.92 (1.16), SF=1.02 (0.83), F/C=1.11 (0.83), FM=0.58 (0.49), and FV=3.06 (1.95); adolescents (n=1,396) had values: JF=1.86 (1.21), SSB=1.29 (1.25), SF=1.42 (1.07), F/C=1.49 (1.01), FM=0.75 (0.59), and FV=2.79 (2.04). For parents and adolescents, healthy weight status (i.e., BMI 18.5 to <25 for parents; 5th to <85th percentile for adolescents) was associated with increased FV (ps=0.001 – <0.001) intake‐frequency, and decreased FM (ps<0.001) and F/C (ps<0.001–0.047) intake‐frequency. In parents only, healthy weight status was associated with decreased intake‐frequency of SSB (p<0.001). SF and JF were not significantly associated with weight status. Similar findings were seen for health status, with the exception of “good‐excellent” general health being associated with decreased adolescent intake of SSB (p=0.003).ConclusionsFindings indicate increased FV intake‐frequency and decreased FM, F/C, and SSB intake‐frequency are associated with healthy weight status and “good‐excellent” perceived health in parents and adolescents. While screeners have some inherent limitations, such as precluding estimation of energy intake and specific nutrients, they are useful at providing reasonably accurate estimates of food groups and patterns of eating behaviors. Creating summed daily intake‐frequency scores from dietary screener data offers a way to investigate patterns of intake‐frequencies for major food groups and relationships with health outcomes.Support or Funding InformationNational Cancer Institute (NCI)