1 National Food Institute, Technical University of Denmark2 Division of Nutrition, National Food Institute, Technical University of Denmark3 Department of Applied Mathematics and Computer Science, Technical University of Denmark4 Statistics and Data Analysis, Department of Applied Mathematics and Computer Science, Technical University of Denmark5 University of Oslo
Background and aim The potential health effects of a New Nordic Diet (NND) are to be tested in the Danish OPUS (Optimal well-being, development and health for Danish children through a healthy New Nordic Diet) School Meal study among 8-11- year-old school-children. Valid and reliable dietary assessment methods are essential for identifying how eating habits may change in response to the intervention and for identifying the impact of the children’s dietary habits on their health and weight status. Several challenges are connected to collecting dietary data from children including their cognitive ability and social desirability which in addition is influenced by the OPUS study. Furthermore, they are untrained in the task and they may not be involved in food shopping or preparation and therefore have little insight into the foods they eat. The overall aim of the present project was to deliver a validated and suitable dietary assessment tool that could be used by 8-11-year-old Danish school children to assess dietary intake in the OPUS School Meal Study. The specific objectives were to develop a dietary assessment tool suitable for 8-11-year-old children feasible to be used in the OPUS School Meal Study (Paper 1), to validate the developed dietary assessment tool by a combination of validation methods in order to obtain information about the reporting accuracy including the acceptability, under-reporting and over-reporting, and repeatability of WebDASC (Paper 2), and the overall- and lunch specific reporting accuracy of fruit, juice and vegetables (Paper 3). Materials and methods The development of a Web-based Dietary Assessment Software for 8-11-year-old Children (WebDASC) followed a prototyping approach: Considerations about factors connected to the OPUS study aim of relevance to the dietary assessment, data level, available resources, and input from professionals, focus groups, literature review, and usability tests preceded its release. Special consideration was given to age-appropriate design issues. In the validation study, which were conducted as a part of the OPUS School Meal pilot study, 81 school children 8-11-year-old, assisted by parents, recorded their diet in the WebDASC and wore an accelerometer on the same 7 consecutive days twice: at baseline with the habitual diet, and at intervention with the NND. On the same 5 school days as they reported their diet in WebDASC the children’s school lunch was photographed and weighed before and after lunch. During the week after the baseline food- and activity recording fasting blood samples were taken. The acceptability of WebDASC was assessed with a questionnaire. Energy intake (EI) estimated with WebDASC was evaluated against accelerometer-estimated energy expenditure using Bland-Altman plot, correlation and Kappa statistics. The repeatability of EI was assessed using Intraclass correlation coefficient. Furthermore, the accuracy of self-reported fruit, juice and vegetable (FJV) intake was evaluated by comparing intake to plasma carotenoids concentration using correlations and Kappa statistics. Intervention effect, weekday and meal effect in FJV intake, and effects of background factors were assessed using Linear Mixed Models. Finally, the accuracy of reporting FJV intake at school lunch were measured by scoring the reported intake in WebDASC against FJV actually eaten observed by the digital photos as either matches, intrusion, omission and faults. Main findings WebDASC was developed as an intuitive, cost-effective, and engaging method to collect detailed dietary data from 8-11-year-old children. Results from the acceptability questionnaire demonstrated that it was well accepted among children and adults. Results from the validation study showed that on group level reported EI was in agreement with total energy expenditure (TEE). However, 20% was classified as under-reporters and 20% as over-reporters. Mis-reporting was associated to weight status and a higher body mass index (BMI) characterized under-reporters, and a lower BMI characterized over-reporters compared to acceptable-reporters. The repeatability of EI was fair. Reporting that illness affected eating influenced reported EI and FJV intake. The WebDASC estimated intake of FJV was significantly correlated with carotenoid plasma concentration, and Spearman and Partial correlation coefficients adjusted for gender, BMI, and TEE showed correlations of 0.58 and 0.49 respectively (p<0.01). Fruits and juice showed higher correlations than vegetables with plasma carotenoid concentration. The results from the photographic observations of school lunch demonstrated that WebDASC attained 82% reporting matches overall and higher percent match for reporting fruits compared to beverages. Intrusions (reporting of FJV not eaten or reporting too large portion size) were the most common reporting mistake (90%). Among intrusion it was more common to report fruit and vegetables not eaten (65%) than reporting a portion size image illustrating a larger portion than the eaten portion size (35%). Conclusions and implications The thesis demonstrated that it is possible to develop a child appealing web-based dietary assessment tool that can be used at home on the family’s home computer. The developed WebDASC was acceptable to use for both 8-11-year-old children and their parents, and feasible to use in the OPUS School Meal pilot study. The WebDASC provides good estimates of average energy intake compared to the estimated total energy expenditure. Moreover a moderate repeatability of EI was observed. The ability of the WebDASC to rank participants according to energy intake was fair. The validation study demonstrated that under-reporting and over-reporting was associated to the weight status and BMI of the children. Possible causes may be the weight and health focus of the study, social desirability and the diet reporting itself. When using plasma carotenoid concentrations as a reference, the WebDASC’s ability to rank participants according to FJV intake was good and the WebDASC obtained a high percent matches for FJV intake and overall intake at school lunch. In conclusion the WebDASC is both acceptable and feasible to use to collect dietary data from 8-11-year-old children in intervention studies. This project demonstrated that, in the study population, data could be used to estimate energy intake on group level and to rank individuals according to EI, and to rank FJV intake both overall and on school meal level, and thereby contribute to the understanding about associations of fruit and vegetable intake, which is an important nutritional indicator for healthy eating habits, and the development of lifestyle diseases.
Main Research Area:
Frost Andersen, Lene, Brockhoff, Per B., Tetens, Inge, Trolle, Ellen