Current dietary assessment methods are affected by memory-related bias and heavily burden respondents. To reduce respondent burden and associated measurement error, we recently developed an innovative smartphone-based 2-hour recall (2hR) method. Evaluation of the use of random 2hRs, administered via the smartphone-based dietary assessment tool 'Traqq', for assessing habitual intake against a validated FFQ and blood concentration markers. Dietary intake was assessed in 215 Dutch adults by repeated 2hRs on randomly selected days and times (i.e., equivalent to three full days of 2hRs) over a four-week period. At the end of the study period, participants completed a validated semi-quantitative FFQ. Sixty-five random participants also provided two fasting blood samples, to assess plasma carotenoid and plasma n-3 PUFA concentrations. Intake estimates of energy (2,132±665 kcal vs. 2,017±572 kcal) were slightly higher with 2hRs than FFQ, while the percentage energy intake from macronutrients where similar (i.e., protein: 37±14 en% vs. 37±13 en%; fat: 35±6 en% vs. 44±7 en%; carbohydrates: 44±7 en% vs. 44±6 en%). For food groups, a larger variation in intake estimates was found ranging from -65% (legumes) to 62% (pastry, cake and biscuits). Spearman correlations between 2hRs and FFQ ranged from 0.33 to 0.69 for energy and macronutrients, from 0.32 to 0.58 for micronutrients, and from 0.27 to 0.67 for food groups. For all nutrients and food groups, ≥70% of the participants were classified in the same/adjacent quartile. Spearman correlations between 2hRs and plasma carotenoids and n-3 PUFA ranged between 0.34 and 0.57 and cross-classification ranged between 61% to 83% in the same/adjacent quartile. Comparing 2hRs with FFQ and blood concentration markers showed good ranking ability for energy, most nutrients, and most frequently consumed foods. More variation was seen for episodically consumed foods and nutrients.
Read full abstract