Conventional analyses of data from dietary validation studies may misestimate reporting accuracy: illustration from a study of the effect of interview modality on children's reporting accuracy

Abstract

Objective: To compare two approaches to analysing energy- and nutrient-converted data from dietary validation (and relative validation) studies - conventional analyses, in which the accuracy of reported items is not ascertained, and reporting-error-sensitive analyses, in which reported items are classified as matches (items actually eaten) or intrusions (items not actually eaten), and reported amounts are classified as corresponding or overreported.

Design: Subjects were observed eating school breakfast and lunch, and interviewed that evening about that day's intake. For conventional analyses, reference and reported information were converted to energy and macronutrients; then t-tests, correlation coefficients and report rates (reported/reference) were calculated. For reporting error-sensitive analyses, reported items were classified as matches or intrusions, reported amounts were classified as corresponding or overreported, and correspondence rates (corresponding amount/reference amount) and inflation ratios (overreported amount/reference amount) were calculated.

Subjects: Sixty-nine fourth-grade children (35 girls) from 10 elementary schools in Georgia (USA).

Results: For energy and each macronutrient, conventional analyses found that reported amounts were significantly less than reference amounts (every P < 0.021; paired t-tests); correlations between reported and reference amounts exceeded 0.52 (every P < 0.001); and median report rates ranged from 76% to 95%. Analyses sensitive to reporting errors found median correspondence rates between 67% and 79%, and that median inflation ratios, which ranged from 7% to 17%, differed significantly from 0 (every P < 0.0001; sign tests).

Conclusions: Conventional analyses of energy and nutrient data from dietary reporting validation (and relative validation) studies may overestimate accuracy and mask the complexity of dietary reporting error.

Figure 1

Dietary reporting accuracy in a validation (or relative validation) study should be assessed by evaluating the congruence between reported and reference information. In this approach, reported and reference items are classified as intrusions (reported items not in the reference set), matches (reported items in the reference set), and omissions (reference set items not reported). Amounts—of servings or of energy and nutrients—are classified as overreported (reported but not in the reference information), corresponding (reported and in the reference information), and unreported (in the reference information, but not reported).