Fourth-grade children's dietary recall accuracy is influenced by retention interval (target period and interview time)

Abstract

Background: For a 24-hour dietary recall, two possible target periods are the prior 24 hours (24 hours immediately preceding the interview time) and previous day (midnight to midnight of the day before the interview), and three possible interview times are morning, afternoon, and evening. Target period and interview time determine the retention interval (elapsed time between to-be-reported meals and the interview), which, along with intervening meals, can influence reporting accuracy.

Objective: The effects of target period and interview time on children's accuracy for reporting school meals during 24-hour dietary recalls were investigated. DESIGN AND SUBJECTS/SETTING: During the 2004-2005, 2005-2006, and 2006-2007 school years in Columbia, SC, each of 374 randomly selected fourth-grade children (96% African American) was observed eating two consecutive school meals (breakfast and lunch) and interviewed to obtain a 24-hour dietary recall using one of six conditions defined by crossing two target periods with three interview times. Each condition had 62 or 64 children (half boys).

Main outcome measures: Accuracy for reporting school meals was quantified by calculating rates for omissions (food items observed eaten but unreported) and intrusions (food items reported eaten but unobserved); a measure of total inaccuracy combined errors for reporting food items and amounts.

Statistical analyses performed: For each accuracy measure, analysis of variance was conducted with target period, interview time, their interaction, sex, interviewer, and school year in the model.

Results: There was a target-period effect and a target-period by interview-time interaction on omission rates, intrusion rates, and total inaccuracy (six P values <0.004). For prior-24-hour recalls compared to previous-day recalls, and for prior-24-hour recalls in the afternoon and evening compared to previous-day recalls in the afternoon and evening, omission rates were better by one third, intrusion rates were better by one half, and total inaccuracy was better by one third.

Conclusions: To enhance children's dietary recall accuracy, target periods and interview times that minimize the retention interval should be chosen.

Figure 1. Illustration of intake covered in a 24-hour dietary recall using each of six conditions defined by crossing two target periods with three interview times

The prior-24-hour target period consists of the 24 hours immediately preceding the interview time. The previous-day target period is midnight to midnight of the day before the interview. In this figure: the solid line inside cells indicates intake to be covered in that 24-hour dietary recall, “B” indicates when a breakfast that is to be reported eaten in the 24-hour dietary recall is eaten, “L” indicates when a lunch that is to be reported eaten in the 24-hour dietary recall is eaten, “D” indicates when a dinner that is to be reported eaten in the 24-hour dietary recall is eaten, “xB” indicates when an intervening breakfast occurs between the to-be-reported breakfast and the interview, “xL” indicates when an intervening lunch occurs between the to-be-reported lunch and the interview, “xD” indicates when an intervening dinner occurs between the to-be-reported dinner and the interview, and “Int” indicates when an interview to obtain a 24-hour dietary recall is conducted.

Figure 2

Multiple-pass interview protocols patterned after the Nutrition Data System for Research (44) and used to obtain 24-hour dietary recalls from fourth-grade children during the 2004–05, 2005–06, and 2006–07 school years in (city), (state)

Figure 2

Multiple-pass interview protocols patterned after the Nutrition Data System for Research (44) and used to obtain 24-hour dietary recalls from fourth-grade children during the 2004–05, 2005–06, and 2006–07 school years in (city), (state)

Figure 2

Multiple-pass interview protocols patterned after the Nutrition Data System for Research (44) and used to obtain 24-hour dietary recalls from fourth-grade children during the 2004–05, 2005–06, and 2006–07 school years in (city), (state)

Figure 3. Information, by two target periods and six conditions, ^a about omission rates, ^b intrusion rates, ^c and total inaccuracy^d for the school-meal (breakfast and lunch) parts of fourth-grade children’s 24-hour dietary recalls validated with school-meal observations during the 2004–05, 2005–06, and 2006–07 school years in (city), (state)

^a The six conditions created by crossing two target periods (prior 24 hours, previous day) with three interview times (morning, afternoon, evening) are defined and illustrated in Figure 1. Each section of this figure shows the mean and standard deviation above the bar for each target period and each condition; each sample size is shown in parentheses below the label for each bar. Accuracy was assessed for only the school-meal parts of 24-hour dietary recalls because only school meals were observed. Each item observed eaten and/or reported eaten at a school meal was weighted by meal component (see Table 1, footnote b) and scored as a match, omission, or intrusion (see Table 1, footnote e). ^b Omission rate is the percentage of items observed eaten that was not reported eaten. For an interviewed child, omission rate = (sum of weighted omissions/[sum of weighted omissions + sum of weighted matches]) × 100%. Values may range from 0% (indicating no omissions) to 100% (indicating that no items observed eaten were reported eaten). Lower omission rates indicate better reporting accuracy. Fifteen children (6 boys) had perfect omission rates of 0% (12 prior-24-hour recalls [1 morning, 4 afternoon, 7 evening]; 3 previous-day recalls [all morning]). Analysis of variance (ANOVA) showed an effect of target period on omission rates (P<0.0001), which were better for prior-24-hour recalls than previous-day recalls. Also, ANOVA showed a target-period by interview-time interaction on omission rates (P=0.001); pairwise comparisons showed that omission rates were better for prior-24-hour recalls in the afternoon and evening than previous-day recalls in the morning, afternoon, and evening (six P values <0.00093). ^c Intrusion rate is the percentage of items reported eaten that was not observed eaten. For an interviewed child, intrusion rate = (sum of weighted intrusions/[sum of weighted intrusions + sum of weighted matches]) × 100%. Values were undefined for 39 children (21 boys) whose recalls failed to meet criteria for both school breakfast and school lunch (see Table 1, footnote b). Defined values for intrusion rate may range from 0% (indicating no intrusions) to 100% (indicating that no items reported eaten were observed eaten). Lower intrusion rates indicate better reporting accuracy. Forty-two children (23 boys) had perfect intrusion rates of 0% (33 prior-24-hour recalls [8 morning, 15 afternoon, 10 evening]; 9 previous-day recalls [3 morning, 2 afternoon, 4 evening]). ANOVA showed an effect of target period on intrusion rates (P<0.0001), which were better for prior-24-hour recalls than previous-day recalls. Also, ANOVA showed a target-period by interview-time interaction on intrusion rates (P<0.0001); pairwise comparisons showed that intrusion rates were better for prior-24-hour recalls in the morning and evening than previous-day recalls in the afternoon and evening; for prior-24-hour recalls in the afternoon than previous-day recalls in the morning, afternoon, and evening; and for previous-day recalls in the morning than previous-day recalls in the afternoon and evening (nine P values <0.0003). ^d Total inaccuracy, for an interviewed child, was the sum of three components: a) the sum, over matches, of the absolute difference between amounts observed eaten and reported eaten for each match times the weight; b) the sum, over intrusions, of each intruded amount times the weight; and c) the sum, over omissions, of each omitted amount times the weight. This single measure cumulates errors (in servings) for all food items whether they are matches, omissions, or intrusions. It provides a composite score for both food items and amounts, but does not indicate whether errors are due to omissions, intrusions, or incorrect amounts for matches. Lower total inaccuracy values indicate better reporting accuracy. ANOVA showed an effect of target period on total inaccuracy (P<0.0001) which was better for prior-24-hour recalls than previous-day recalls. Also, ANOVA showed a target-period by interview-time interaction on total inaccuracy (P=0.004); pairwise comparisons showed that total inaccuracy was better for prior-24-hour recalls in the morning than previous-day recalls in the afternoon and evening; for prior-24-hour recalls in the afternoon and evening than previous-day recalls in the morning, afternoon, and evening; and for previous-day recalls in the morning than previous-day recalls in the evening (nine P values <0.0024).

Figure 3. Information, by two target periods and six conditions, ^a about omission rates, ^b intrusion rates, ^c and total inaccuracy^d for the school-meal (breakfast and lunch) parts of fourth-grade children’s 24-hour dietary recalls validated with school-meal observations during the 2004–05, 2005–06, and 2006–07 school years in (city), (state)

^a The six conditions created by crossing two target periods (prior 24 hours, previous day) with three interview times (morning, afternoon, evening) are defined and illustrated in Figure 1. Each section of this figure shows the mean and standard deviation above the bar for each target period and each condition; each sample size is shown in parentheses below the label for each bar. Accuracy was assessed for only the school-meal parts of 24-hour dietary recalls because only school meals were observed. Each item observed eaten and/or reported eaten at a school meal was weighted by meal component (see Table 1, footnote b) and scored as a match, omission, or intrusion (see Table 1, footnote e). ^b Omission rate is the percentage of items observed eaten that was not reported eaten. For an interviewed child, omission rate = (sum of weighted omissions/[sum of weighted omissions + sum of weighted matches]) × 100%. Values may range from 0% (indicating no omissions) to 100% (indicating that no items observed eaten were reported eaten). Lower omission rates indicate better reporting accuracy. Fifteen children (6 boys) had perfect omission rates of 0% (12 prior-24-hour recalls [1 morning, 4 afternoon, 7 evening]; 3 previous-day recalls [all morning]). Analysis of variance (ANOVA) showed an effect of target period on omission rates (P<0.0001), which were better for prior-24-hour recalls than previous-day recalls. Also, ANOVA showed a target-period by interview-time interaction on omission rates (P=0.001); pairwise comparisons showed that omission rates were better for prior-24-hour recalls in the afternoon and evening than previous-day recalls in the morning, afternoon, and evening (six P values <0.00093). ^c Intrusion rate is the percentage of items reported eaten that was not observed eaten. For an interviewed child, intrusion rate = (sum of weighted intrusions/[sum of weighted intrusions + sum of weighted matches]) × 100%. Values were undefined for 39 children (21 boys) whose recalls failed to meet criteria for both school breakfast and school lunch (see Table 1, footnote b). Defined values for intrusion rate may range from 0% (indicating no intrusions) to 100% (indicating that no items reported eaten were observed eaten). Lower intrusion rates indicate better reporting accuracy. Forty-two children (23 boys) had perfect intrusion rates of 0% (33 prior-24-hour recalls [8 morning, 15 afternoon, 10 evening]; 9 previous-day recalls [3 morning, 2 afternoon, 4 evening]). ANOVA showed an effect of target period on intrusion rates (P<0.0001), which were better for prior-24-hour recalls than previous-day recalls. Also, ANOVA showed a target-period by interview-time interaction on intrusion rates (P<0.0001); pairwise comparisons showed that intrusion rates were better for prior-24-hour recalls in the morning and evening than previous-day recalls in the afternoon and evening; for prior-24-hour recalls in the afternoon than previous-day recalls in the morning, afternoon, and evening; and for previous-day recalls in the morning than previous-day recalls in the afternoon and evening (nine P values <0.0003). ^d Total inaccuracy, for an interviewed child, was the sum of three components: a) the sum, over matches, of the absolute difference between amounts observed eaten and reported eaten for each match times the weight; b) the sum, over intrusions, of each intruded amount times the weight; and c) the sum, over omissions, of each omitted amount times the weight. This single measure cumulates errors (in servings) for all food items whether they are matches, omissions, or intrusions. It provides a composite score for both food items and amounts, but does not indicate whether errors are due to omissions, intrusions, or incorrect amounts for matches. Lower total inaccuracy values indicate better reporting accuracy. ANOVA showed an effect of target period on total inaccuracy (P<0.0001) which was better for prior-24-hour recalls than previous-day recalls. Also, ANOVA showed a target-period by interview-time interaction on total inaccuracy (P=0.004); pairwise comparisons showed that total inaccuracy was better for prior-24-hour recalls in the morning than previous-day recalls in the afternoon and evening; for prior-24-hour recalls in the afternoon and evening than previous-day recalls in the morning, afternoon, and evening; and for previous-day recalls in the morning than previous-day recalls in the evening (nine P values <0.0024).

Figure 3. Information, by two target periods and six conditions, ^a about omission rates, ^b intrusion rates, ^c and total inaccuracy^d for the school-meal (breakfast and lunch) parts of fourth-grade children’s 24-hour dietary recalls validated with school-meal observations during the 2004–05, 2005–06, and 2006–07 school years in (city), (state)

^a The six conditions created by crossing two target periods (prior 24 hours, previous day) with three interview times (morning, afternoon, evening) are defined and illustrated in Figure 1. Each section of this figure shows the mean and standard deviation above the bar for each target period and each condition; each sample size is shown in parentheses below the label for each bar. Accuracy was assessed for only the school-meal parts of 24-hour dietary recalls because only school meals were observed. Each item observed eaten and/or reported eaten at a school meal was weighted by meal component (see Table 1, footnote b) and scored as a match, omission, or intrusion (see Table 1, footnote e). ^b Omission rate is the percentage of items observed eaten that was not reported eaten. For an interviewed child, omission rate = (sum of weighted omissions/[sum of weighted omissions + sum of weighted matches]) × 100%. Values may range from 0% (indicating no omissions) to 100% (indicating that no items observed eaten were reported eaten). Lower omission rates indicate better reporting accuracy. Fifteen children (6 boys) had perfect omission rates of 0% (12 prior-24-hour recalls [1 morning, 4 afternoon, 7 evening]; 3 previous-day recalls [all morning]). Analysis of variance (ANOVA) showed an effect of target period on omission rates (P<0.0001), which were better for prior-24-hour recalls than previous-day recalls. Also, ANOVA showed a target-period by interview-time interaction on omission rates (P=0.001); pairwise comparisons showed that omission rates were better for prior-24-hour recalls in the afternoon and evening than previous-day recalls in the morning, afternoon, and evening (six P values <0.00093). ^c Intrusion rate is the percentage of items reported eaten that was not observed eaten. For an interviewed child, intrusion rate = (sum of weighted intrusions/[sum of weighted intrusions + sum of weighted matches]) × 100%. Values were undefined for 39 children (21 boys) whose recalls failed to meet criteria for both school breakfast and school lunch (see Table 1, footnote b). Defined values for intrusion rate may range from 0% (indicating no intrusions) to 100% (indicating that no items reported eaten were observed eaten). Lower intrusion rates indicate better reporting accuracy. Forty-two children (23 boys) had perfect intrusion rates of 0% (33 prior-24-hour recalls [8 morning, 15 afternoon, 10 evening]; 9 previous-day recalls [3 morning, 2 afternoon, 4 evening]). ANOVA showed an effect of target period on intrusion rates (P<0.0001), which were better for prior-24-hour recalls than previous-day recalls. Also, ANOVA showed a target-period by interview-time interaction on intrusion rates (P<0.0001); pairwise comparisons showed that intrusion rates were better for prior-24-hour recalls in the morning and evening than previous-day recalls in the afternoon and evening; for prior-24-hour recalls in the afternoon than previous-day recalls in the morning, afternoon, and evening; and for previous-day recalls in the morning than previous-day recalls in the afternoon and evening (nine P values <0.0003). ^d Total inaccuracy, for an interviewed child, was the sum of three components: a) the sum, over matches, of the absolute difference between amounts observed eaten and reported eaten for each match times the weight; b) the sum, over intrusions, of each intruded amount times the weight; and c) the sum, over omissions, of each omitted amount times the weight. This single measure cumulates errors (in servings) for all food items whether they are matches, omissions, or intrusions. It provides a composite score for both food items and amounts, but does not indicate whether errors are due to omissions, intrusions, or incorrect amounts for matches. Lower total inaccuracy values indicate better reporting accuracy. ANOVA showed an effect of target period on total inaccuracy (P<0.0001) which was better for prior-24-hour recalls than previous-day recalls. Also, ANOVA showed a target-period by interview-time interaction on total inaccuracy (P=0.004); pairwise comparisons showed that total inaccuracy was better for prior-24-hour recalls in the morning than previous-day recalls in the afternoon and evening; for prior-24-hour recalls in the afternoon and evening than previous-day recalls in the morning, afternoon, and evening; and for previous-day recalls in the morning than previous-day recalls in the evening (nine P values <0.0024).