Cognitive research enhances accuracy of food frequency questionnaire reports: results of an experimental validation study

doi:10.1016/S0002-8223(02)90050-7

Journal of the American Dietetic Association

Volume 102, Issue 2, February 2002, Pages 212-225

https://doi.org/10.1016/S0002-8223(02)90050-7 Get rights and content

Abstract

Objective To test whether changing a food frequency questionnaire (FFQ) on the basis of cognitive theory and testing results in greater accuracy. Accuracy was examined for 4 design issues: a) Grouping: asking about foods in a single vs multiple separate questions; b) different forms of a food: asking consumption frequency of each form of a food (eg, skim, 2%, whole milk) vs a nesting approach—asking frequency of the main food (eg, milk) and proportion of times each form was consumed; c) additions (eg, sugar to coffee): asking independent of the main food vs nested under the main foods; d) units: asking frequency and portion size vs frequency of units (eg, cups of coffee).

Design Participants in two randomly assigned groups completed 30 consecutive daily food reports (DFRs), followed by 1 of 2 FFQs that asked about foods consumed in the past month. One was a new, cognitively-based National Cancer Institute (NCI) Diet History Questionnaire; the other was the 1992 NCI-Block Health Habits and History Questionnaire.

Subjects/setting 623 participants, age range 25 to 70 years, from metropolitan Washington, DC.

Statistical analyses performed Accuracy was assessed by comparing DFR and FFQ responses using categorical (percent agreement) and continuous (rank order correlation, discrepancy scores) agreement statistics.

Results Grouping: accuracy was greater using separate questions. Different forms of food: accuracy was greater using nesting. Additions: neither approach was consistently superior; accuracy of the addition report was affected by accuracy of the main food report. Units: both approaches were similarly accurate.

Conclusions Accuracy of FFQ reporting can be improved by restructuring questions based on cognitive theory and testing. J Am Diet Assoc. 2002;102:212–218,223–225,

Section snippets

Experimental Design

After stratified random assignment into 2 groups, study participants completed 30 days of criterion information, and then completed their respective FFQs (either the HHHQ or the test-DHQ) for the reference period of the previous month. Table 1 shows the items used to evaluate each question-design issue. Additional items were queried similarly on both FFQs (control items).

Instruments

The Daily Food Report (DFR), is a 1-page (2-sided) machine-scannable list of 88 food and drink items. Each participant filled in bubbles next to each item to indicate how many times the item was consumed that day. Items were placed on the DFR: a) if they were relevant to the 4 question-design issues under study; and b) if, based on other food intake data (USDA's Continuing Survey of Food Intakes of Individuals 1994–1996), it was expected that they would be consumed frequently enough within 30

Sample and Study Procedures

Reliably prompt mail was required to monitor daily completion of DFRs, so participants were sampled from a restricted geographical area. The sampling frame consisted of households with addresses listed in residential telephone directories in Washington, DC; Alexandria, Arlington, and Fairfax counties, Virginia; and Montgomery, Prince George's, and Frederick counties (excluding Frederick City), Md.

In August 1996 postcards were sent that provided brief information about the study and indicated

Analytical Procedures

On the DFRs, participants reported the number of times each item was consumed each day. On the FFQs, participants reported a rate of intake (ie, number of times per time period). Thus, the DFR information was continuous, whereas the FFQ information was categorical. To compare the two types of information required constructing new variables: category-converted DFR variables and continuous-converted FFQ variables.

Variables: DFR For each food, daily reported frequencies of intake on the DFR were

Statistical Analyses

With the categorical variables, three statistics were calculated to assess the accuracy of FFQ responses: the percentage of individuals whose FFQ responses agreed exactly with their category-converted DFR information, the percentage of those who never recorded consuming an item (on the DFR) and who reported never eating that item (on the FFQ), and the percentage of those who ever recorded consuming the food (on the DFR) and who reported consuming the food (on the FFQ). The continuity-adjusted χ²

Results

Of the 650 participants enrolled in the study, 623 (95.8%) returned at least 26 DFRs and the FFQ. Of these, 64% were female; 76% were white, 14% African-American, 4% Latino, and 6% other or unknown; 31% were ages 25 to 39 years, 43% ages 40 to 54, and 26% ages 55 to 70; 1% had less than high school education, 12% had no higher than high school education, and 86% had more than high school education; and 13% resided in Washington, DC, 31% in Virginia, 10% in rural Frederick, Md, and 44% in other

Comparability of HHHQ and Test-DHQ Samples

To examine comparability between participants completing the HHHQ and the test-DHQ, the demographic characteristics, FFQ completion rate, dietary intake recorded on the DFR instrument, and accuracy of FFQ reports for similarly worded foods were compared.

The HHHQ and test-DHQ groups were not significantly different in gender, age, race/ethnicity, educational level, or residence. Of the 623 participants in the study, 314 completed the HHHQ and 309 the test-DHQ, indicating that the greater length

Food Grouping: Single vs Separate Questions

Response accuracy of frequency reports for foods (Table 1) asked about in a single question vs asked about in separate questions were compared. In most but not all comparisons, the test-DHQ separated food question approach was superior to the HHHQ grouped food single question approach (Table 2).

Although there were no statistically significant differences between the two approaches in agreement for those respondents who did not consume the item, agreement for those who did consume the item was

Different Forms of Foods: Multiple Separate Questions vs Nesting

The HHHQ asks separate questions about frequency of consumption and portion size for different forms of certain foods. The test-DHQ uses a nesting approach in which questions about the frequency of consumption and portion size for the main food are asked, followed by questions about the relative proportions of time subordinate forms are consumed (Table 1).

In general, agreement statistics for the total samples indicate that frequency estimates for the main food were better with the nesting

Additions to Foods: Single Independent Question vs Nesting

The third design area examined—whether to ask about the use of an addition to food items in a single independent question (the HHHQ approach) or to nest questions about the use of the addition underneath each main food (the test-DHQ approach) (Table 1)—produced inconsistent results across the examined foods (Table 5). For milk on cereal, the test-DHQ nesting approach was somewhat better than the HHHQ single independent question approach. For salad dressing and mayonnaise, no differences between

Foods Conceptualized in Units: Frequency and Portion Size vs Frequency of Units

The last design issue examined was the utility of asking about foods in their usual serving units compared with asking about frequency and portion size (Table 1). Accuracy statistics for the two FFQs were generally similar (data not shown). Accuracy was relatively high for both instruments and for all items for nonconsumers and consumers. For example, for cups of coffee, correlations were 0.94 (HHHQ) and 0.92 (test-DHQ); percent agreement among nonconsumers was 95.3 (HHHQ) and 90.6 (test-DHQ);

Discussion

Many characteristics of FFQs likely affect reporting accuracy, including the number and nature of line items, wording, formatting, and response categories. The results of our study illustrate the complexity in the interplay of these characteristics.

The first question-design issue investigated in this study was the effect on accuracy of grouping several food items into a single question. In closed-ended dietary questionnaires like FFQs, it is impossible to ask about consumption of the entire

Applications/Conclusions

Cognitive research can reveal superior ways to ask about particular foods on an FFQ and other dietary assessment instruments. Dietitians should be aware of cognitive issues when selecting or modifying an FFQ for a particular target audience. The findings of this research were used to update and complete a version of the DHQ for public use (14).

References (15)

A.F. Subar et al.
Improving food frequency questionnairesA qualitative approach using cognitive interviewing
J Am Diet Assoc.
(1995)
S.M. Krebs-Smith et al.
Using food frequency questionnaires to estimate fruit and vegetable intakeAssociation between the number of questions and total intakes
J Nutr Ed.
(1995)
E. Rosch et al.
Basic objects in natural categories
Cognit Psychol.
(1976)
N.E. Johnson et al.
Development of a dietary intake monitoring system for nursing homes
J Am Diet Assoc.
(1982)
Dietary assessment calibration/validation register. Available at: http://riskfactor.cancer.gov/DACV. Accessed: December...
A.F. Smith
Cognitive processes in long-term dietary recall
Vital and Health Statistics.
(1991)
G.B. Willis et al.
The use of verbal report methods in the development and the testing of survey questionnaires
Appl Cognit Psychol.
(1991)

There are more references available in the full text version of this article.

Cited by (232)

The Role of Dietary Acid Load on Progression of Estimated Glomerular Filtration Rate Among Individuals Diagnosed With Chronic Kidney Disease
2024, Journal of Renal Nutrition
Individuals with chronic kidney disease (CKD) are at an increased risk for developing metabolic acidosis. Metabolic acidosis has been shown to worsen kidney function and exacerbate systemic inflammation. Diets high in protein foods can exacerbate metabolic acidosis as protein foods tend to be more acidic, while fruits and vegetables are more alkalotic. The main objective of this systematic review was to determine if higher consumption of fruits and vegetables in adults with CKD stages 1-5 reduces the rate of decline of estimated glomerular filtration rate.
Searches of Cumulated Index to Nursing and Allied Health Literature (CINAHL -Elton B. Stephens Company [EBSCO]), Cochrane Library (Wiley), Dissertation & Thesis Global (ProQuest), Embase (Elsevier), Medline (OVID), PubMed (National Library of Medicine), Scopus (Elsevier), and Web of Sciences (Clarivate) identified 1,451 articles published between January 2015 and June 2023.
After independent review, 7 total studies were included. Six of the studies found an association between dietary acid load and progression of CKD.
Dietary counseling focusing on decreasing dietary acid load may be beneficial for individuals with CKD.
Relationship Between Diet Quality and Statin Use Among Adults With Metabolic Syndrome From the CARTaGENE Cohort
2024, CJC Open
In metabolic syndrome (MetS), cardiovascular disease (CVD) risk reduction relies on the complementary use of diet and lipid-lowering medication. Evidence suggests that initiating such medication may impede diet quality. The objective of this study was to evaluate the relationship between diet quality and statin use among adults with MetS and free of CVD from the Province of Québec.
This cross-sectional study included 2481 adults with MetS (40-69 years of age) from the CARTaGENE Québec population-based cohort, of whom 463 self-reported using statin monotherapy. Diet was assessed using the Canadian Dietary History Questionnaire II, a food- frequency questionnaire, and diet quality was assessed using the Alternative Healthy Eating Index (AHEI).
In multivariable-adjusted linear regression models, statin users had lower AHEI (%) compared with nonusers (users: 40.0; 95% confidence interval [CI], 38.9, 41.2 vs nonusers: 41.2; 95% CI, 40.4, 42.0; P = 0.03] because of a lower consumption of vegetables and whole grains. Stratified interaction analyses showed that the lower diet quality among statin users was mostly prevalent among men aged ≥ 50 years and women aged ≥ 60 years, among individuals with annual household incomes of < $50,000 and persons who self-reported history of high blood pressure.
In this cohort of adults with MetS from Quebéc, the use of statin monotherapy in primary prevention of CVD was associated with a slightly lower diet quality. These data suggest suboptimal complementarity between diet quality and use of cholesterol-lowering medication in primary prevention of CVD in MetS.
Dans le syndrome métabolique, la réduction du risque de maladie cardiovasculaire repose sur la complémentarité entre une saine alimentation et l'utilisation d’hypolipidémiants. Des évidences suggèrent que l'initiation d’un traitement médicamenteux hypolipémiant influencerait négativement la qualité de l'alimentation. Cette étude avait pour objectif de déterminer la relation entre la qualité de l’alimentation et l’utilisation de statines chez des adultes avec un syndrome métabolique, mais sans maladie cardiovasculaire, au Québec.
Cette étude transversale comptait 2481 adultes avec un syndrome métabolique (âgés de 40 à 69 ans) provenant de la cohorte CARTaGENE, représentative de la population du Québec, dont 463 sujets ayant autodéclaré qu’ils prenaient une statine en monothérapie. L’alimentation des sujets a été évaluée à l’aide du Canadian Dietary History Questionnaire II, un questionnaire de fréquence alimentaire, et la qualité de l’alimentation a été évaluée à l’aide de l’Alternative Healthy Eating Index (AHEI).
Dans des modèles de régression linéaire multivariée, les utilisateurs de statines ont présenté un indice AHEI (%) plus faible en comparaison aux non-utilisateurs (utilisateurs : 40,0; intervalle de confiance [IC] à 95 % : 38,9-41,2 vs non-utilisateurs : 41,2; IC à 95 % : 40,4-42,0; p = 0,03), en raison d’une plus faible consommation de légumes et de grains entiers. Selon des analyses d’interaction stratifiées, la plus faible qualité nutritionnelle chez les utilisateurs de statines était particulièrement prévalente chez les hommes de ≥ 50 ans et les femmes de ≥ 60 ans, chez les personnes dont le revenu annuel du ménage était < 50 000 dollars et chez les personnes ayant autodéclaré des antécédents d’hypertension.
Dans cette cohorte d’adultes du Québec avec un syndrome métabolique, l’utilisation de statines en monothérapie dans la prévention des maladies cardiovasculaires était associée à une alimentation dont la qualité était légèrement plus faible. Ces données suggèrent une complémentarité sous-optimale entre la qualité de l’alimentation et l’utilisation d’hypocholestérolémiants en prévention des maladies cardiovasculaires chez les personnes avec un syndrome métabolique.
Relationship Between Diet Quality and Glucose-Lowering Medication Intensity Among Adults With Type 2 Diabetes: Results From the CARTaGENE Cohort
2024, CJC Open
In real-world settings, whether diet and medication are used as complements for glycemic management in type 2 diabetes (T2D) remains unclear. This study assessed the relationship between diet quality and intensity of glucose-lowering medication among adults with T2D.
This cross-sectional study included 352 adults with T2D from the CARTaGENE Québec population-based cohort. Diet quality was assessed using the Healthful Plant-Based Diet Index (hPDI). Glucose-lowering medication intensity was graded according to self-reported information on the type and number of drugs: no medication; oral monotherapy; oral polytherapy; and insulin with and without oral medication. In the subsample of 239 individuals who reported the medication dosages, intensity was also graded using the Medication Effect Score (MES).
In multivariable-adjusted models, we found no evidence of a relationship between the hPDI and medication intensity, assessed using the categorical approach (P_{between-group} = 0.25) or the MES (P = 0.43). However, the hPDI was inversely associated with the MES among men < 50 years of age and women < 60 years (β_{1-point MES} = –2.24 [95% confidence interval, –4.46, –0.02] hPDI points), but not among older individuals (β = –0.03 [–1.28, 1.21] hPDI points). Evidence of a nonsignificant inverse relationship between the hPDI and HbA1c was observed (β_{10-point hPDI} = –0.23% [–0.63, 0.17]), whereas a positive and significant association between the MES and hemoglobin (Hb)A1c was found (β_{1-point MES} = 0.30% [0.10, 0.51]).
In this cohort of adults with T2D, there was an overall lack of complementarity between diet quality and intensity of glucose-lowering medication. The issue was particularly important among younger adults for whom diet quality was inversely associated with intensity of medication.
Il n'a jamais été étudié, en contexte réel, si la qualité de l'alimentation et la médication sont utilisées de façon complémentaire dans la gestion de la glycémie chez des personnes avec le diabète de type 2 (DT2). La présente a évalué la relation entre la qualité de l'alimentation et l’intensité de la médication hypoglycémiante chez des adultes avec DT2.
Cette étude transversale portait sur 352 adultes avec DT2 participant à la cohorte populationnelle québécoise CARTaGENE. Nous avons évalué la qualité du régime à l’aide du Healthful Plant-based Diet Index (hPDI, soit l’indice d’un régime alimentaire à base de plantes). Nous avons évalué l’intensité des médicaments hypoglycémiants à partir des renseignements fournis sur le type et le nombre de médicaments : aucun médicament, monothérapie orale, polythérapie orale, insuline avec ou sans médicaments par voie orale. Dans le sous-échantillon de 239 individus qui ont mentionné les posologies de médicaments, nous avons aussi évalué l’intensité au moyen du Medication Effect Score (MES).
Dans les modèles multivariés, nous n’avons observé aucune évidence de relation entre le hPDI et l’intensité des médicaments au moyen de l’approche catégorielle (P_{entre les groupes} = 0,25) ou du MES (P = 0,43). Toutefois, le hPDI était inversement associé au MES chez les hommes < 50 ans et chez les femmes < 60 ans (β_{MES 1 point} = –2,24 [intervalle de confiance à 95 %, –4,46, –0,02] points hPDI), mais non chez les personnes plus âgées (β = –0,03 [–1,28, 1,21] point hPDI). Nous avons observé une relation inverse non significative entre le hPDI et l’hémoglobine (Hb)A1c (β_{hPDI 10 points} = –0,23 % [–0,63, 0,17]) et une association positive et significative entre le MES et l’HbA1c (β_{MES 1 point} = 0,30 % [0,10, 0,51]).
Au sein de cette cohorte d’adultes avec DT2, nous avons constaté une absence globale de complémentarité entre la qualité de l'alimentation et l’intensité des médicaments hypoglycémiants. Cet enjeu était d’autant plus important chez les personnes plus jeunes pour lesquels la qualité du régime était inversement associée à l’intensité des médicaments.
Efficacy of a self-management program on quality of life in colorectal cancer patients: A randomized controlled trial
2023, European Journal of Oncology Nursing
To test the efficacy of a self-management program based on acceptance and commitment therapy on quality of life, emotional distress, fatigue, physical activity, and fruit and vegetable intake in patients with colorectal cancer.
The study was a randomized controlled trial. A sample of 156 patients with colorectal cancer (stage I–III) was recruited by convenience sampling and participants were allocated randomly assigned to control or intervention groups. The intervention included a colorectal cancer self-management information booklet, two personal skills training sessions, and 12 follow-up telephone calls. The control group received health education leaflets. Outcome variables were assessed in both groups at baseline and every two months thereafter during the six-month follow-up period.
The mean age of participants was 62 years (range: 30–89 years). Generalized estimation equations analyses revealed significant differences over time in changes in anxiety (β = −2.22, p = 0.001), depression (β = −1.48, p = 0.033), fatigue (β = 4.46, p = 0.001), physical and functional measures (β = 6.16, p = 0.005), and colorectal-cancer–specific quality of life (β = 7.45, p = 0.012). However, there were no significant differences in changes in physical activity or fruit and vegetable intake over time.
The self-management skills provided by oncology nurses, including symptom management, psychological adjustment, and relaxation exercises, help colorectal cancer patients to overcome the challenges of cancer survivorship, accelerate their recovery, and improve their quality of life.
NCT03853278 registered on ClinicalTrials.gov.
Motives and food craving: Associations with frequency of hyper-palatable food intake among college students
2023, Eating Behaviors
College food environments provide wide access to hyper-palatable foods (HPF). Palatable eating motives and craving are individual factors that may impact HPF intake. Nevertheless, no study to date has examined the relationship between these factors and frequency of HPF intake in a dormitory all-you-can-eat-style cafeteria meal, a typical college food setting.
College students (N = 218) self-reported frequency of HPF intake at a college cafeteria meal and their frequency of HPF intake in the overall diet. Quasipoisson and logistic regression models tested whether palatable eating motives (assessed via the Palatable Eating Motives Scale) and craving (assessed via the Food Craving Inventory) were associated with the rate of HPF intake during the cafeteria meal and excess frequency of HPF intake in the overall diet.
Craving was associated with a 19% higher rate of HPF intake during the cafeteria meal (RR = 1.19; p = .015) and a >2-fold higher likelihood of excess frequency of HPF intake the overall diet (OR = 2.69; p < .001). Coping motive was associated with an 87% higher likelihood of excess frequency of HPF intake in the overall diet (OR = 1.87; p = .002), but not in the cafeteria meal. Reward enhancement, social and conformity motives were not significantly associated with the frequency of HPF intake in either setting.
Results highlight the role of craving and coping motive in the frequency of HPF intake among college students. Prevention efforts may consider targeting craving and eating to cope to potentially address frequency of HPF intake among college students.
DHA, nutrient intake, and maternal characteristics as predictors of pregnancy outcomes in a randomised clinical trial of DHA supplementation
2023, Clinical Nutrition
To investigate the relationships among docosahexaenoic acid (DHA) intake, nutrient intake, and maternal characteristics on pregnancy outcomes in a phase III randomised clinical trial designed to determine the effect of a DHA dose of 1000 mg/day compared to 200 mg/day on early preterm birth (<34 weeks gestation).
A secondary aim of the phase III randomised trial was to explore the relationships among pregnancy outcomes (maternal red blood cell phospholipid (RBC-PL) DHA at delivery, preterm birth, gestational age at delivery, labor type, birth anthropometric measures, low birth weight, gestational diabetes, pre-eclampsia, and admission to a neonatal intensive care unit) in participants (n = 1100). We used Bayesian multiple imputation and linear and logistic regression models to conduct an analysis of five general classes of predictor variables collected during the trial: a) DHA intake, b) nutrient intake from food and supplements, c) environmental exposure to tobacco and alcohol, d) maternal demographics, and e) maternal medical history.
DHA supplementation lowered the risk of preterm birth and NICU admission, and increased gestation and birth weight as observed in the primary analysis. Higher maternal RBC-PL-DHA at delivery was associated with DHA supplementation and formal education of a bachelor's degree or higher. DHA supplementation and maternal age were associated with a higher risk of gestational diabetes. Total vitamin A intake was associated with longer gestation, while fructose and intake of the long chain omega-6 fatty acid, arachidonic acid, were associated with shorter gestation. Risk of preterm birth was associated with a history of low birth weight, preterm birth, pre-eclampsia, and NICU admission.
Bayesian models provide a comprehensive approach to relationships among DHA intake, nutrient intake, maternal characteristics, and pregnancy outcomes. We observed previously unreported relationships between gestation duration and fructose, vitamin A, and arachidonic acid that could be the basis for future research.
ClinicalTrials.gov (NCT02626299); December 10, 2015.

View all citing articles on Scopus

View full text