A new statistical method for estimating the usual intake of episodically consumed foods with application to their distribution

OBJECTIVE: We propose a new statistical method that uses information from two 24-hour recalls to estimate usual intake of episodically consumed foods. STATISTICAL ANALYSES PERFORMED: The method developed at the National Cancer Institute (NCI) accommodates the large number of nonconsumption days that occur with foods by separating the probability of consumption from the consumption-day amount, using a two-part model. Covariates, such as sex, age, race, or information from a food frequency questionnaire, may supplement the information from two or more 24-hour recalls using correlated mixed model regression. The model allows for correlation between the probability of consuming a food on a single day and the consumption-day amount. Percentiles of the distribution of usual intake are computed from the estimated model parameters. RESULTS: The Eating at America's Table Study data are used to illustrate the method to estimate the distribution of usual intake for whole grains and dark-green vegetables for men and women and the distribution of usual intakes of whole grains by educational level among men. A simulation study indicates that the NCI method leads to substantial improvement over existing methods for estimating the distribution of usual intake of foods. CONCLUSIONS: The NCI method provides distinct advantages over previously proposed methods by accounting for the correlation between probability of consumption and amount consumed and by incorporating covariate information. Researchers interested in estimating the distribution of usual intakes of foods for a population or subpopulation are advised to work with a statistician and incorporate the NCI method in analyses.     

Statistical methods for estimating usual intake of nutrients and foods: a review of the theory

Although 24-hour recalls are frequently used in dietary assessment, intake on a single day is a poor estimator of long-term usual intake. Statistical modeling mitigates this limitation more effectively than averaging multiple 24-hour recalls per respondent. In this article, we describe the statistical theory that underlies the four major modeling methods developed to date, then review the strengths and limitations of each method. We focus on the problem of estimating the distribution of usual intake for a population from 24-hour recall data, giving special attention to the problems inherent in modeling usual intake for foods or food groups that a proportion of the population does not consume every day (ie, episodically consumed foods). All four statistical methods share a common framework. Differences between the methods arise from different assumptions about the measurement characteristics of 24-hour recalls and from the fact that more recently developed methods build upon their predecessor(s). These differences can result in estimated usual intake distributions that differ from one another. We also demonstrate the need for an improved method for estimating usual intake distributions for episodically consumed foods.     

Comparison of intakes of US Chinese women based on food frequency and 24-hour recall data

The objective of this cross-sectional study was to compare dietary reports from a food frequency questionnaire (FFQ) for US Chinese women with 24-hour recall estimates. The subjects were 56 women recruited through organizations in Philadelphia's Chinese community. Spearman correlations were used to describe FFQ estimates of food servings per month and nutrient intake per day vs estimates from three 24-hour recalls over 1 month. On average, women reported at least weekly consumption of 28 of 96 FFQ food items. The three most frequently consumed were rice (38 times/month), tea (29 times/month), and dark green, leafy vegetables (18 times/month). Comparing reported frequencies of the 28 foods to 24-hour recall estimates, the median Spearman correlation was 0.36. For nutrient estimates, correlations were high (r >0.5) for dietary fiber and calcium; moderate ( r =0.25 to 0.5) for energy, saturated fat, cholesterol, carbohydrates, protein, folic acid, and iron; but poor (r <0.25) for total fat, vitamin C, vitamin A, and carotene. These findings provide some assurance of the FFQ's adequacy for describing US Chinese women's intake of commonly consumed foods and selected nutrients. They also provide a basis for further improvements to, and evaluations of, the FFQ.     

Assessing the accuracy of a food frequency questionnaire for estimating usual intake of phytoestrogens

The primary objective of this study was to assess the accuracy of a modified Block food frequency questionnaire (FFQ) with respect to its ability to assess usual dietary intakes of daidzein and genistein. Participants were a convenience sample of 51 Japanese and 18 Caucasian women. All interviews were conducted between February 1997 and October 1997. At each of the four study visits, participants provided a 24-hour urine specimen and a 48-hour dietary recall. At the first visit, participants also completed an interviewer-administered modified Block FFQ. The daidzein and genistein intakes estimated using the FFQ were moderately correlated with the mean estimates of daidzein and genistein intake calculated from four 48-hour dietary recalls (correlation for daidzein = 0.49-0.58 and correlation for genistein = 0.45-0.54) and estimates of urinary concentrations of these compounds calculated from four collections (correlations for daidzein and genistein = 0.49 and 0.30, respectively). The accuracy of the modified Block FFQ for assessment of usual daidzein and genistein intakes is supported by this study. These results support the use of this instrument in epidemiological studies as an easy and low-cost method to assess the usual dietary daidzein or genistein intake.     

Validation of a food frequency questionnaire for assessment of calcium and bone-related nutrient intake in rural populations

OBJECTIVE: To assess the ability of a semiquantitative food frequency questionnaire (FFQ) to measure calcium and bone-related nutrient intakes in a rural South Dakota population. DESIGN: Intake estimates from FFQ were compared with four 24-hour recalls obtained quarterly during the preceding year. SUBJECTS: Convenience sample of 100 participants of the South Dakota Rural Bone Health Study were recruited, with 81 completing the FFQ. MAIN OUTCOME MEASURES: Calcium and bone-related nutrient intakes were expressed as milligrams per day, milligrams per 1,000 kcal, or quartiles. STATISTICAL ANALYSES PERFORMED: Intakes by FFQ and 24-hour recalls were compared using paired t test and quartiles were formed to examine cross-classification. RESULTS: Calcium intakes from FFQ and recalls were 1,287 and 1,141 mg/day (P=0.01), but calcium per 1,000 kcal did not differ. Calcium intake by FFQ correlated with intake by recall when expressed as milligrams per day (r=0.49, P<0.001) or milligrams per 1,000 kcal (r=0.59, P<0.001). Bland-Altman graphs indicated fairly good agreement between methods. Seventy-eight percent of subjects fell into the same or within one quartile category when calcium intake was expressed as milligrams per day and 83% when expressed as milligrams per 1,000 kcal. Gross misclassification occurred in 0% to 4% of the nutrients. CONCLUSIONS: Although FFQ may not be a valid indicator of an individual's intake, it does adequately classify rural populations into quartiles of calcium and bone-related nutrient intakes, making it a useful tool for assessing dietary calcium and bone related intake in rural populations.     

Validity and reliability of a quantitative food frequency questionnaire measuring n-3 fatty acid intakes in cardiac patients in the Midwest: a validation pilot study

The purpose of this pilot study was to test the validity and reliability of a quantitative n-3 fatty acid food frequency questionnaire (FFQ) for later use with larger groups of individuals. A convenience sample of heart patients provided dietary data via three 24-hour food recalls and FFQs. Participants were women (n=17) and men (n=11), 43 to 77 years of age. The association of mean daily intake of n-3 fatty acids obtained using food recalls and the FFQ was assessed by Pearson correlation. The reliability of the FFQ was assessed using coefficient alpha. Correlation of n-3 fatty acid intake using the food recalls and the FFQ was r=0.42 (P<0.05). The coefficient alpha for the test-retest of the FFQ was .83. The top two foods, walnuts and flaxseed, contributed 58% of the n-3 fatty acid intake, and the third food, salmon, contributed 5%. This quantitative n-3 FFQ is a valid instrument for use in place of food recalls for estimating n-3 fatty acid intakes in heart patients and is a reliable instrument to estimate n-3 fatty acid intakes from plant, animal, and seafood sources. The FFQ should be tested in a larger population. Registered dietitians can use this FFQ to screen for intakes, educate patients on food sources, and measure change in intakes after nutrition intervention.     

A 7-item versus 31-item food frequency questionnaire for measuring fruit, juice, and vegetable intake among a predominantly African-American population

OBJECTIVE: This study sought to determine which of 2 fruit and vegetable food frequency questionnaires (FFQs) most closely approximated intake measured by the average of four 24-hour dietary recalls. DESIGN: Participants completed either a 31-item FFQ (n = 70) or 7-item FFQ (n = 76) on 2 occasions approximately 2 weeks apart. During the interval between FFQs1 participants provided four 24-hour dietary recalls via telephone interview. SUBJECTS/SETTING: Participants were 146 persons with food preparation responsibilities in families of students in grades 3 through 5. Respondents were predominantly African-American women in Atlanta, Ga. STATISTICAL ANALYSIS: Pearson correlation coefficients of log-transformed values estimated the reliability of each FFQ and compared FFQ estimates to reference values. The intraclass correlation coefficient evaluated consistency across 24-hour recalls. RESULTS: The first FFQs overestimated intake approximately twofold. The 31-item FFQ estimates exceeded 7-item FFQ estimates by approximately 30% . Correlations with recall estimates were high for the 7-item FFQ and moderate to low for the 31-item FFQ. The second FFQ estimates were more highly correlated to reference values. From the first to the second administration, 7-item FFQ estimates dropped from 5.2 to 2.7 servings, and 31-item FFQ estimates dropped from 6.7 to 3.5 servings. Neither FFQ produced highly reliable estimates. CONCLUSIONS: Mean total fruit and vegetable consumption was closer to reference estimates for the first 7-item FFQ and the second 31-item FFQ. The 7-item FFQ correlated more highly with reference estimates than did the 31-item FFQ. Therefore, we conclude that for African-American adults, a 1-time-administered FFQ using 7 broad food categories correlates more highly with reference values than a FFQ using 31 individual fruit and vegetable items.     

The CARDIA dietary history: development, implementation, and evaluation

To meet the objectives for dietary assessment in the Coronary Artery Risk Development in Young Adults (CARDIA) prospective study, we developed a dietary history to provide accurate and reliable quantitative data on habitual individual nutrient intakes at baseline. The CARDIA dietary history was an interviewer-administered method that included a short questionnaire regarding general dietary practices followed by a comprehensive food frequency questionnaire about typical intake of foods using the previous month as a reference for recall. For each broad category of foods, participants were questioned in detail about specific foods only if they indicated that they consumed foods from that category. Follow-up questions for selected foods concerned serving size, frequency of consumption, and common additions to these foods. Provision was made for reporting foods not found in the food frequency list. The interview took approximately 45 minutes. Cue cards prompted responses and plastic food models assisted in estimating usual amounts consumed. A precoded format standardized coding for reported items and established the detail needed for recall during the interview. Baseline nutrient analyses from the CARDIA dietary history provided estimates that agreed reasonably well with expected caloric intake for body mass index according to the age- and sex-specific Recommended Dietary Allowances, but were higher than those reported from 24-hour recalls for comparable age, sex, and race groups in the second National Health and Nutrition Examination Survey. The CARDIA dietary history is a comprehensive assessment tool that can provide a dietitian with detailed information regarding habitual eating patterns and nutrient intakes among adults.     

Validity and reproducibility of the food frequency questionnaire used in the Shanghai Women's Health Study

To evaluate the validity and reliability of the food frequency questionnaire (FFQ) used in the Shanghai Women's Health Study (SWHS), 200 SWHS participants were recruited for a dietary calibration study. Study participants completed an FFQ at baseline and 24-h dietary recalls (24-HDR) twice per month consecutively for 12 months. At the end of the study, a second FFQ was administered. Of the 200 study participants, 196 completed 24 or more days of 24-h dietary recalls, 191 completed two FFQs from whom the results of this report were based. The FFQ included the foods that accounted for 86% of the foods recorded in the 24-HDR surveys. Validity of the FFQ was evaluated by comparing intake levels of major nutrients and foods obtained from the second FFQ with those derived from the multiple 24-HDR. The median intake for major nutrients, rice, poultry and meat derived from the second FFQ and the 24-HDR was similar, with the differences ranging from 1.3 to 12.1%. The FFQ tended to overestimate the intake level of total vegetables and total fruits, and the differences were explained mainly by over-reporting seasonal vegetables and fruits consumption in the FFQ. Nutrient and food intake assessed by the FFQ and the multiple 24-HDR correlated very well, with the correlation coefficients being 0.59-0.66 for macronutrients, 0.41-0.59 for micronutrients, and 0.41-0.66 for major food groups. The reliability of the FFQ was assessed by comparing the correlation and median intake of nutrients and food groups obtained from the two FFQs that were administered approximately 2 y apart. The median intake levels for selected nutrients and food groups derived from the two FFQs were similar with differences below 10%. At the individual level, the intake levels of these dietary variables obtained from two FFQs also correlated well. When nutrient and food group intakes were categorized into quartiles, FFQ and 24-HDR produced exact agreement rates between 33 and 50%. Misclassification to adjacent quartile was common, ranging from 34-48%, while misclassification to an extreme quartile was rare (1-6%). These data indicate that the SWHS FFQ can reliably and accurately measure usual intake of major nutrients and food groups among women in Shanghai.     

Validation of a simplified food frequency questionnaire as used in the Nutrition and Health Survey in Taiwan (NAHSIT) for the elderly

A 28-item simplified food frequency questionnaire (SFFQ) combined with 9 open questions about basic commodities was designed for the Elderly Nutrient and Health Survey in Taiwan (NAHSIT) to collect participants' usual intake of the previous month. We have examined the validity this SFFQ as assessed by multiple 24-hour dietary recall (n=81) and biomarkers (n=1473). All questionnaires were completed by face-to-face interview and fasting blood taken. Thirty seven males and 44 females were randomly selected from NAHSIT participants. Of these, 31 and 50 subjects completed 2 or 3 24-hour dietary recalls within one month, respectively. Mean daily intake frequencies for each food group were calculated from the SFFQ and 24-hr recalls, respectively. Spearman rank correlation coefficients between frequencies of food intake from the FFQ and dietary recalls ranged from 0.132 to 0.678 for men; 0.052 to 0.759 for women. For food weight, the trends were similar to frequency findings. For biomarkers, the most correlated were dairy intake frequency judged by vitamin B-2 and calcium intakes and by erythrocyte glutathione reductase (EGRAC) for B-2 functionality, where the correlation coefficients were, respectively, 0.533, 0.518 and -0.205 for men; 0.494, 0.475 and -0.174 for women; fish and fruit followed in overall validity. The SFFQ measured the food patterns of elderly NAHSIT elders with validity high for dairy and good for fish, and fruit intakes in both genders.     

Vegetable and fruit food frequency questionnaire serves as a proxy for quantified intake

BACKGROUND: Public health practitioners need valid tools to survey trends in dietary intake. The Rapid Risk Factor Surveillance System (RRFSS) includes an optional six-item vegetable and fruit intake food frequency questionnaire (FFQ) module. Our objectives were 1) to compare reported vegetable and fruit consumption from the FFQ to quantified servings (portions) defined by Canada's Food Guide to Healthy Eating and ascertained by a reference method, and 2) to compare the FFQ with the reference method for their classification of the proportion of respondents consuming five or more servings of vegetables and fruit per day. METHODS: Dietitians administered 24-hour recalls to each of 174 adult respondents who had completed the FFQ as part of the RRFSS. Recalls were conducted over the telephone on three separate occasions using an adaptation of the multiple pass method. RESULTS: The mean total intake of vegetables and fruit for the group was 4.6 times/day from the FFQ versus 4.8 servings/day from the recalls (paired t-test; p = 0.92). Thirty-seven percent of respondents were classified as consuming five or more times/day by the FFQ versus 35% by the 24-hour recall servings. CONCLUSION: The FFQ tool can be used as a proxy for quantified intake of vegetable and fruit consumption.     

Dietary assessment instruments for preschool children: reliability of parental responses to the 24-hour recall and a food frequency questionnaire

Nutrient intakes of preschool-age children were assessed with a 24-hour dietary recall and a 3-month food frequency questionnaire (FFQ). Parents of 55 preschoolers (mean age = 4.25 +/- 0.59 years) completed the recall and FFQ on two occasions 1 week apart. The recalls and FFQs were analyzed for energy, cholesterol, protein, total carbohydrate, calcium, sodium, potassium, and saturated, polyunsaturated, and monounsaturated fats; each nutrient was expressed as raw value, value per kg body weight, and value per 1,000 kcal. Test-retest reliability estimates for the 24-hour recall indicated significant variability in reported total energy intake, but stable reports of intake were observed for one or all units of expression for polyunsaturated fats, cholesterol, protein, total carbohydrate, calcium, and potassium. The FFQ showed significant positive test-retest reliability estimates for all nutrients for all units of expression. Comparison of the recall and FFQ data showed similar percentages of intakes of energy from fat, carbohydrate, and protein and significant correlations for reported intakes of cholesterol, protein, calcium, and potassium. Comparison of the recall and FFQ data with recall data from a comparable cohort showed lower reported intakes for our sample, with the exception of protein, carbohydrate, calcium, and potassium. If validation studies are successful, the FFQ may be useful in epidemiological studies of preschoolers' intakes over extended periods. The recall may prove to be a useful tool in the assessment of day-to-day variations in macronutrient intakes.     

Validity of a food frequency questionnaire and a food diary in a short-term recall situation

The purpose of this study was to compare intakes of eight nutrients computed from a food frequency questionnaire (FFQ) and from 3-day food diaries with known values. Total food intake of 19 healthy adult volunteers (10 men and 9 women) was monitored for 2 weeks. Each week, subjects completed a 3-day diary and a 1-week FFQ. Subjects had previously completed one diary and one FFQ with respect to their usual diets. FFQs and diaries were analyzed for calories, protein, fat, cholesterol, calcium, iron, vitamin A, and vitamin C. All nutrients were underestimated by the FFQ (all p less than .05). Subjects omitted foods from the FFQ and incorrectly reported consumption frequencies. Vitamin A (p less than .001) and calories (p less than .05) were underestimated by the food diary. It is concluded that factors such as frequency of eating, pairing of foods, and explicitness of FFQ items affect recall accuracy. Those factors could influence the choice of FFQ design and administration.     

Fruit and vegetable assessment: performance of 2 new short instruments and a food frequency questionnaire

Objective To evaluate the ability of 2 new short assessment instruments and a food frequency questionnaire (FFQ) to measure intake of fruit and vegetables. The “All-Day” screener asks frequency and portion size questions about 9 food items. The “By-Meal” screener is similar, except that it asks about 2 of those 9 food items in terms of mealtime.

Design Survey participants completed 4 telephone-administered 24-hour dietary recalls over 1 year, a self-administered FFQ 1 to 2 months later, and 1 of 2 self-administered screeners after an additional 7 months.

Subjects/setting Participating were 202 men and 260 women aged 20 to 70 years living throughout the United States.

Statistical analyses Fruit and vegetable intakes measured by each screener and the FFQ were compared with true usual intake based on a measurement error model with 24-hour dietary recalls as the reference instrument.

Results Estimates of median daily servings of fruit and vegetables were as follows: For men: True intake (5.8) vs All-Day screener (5.0), By-Meal screener (5.5), and FFQ (6.6); for women: true intake (4.2) vs All-Day screener (5.0), By-Meal screener (5.4), and FFQ (6.2). Estimated correlations between the test instruments and true intake were as follows: For men: All-Day screener (0.66), By-Meal screener (0.67), FFQ (0.68); for women: All-Day screener (0.51), By-Meal screener (0.53), and FFQ (0.54).

Applications/conclusions Both screeners might be useful to estimate median intakes of fruit and vegetable servings in US populations, but they might be less useful in accurately ranking individuals. More research is needed before using the screeners in ethnic or low-literacy populations. J Am Diet Assoc. 2002;102:1764-1772.     

Validation of three food frequency questionnaires and 24-hour recalls with serum carotenoid levels in a sample of African-American adults

The validity of self-reported fruit and vegetable intake in minority populations has not been adequately established. In this study, the authors examined the association of three food frequency questionnaires (FFQs) and 24-hour dietary recalls with serum carotenoid levels. Approximately 1,000 African-American adults recruited from 15 churches in Atlanta, Georgia (1997-1998) completed three fruit and vegetable FFQs: a seven-item instrument assessing intake during the past month; a two-item measure assessing usual intake; and a 36-item measure adapted from the Health Habits and History Questionnaire. A total of 414 participants received a 24-hour recall by telephone, and 105 of them received two additional recalls. Serum levels of lycopene, lutein, cryptoxanthin, alpha-carotene, and beta-carotene were assessed in 813 participants and used as the validity criterion. The correlations of fruit and vegetable servings with specific and total serum carotenoid levels were generally higher for the 36-item FFQ than for the two-item and seven-item instruments. The strongest correlation of fruit and vegetable servings with total carotenoid levels was observed for the three recalls (r = 0.42), with the 36-item FFQ and the single 24-hour recall yielding comparable correlations (r = 0.35 and r = 0.37, respectively). The validity of the 36-item fruit and vegetable FFQ was generally as strong as the validity of both 1 and 3 days of recalls. Given the lower cost and time needed for administration relative to recalls, it appears that the 36-item FFQ has merit for evaluating fruit and vegetable health interventions.     

Taking advantage of the strengths of 2 different dietary assessment instruments to improve intake estimates for nutritional epidemiology

With the advent of Internet-based 24-hour recall (24HR) instruments, it is now possible to envision their use in cohort studies investigating the relation between nutrition and disease. Understanding that all dietary assessment instruments are subject to measurement errors and correcting for them under the assumption that the 24HR is unbiased for usual intake, here the authors simultaneously address precision, power, and sample size under the following 3 conditions: 1) 1-12 24HRs; 2) a single calibrated food frequency questionnaire (FFQ); and 3) a combination of 24HR and FFQ data. Using data from the Eating at America's Table Study (1997-1998), the authors found that 4-6 administrations of the 24HR is optimal for most nutrients and food groups and that combined use of multiple 24HR and FFQ data sometimes provides data superior to use of either method alone, especially for foods that are not regularly consumed. For all food groups but the most rarely consumed, use of 2-4 recalls alone, with or without additional FFQ data, was superior to use of FFQ data alone. Thus, if self-administered automated 24HRs are to be used in cohort studies, 4-6 administrations of the 24HR should be considered along with administration of an FFQ.     

Evaluation of a computerized food frequency questionnaire to estimate calcium intake of Asian, Hispanic, and non-Hispanic white youth

The objective of this study was to evaluate a computerized food frequency questionnaire (FFQ) that estimates calcium intake among Asian, Hispanic, and non-Hispanic white youth. A computerized FFQ based on a list of 80 foods with corresponding food photos was evaluated for 4 consecutive weeks. The evaluation study consisted of computerized FFQs during weeks 1 and 4, and 24-hour dietary recalls during each of weeks 2 and 3. Subjects were a convenience sample of Asian (29%), Hispanic (36%), and non-Hispanic white (35%) youth, age 11 to 18 years, living in northern Utah (N=161). Paired t tests, percent agreement, Pearson correlation coefficients of transformed calcium intake values (using ladder of transformation), deattenuated Pearson correlation coefficients, and Spearman correlation coefficients were used to evaluate the computerized FFQ. The correlation of calcium intakes estimated by the first and second computerized FFQ, 1 month apart, was 0.72 (transformed Pearson's r) for the total sample (N=161). Correlations within subgroups were: for males, r=0.59; females, r=0.81; 11- to 14-year-olds, r=0.66; 15- to 18-year-olds, r=0.82; Asians, r=0.73; Hispanics, r=0.76; and non-Hispanic whites, r=0.61. The correlation of calcium intakes estimated by the second computerized FFQ and the mean of two 24-hour dietary recalls was 0.56 (deattenuated, transformed Pearson's r) for the total sample. Correlations were also significant for males (r=0.50), females (r=0.57), 11- to 14-year-olds (r=0.56), 15- to 18-year-olds (r=0.59), Asians (r=0.63), Hispanics (r=0.55), and non-Hispanic whites (r=0.57). This computerized FFQ was found to be reliable in estimating calcium intake among a multiethnic youth population in the United States.     

Development and validation of a food frequency questionnaire to estimate the intake of genistein in Malaysia

Abstract

Objective: To develop and validate a food frequency questionnaire (FFQ) to estimate the genistein intake in a Malaysian population of pregnant women. Method: A single 24-h dietary recall was obtained from 40 male and female volunteers. A FFQ of commonly consumed genistein-rich foods was developed from these recalls, and a database of the genistein content of foods found in Malaysia was set up. The FFQ was validated against 7-d food diary (FD) kept by 46 pregnant women and against non-fasting serum samples obtained from 64 pregnant women. Reproducibility was assessed by comparing the responses on two FFQs administered approximately 1 month apart. Results: The Pearson correlation coefficient between FFQ1 and FD was 0.724 and that between FFQ2 and FD was 0.807. Classification into the same or adjacent quintiles was 78% for FFQ1 versus FD and 88% for FFQ2 versus FD. A significant dose -- response relation was found between FFQ-estimated genistein intake and serum levels. Conclusion: The FFQ developed is a reliable, valid tool for categorising people by level of genistein intake.     

Development of a food frequency questionnaire to estimate calcium intake of Asian, Hispanic, and white youth

OBJECTIVE: To develop a food frequency questionnaire (FFQ) that estimates calcium intake of Asian, Hispanic, and white youth living in the western United States. DESIGN: A list of 80 foods was assembled to create an FFQ to measure calcium intake. Evaluation of the FFQ spanned four consecutive weeks. An FFQ was completed during Week 1 and Week 4, and a 24-hour dietary recall was completed during Week 2 and Week 3.Subjects/setting A convenience sample of 162 Asian, Hispanic, and white youth ages 10 to 18 years was selected.Statistical analyses performed Percent agreement, paired t tests, Pearson correlation coefficients of cube-root transformed values, and deattenuated Pearson correlation coefficients of cube-root transformed values were used to evaluate the FFQ. RESULTS: The correlation between calcium intake estimates, when measured by first and second administrations of the FFQ, was 0.68 (Pearson's r) for the total sample. Correlations differed by age, sex, and ethnic subgroups as follows: 10 to 13 years (r=0.62), 14 to 18 years (r=0.73), male (r=0.73), female (r=0.64), Asian (r=0.77), Hispanic (r=0.72), and white (r=0.48). The correlation between calcium intakes as estimated by the second FFQ vs the average of the two 24-hour dietary recalls was 0.54 (deattenuated Pearson's r) for the total sample. This correlation differed by age, sex, and ethnic subgroups as follows: 10 to 13 years (r=0.46), 14 to 18 years (r=0.59), male (r=0.65), female (r=0.45), Asian (r=0.64), Hispanic (r=0.18), and white (r=0.57). CONCLUSIONS: A unique dietary survey has been developed to estimate calcium intake among Asian, Hispanic, and white youth in the United States.     

A food frequency questionnaire for adolescents: defining eating patterns.

A self-administered food frequency questionnaire (FFQ) was developed to indicate weekly consumption of 64 foods. Reliability, validity, and usefulness of the tool to define a protein eating pattern were determined. Adolescents (N = 1,108) completed the FFQ during the Bogalusa Heart Study. Two-hour and 2-week reliability measures demonstrated consistency of intake of specific foods. Frequency of foods obtained from seven consecutive 24-hour recalls was compared with frequency obtained from the FFQ. A mean 50% agreement for both frequency and quantity of food intake was observed. Geometric means showed differences in mean number of protein foods by age of adolescent but the only significant difference was for beef intake of 15-year-olds. White children reported more servings of beef, cheese, and vegetables with meat than did black children. Black children reported more servings of eggs, luncheon meat, pork, poultry, and total protein than did white children. Boys reported a greater frequency of total protein foods, specifically, eggs, milk, and poultry, than did girls. Significant correlations were noted between low-density-lipoprotein cholesterol and intake of eggs and luncheon meat. We were able to quantitate the reliability and validity of the FFQ and to use it to explore the association of specific eating patterns with cardiovascular disease risk.