Contents - Next

This is the old United Nations University website. Visit the new site at


Food and Nutrition Bulletin
Published by the United Nations University Press, Tokyo, Japan.

Editorial address:

Food and Nutrition Bulletin
9 Bow Street
Cambridge, MA 02138, USA
Tel: (617) 495-0417
Fax: (617) 495-5418
Telex/cable: 92 1496

Subscription address:

United Nations University Press
The United Nations University
Toho Seimei Building
15-1 Shibuya 2-chome, Shibuya-ku
Tokyo 150, Japan
Tel.: (03) 3499-2811. Fax: (03) 3499-2828.
Telex: J25442. Cable: UNATUNIV TOKYO.

The Food and Nutrition Bulletin incorporates and continues the PAG Bulletin of the former Protein-Calorie Advisory Group of the United Nations system and is published quarterly by the United Nations University Press in collaboration with the United Nations ACC Sub-committee on Nutrition. The views expressed are those of the authors and not necessarily those of the United Nations University or the ACC Sub-committee on Nutrition.

All correspondence concerning the content of the Bulletin, comments, news, and notices should be sent to the editor at the Cambridge office address given above. All material may be freely reproduced provided acknowledgement is given and a copy of the publication containing the reproduction is sent to the Bulletin.

Editorial policy

The Food and Nutrition Bulletin is intended to make available policy analyses. state-of-the-art summaries, and original scientific articles relating to multidisciplinary efforts to alleviate the problems of hunger and malnutrition in the developing world. It is not intended for the publication of scientific articles of principal interest only to individuals in a single discipline or within a single country or region. Notices of relevant books and other publications will be published if they are received for review. The Bulletin is also a vehicle for notices of forthcoming international meetings that satisfy the above criteria and for summaries of such meetings.

The Food and Nutrition Bulletin also serves as the principal outlet for the publication of reports of working groups and other activities of the UN ACC Sub-committee on Nutrition (SCN) and its Advisory Group on Nutrition. The SCN itself is a focal point for co-ordinating activities of FAO, WHO, UNICEF, the UNU, Unesco, the World Bank, the World Food Programme, the World Food Council, the United Nations Environment Programme, and other bodies of the United Nations system which have an interest in food and nutrition.

Unsolicited manuscripts of articles of the type published in this and previous issues may be sent to the editor at the Cambridge office address given above. They must be typed, double-spaced, with complete references and must include original copy for any figures used (see the " Note for contributors" in the back of this issue). All articles submitted will be reviewed promptly and the author will be notified of the editorial decision. Any disciplinary or conceptual approach relevant to problems of world hunger and malnutrition is welcome, and controversy over some of the articles is anticipated. Letters to the editor are encouraged and will be printed if judged to have an adequate basis and to be of sufficient general interest.

It is expressly understood that articles published in the Bulletin do not necessarily represent the views of the United Nations University or of any United Nations organization. The views expressed and the accuracy of the information on which they are based are the responsibility of the authors. Some articles in the Bulletin are reports of various international committees and working groups and do represent the consensus of the individuals involved; whether or not they also represent the opinions or policies of the sponsoring organizations is expressly stated.

The United Nations University (UNU) is an organ of the United Nations established by the General Assembly in 1972 to be an international community of scholars engaged in research, advanced training, and the dissemination of knowledge related to the pressing global problems of human survival, development, and welfare. Its activities focus mainly on peace and conflict resolution, development in a changing world, and science and technology in relation to human welfare. The University operates through a worldwide network of research and postgraduate training centres, with its planning and co-ordinating headquarters in Tokyo, Japan.

The United Nations University Press, the publishing division of the UNU, publishes scholarly books and periodicals in the social sciences, humanities, and pure and applied natural sciences related to the University's research.

Food and Nutrition Bulletin Editor:

Dr. Nevin S. Scrimshaw Assistant Editor: Ms. Edwina B. Murray Editorial Consultant: Ms. Sarah Jeffries Associate Editors: Dr. Herná Delgado, Director, Institute of Nutrition of Central America and Panama (INCAP), Guatemala City, Guatemala Dr. Cutberto Garza, Director and Professor, Division of Nutritional Sciences. Cornell University, Ithaca, N.Y., USA Dr. Peter Pellet, Professor, Department of Food Science and Nutrition, University of Massachusetts, Amherst, Mass.' USA Dr. Aree Valyasevi, Professor and Institute Consultant, Mahidol University, Bangkok, Thailand

Food and Nutrition Bulletin, vol. 13, no. 4

(c) The United Nations University, 1991

United Nations University Press
The United Nations University, Toho Seimei Building
15-1 Shibuya 2-chome, Shibuya-ku, Tokyo 150, Japan
Tel.: (03)3499-2811 Fax: (03)3499-2828
Telex: J25442 Cable: UNATUNIV TOKYO

ISBN 92-8080793-5
ISSN 0379-5721

Typeset by Asco Trade Typesetting Limited, Hong Kong
Printed by Permanent Typesetting and Printing Co. Ltd., Hong Kong


Public health nutrition

Comparison of food-frequency and 24-hour-recall instruments for estimating vitamin A intake

Lilian Portocarrero, Julieta Quan de Serrano, Carmen Yolanda Lopez,
Elien de Zepeda. Alejandrina Vasquez Jesus Bulux, and Noel W. Solomons


The purpose of the study was to compare the 24-hour-recall method with the seven-day food-frequency questionnaire to estimate the dietary intake of vitamin A. Thirty-three women from a pert-urban neighbourhood in (Guatemala City were interviewed using both methods with a seven-day interval. The reproducibility of the results was analysed at both the individual and group levels. The 24-hour-recall method gave consistently higher average values for vitamin A intake, but the median was higher with the frequency method. The correlation coefficients between the repeated interviews were r = .49 for food frequency and r = -.07 for 24-hour-recalls. Diagnostic classification with the two methods was in good agreement at the 750-retinol-equivalents cut-off criterion for adequacy of vitamin A intake, but fell progressively as sequentially lower cutoffs were applied.



Guatemala is a country with a long history of hypo-vitaminosis A [1-3]. For that reason, there is an interest in defining the population's intake of vitamin A. Two of the various approaches to this are the 24-hour-recall method and the food-frequency questionnaire. On a population basis, it is accepted that both can provide a reasonable estimate of central tendency (i.e., mean, median) and of variance (i.e., standard deviation, range) [4-6], but there is serious concern about their ability to provide a reliable estimate for the individual [7, 8]. Dietary intake is not constant from day to day, making a single 24-hour recall subject to being unreliable and non-representative of the person's long-term consumption [7, 8]. In theory, food-frequency assessment would encompass more dietary experience, and its average, expressed on a per day basis, should provide a more stable estimate. In a recent dietary survey conducted in a pert-urban neighbourhood of Guatemala City, we examined the within-individual and inter-method reproducibility of estimates of vitamin A intake, which is believed to be the most variable of all nutrient intakes [9-12].


Population and methods


This research was part of a larger inquiry into dietary vitamin A intake, intestinal parasitosis, biochemical indices of vitamin A status, and haematology of pregnant women in the low-income, marginal, pert-urban settlement of Guajitos on the southern edge of Guatemala City. The control group consisted of 46 non-pregnant neighbours, 14-69 years old. All the women spoke Spanish, and their literacy rate was about 77%.

Food-frequency estimate of vitamin A-rich food

The intake of vitamin A from foods rich in this nutrient over the previous seven days was estimated using a food-frequency questionnaire [13]. This instrument listed 39 food and beverage items from the Guatemalan urban diet, and the subjects were asked how many servings of each item they had consumed in the last seven days. These were converted into standard portions and the corresponding value for vitamin A activity, expressed in retinol equivalents (RE), using the Food Composition Tables for Use in Latin America [14] and Central American food tables [15]. A seven-day tabulation of vitamin A intake was derived, and this was divided by 7 to provide an estimate of average daily intake.

24-hour-recall estimate of vitamin A intake

A history of intake of all foods and beverages consumed at meals and snacks in the previous 24 hours was elicited. The field team consisted of two research nutritionists and two research physicians who were trained in the procedure. The recall began with the last meal consumed and proceeded backwards for 24 hours. An estimate of portion sizes in common household measures was requested, and actual recipes were recorded for complex foods such as soups, stews, and casseroles. Portion sizes were converted to volumes and weights in grams and then into multiples of standard units. The corresponding values for vitamin A activity, expressed in RE, from the same food composition tables [14, 15] were applied to portion sizes.

Field procedures

The interviews were conducted from September to November 1987. Our intention was to apply both instruments to a given subject on the first day of study. and then again seven days later. This would have generated two 24-hour recalls and two food-frequency estimates at a seven-day interval. In practice, it was possible to obtain the full complement of data from only 33 subjects. The same research interviewer worked with the same subject throughout the study.

Data analysis

Descriptive statistics-arithmetic means, standard deviations, medians, and ranges-were calculated. The estimates for each individual resulting from the two food-frequency interviews and the two 24-hour recalls were compared on correlation scattergrams, with the results from the first day on the x axis and those from the second day on the y axis. The same-day pairings of the results from the two instruments were similarly compared with each other, with the food-frequency estimates on the x axis and the 24-hour-recall estimates on the y axis. Pearson's product moment linear regression analyses were performed on the paired within-individual and inter-method data.

Two-by-two contingency tables for the correspondence of diagnostic classification were constructed relative to three assumptions of adequacy: (1) at 100% of the WHO recommended vitamin A intake [16] for adults (750 RE), (2) at 67% of the recommendation (500 RE), and (3) at 50% of the recommendation (375 RE).



Within-individual test-retest reproducibility (precision)

Food-frequency estimates

Figure 1 shows the intakes of vitamin A as estimated using the food-frequency instrument. The mean for the 33 estimates on the first occasion was 413 ± 446 RE, the median was 274 RE, and the range was 131,787 RE. For the second occasion these were respectively 480 ± 420 RE, 351 RE, and 38-1,748 RE. From the first assessment to the second, 17 individuals (52%) reported a greater than 10% increment of vitamin A intake, 12 (36%) reported a decrement of over 10%, and 4 (12%) remained within ± 10% of their original estimate. The maximum increase was seen in a subject whose estimated intake rose by 93% from one week to next.

The maximum decrease was a reduction of 78% over seven days. The correlation coefficient for the two frequency estimates was r = .49. On the basis of the results obtained in the first interview, it would have been possible to predict only 23% of the variation in the results of the second interview (fig. 2).

The two-by-two tables for the reproducibility of the diagnostic classification at the three assumptions of adequacy in relation to the 1967 WHO/FAO recommended levels [16] are shown in figure 3 for the first and second estimates using food frequency. Twenty six individuals (79%) were classified twice in the same category at the 750-RE cut-off criterion.

24-hour recall estimates

Figure 4 shows the estimates of vitamin A intakes from the 24-hour recalls. The mean estimate for the 33 recalls on the first of two occasions was 507 ± 1,609 RE, with a median of 165 RE and a range of 11-9,376 RE. For the second occasion these were respectively 677 ± 1,794 RE, 240 RE, and 0-9,818 RE. Twenty subjects (61%) had an increment of 10% or greater, 12 (36%) had a decrement of 10% or greater, and 1 had a change within ± 10%. The maximum increase was 98% and the maximum decrease 100% at the second recall.

FIG. 1. Estimates of vitamin A intake using the food-frequency instrument (N= 33)

FIG. 2. Correlation coefficient for the two food-frequency estimates (r = .49; y = 294.65 + 0.47 x; N = 33)

FIG. 3. Reproducibility of the diagnostic classification at three assumptions of adequacy (750 RE, 500 RE, and 375 RE)

FIG. 4. Estimates of vitamin A intake using the 24-hour-recall instrument (N= 33)

FIG. 5. Correlation coefficient for the two 24-hour-recall estimates (r = -.07; y = 718.83 - 0.08 x; N = 33)

Figure 5 shows the correlation of the first and second 24-hour results. The correlation coefficient was -.07. The value for the coefficient of determination, r²,, indicated that data obtained on one day's food consumption by the recall method have limited value in predicting the intake values on another day. With the two-by-two analysis for diagnostic classification at various levels of intake adequacy with 24-hour recall, 28 (85%) of the individuals were classified twice in the same category at the 750-RE cut-off; this fell progressively at the 500-RE and 375-RE cut-offs.

Inter-method correspondence

The individual relationship between the 24-hour recall and the food-frequency questionnaire for the same participant on the same day was evaluated as shown in figure 6. The mean estimate of vitamin A intake from 66 of the 24-hour recalls was 592 ± 1,693 RE, compared to 450 ± 431 RE for the food frequency questionnaire. The respective medians of intake were 187 RE and 291 RE. In 25 (38%) of the 66 pairs of interviews, the 24-hour-recall value was higher than that derived from the food-frequency method; it was over 10% higher in 24 pairs of interviews (36%). In one subject the 24-hour recall estimated a value 85% higher than that calculated from the corresponding food-frequency questionnaire on the first occasion.

FIG. 6. Estimates of vitamin A intake using the 24-hour-recall and food-frequency instruments (N = 33)

The food-frequency value was the higher of the two; in 39 women (59%) it was 10% or more higher than the corresponding 24-hour estimate. In one subject the estimated intake obtained using the food frequency method was 100% greater than that obtained using the 24-hour recall. The crude correlation coefficient for the correspondence of the two methods was .36 (fig. 7), but this improved to .58 when the three major outlying data pairs were eliminated from the regression. On the basis of the results obtained by the food-frequency interviews, it would have been possible to predict 34% of the variation in the results of the 24-hour interviews with the optimized data set (eliminating outliers).


As both of the methods used in this study involve recall of dietary intake and can be performed in a relatively short period of time, they represent practical options for field survey work. At the level of description of a population both approaches provide about equally accurate estimates. Madden et al. [4], Gersovitz et al. [6], and others have showed clear relationships between reported and weighed group mean values. Hankin et al. [17] found that group mean values do not vary significantly from day to day. Young et al. [5] compared a 24-hour recall with the average value from a seven-day dietary-intake record recorded by the subjects. They found that the group means given by the two methods were "interchangeable" if the number of subjects was greater than 50. Without having a true gold standard for dietary intake, we cannot address this issue with our data. It is comforting, however, that the median estimates derived from both techniques were fairly consistent.

It is at the individual level that the day-to-day variation in dietary intake reduces the representativeness of short-term estimates of consumption. Much has been written about the need for several 24-hour-recall estimates to provide stable averages for individuals [8, 9, 11, 18]. Theoretically, food-frequency data for seven days would smooth out some of the peaks and valleys of each day's variation in intake. If the week during which the interview is conducted is representative of long-term intake, the average from a seven-day food-frequency interview should be more stable. Data sets based on observations covering several days will present less bias in the estimate of centiles, or of prevalences of high or low intake, than data sets based on a single day's observations [8].

A partial test of this notion examined the degree of correspondence between the first and second weeks in our food-frequency test-retest reproducibility format. The correlation coefficient was of a high order (r = .48), statistically significant with 33 pairs of data, and suggestive of stability for the seven-day interval. However, the correlation between two 24-hour-recall estimates in the same individual is much lower (r = - .07) than for food frequency, which confirms the hierarchical nature of stabilities. Poor correspondence in the case of the 24-hour-recall method is not unexpected in view of the high day-to-day variation. Food intake data on a single day are of limited value in predicting the intake on another day [9, 11]. Adelson [19] interviewed 39 men, who provided records for two consecutive weeks, and concluded that the most substantial differences of nutritive content of food were in vitamin A and ascorbic acid. Anderson et al. [20] pointed out that three to seven days of observations would provide a reasonable picture of the distribution of usual intakes for many nutrients, but for some nutrients with large day-to-day variations, such as vitamin A, many more days of observation would be required before reliable estimates of the distributions could be obtained.

FIG. 7. Correlation coefficient for the food-frequency and 24-hour-recall estimates (r = .36: y = 106.36 + I .13 x; N = 33)

It is one thing to reproduce with accuracy the same numerical value with the same survey instrument, and another to assign the same diagnostic classification in a system of binary assignment of adequate or inadequate. In terms of their ability to classify the same individuals twice in the same adequacy category, the two methods, frequency and recall, had a reasonable concordance at the 750-RE cut-off criterion. With the food-frequency interview 26 individuals (79%) were classified twice in the same category, and 28 (85%) with the 24-hour recall; but this fell progressively as sequentially lower cut-offs were applied. It was also our intention to determine how well the 24-hour recall corresponded to the simultaneous food-frequency estimate. The inter-method coefficient between the two instruments, adjusted to remove gross outliers, was .58, and the correspondence of diagnostic classification was 94% of the individuals classified twice in the same category at the 750-RE cut-off criterion.

The design of this study allowed us to evaluate the use of standard dietary survey instruments under real field conditions in the same individuals simultaneously. Even though we were dealing with the estimation of vitamin A intake, reputedly the single most difficult nutrient to estimate in a precise manner [9-12], the expected hierarchical relationships of stability were confirmed. Moreover, we were able to put quantitative numbers on the strength of the correspondence between the two methods for correlations and binary discrimination.

Given the moderate to poor test-retest reproducibility of the instruments in the same individual, the most important conclusions would be to avoid use of the 24-hour recall, and to exercise caution in seeking to explore biological associations using the estimate of vitamin A from a single seven-day food-frequency questionnaire as the individual value [21].



1. Arroyave G, Bauernfeind JC, Olson JA, Underwood BA. Selection of intervention strategies. In: Guidelines of the eradication of vitamin A deficiency and xerophthalmia: a report of the International Vitamin A Consultative Group. Washington, DC: The Nutrition Foundation, 1976.

2. Institute of Nutrition of Central America and Panama. Evaluación nutricional de la pablación de Centro America y Panama. Guatemala City: INCAP, 1969: V25-V30.

3. Interdepartmental Committee of Nutrition for National Defense, and Institute of Nutrition of Central America and Panama. Nutritional evaluation of the population of Central America and Panama 1965-1967: Regional summary. DHEW Publication no. HSM-72-8120. Washington, DC: DHEW, 1972.

4. Madden JP, Goodman SJ, Guthrie HA. Validity of the 24-hour recall: analysis of data obtained from elderly subjects. J Am Diet Assoc 1976;68: 143-47.

5. Young CM, Hagan GC, Tucker RE, Foster WD. A comparison of dietary study methods: II. Dietary history vs. seven-day record vs. 24-hour recall. J Am Diet Assoc 1952;28:218-21.

6. Gersovitz M, Madden JP, Smiciklas-Wright H. Validity of the 24-hour dietary recall and seven-day record for group comparisons. J Am Diet Assoc 1978;73:48-55.

7. Rasanen L. Nutrition survey of Finnish rural children: VI. Methodological study comparing the 24-hour recall and the dietary history interview. Am J Clin Nutr 1967;32:2560-67.

8. Emmons L, Hayers M. Accuracy of 24-hr recalls of young children. J Am Diet Assoc 1973;62:409-15.

9. Beaton GH, Milner DJ, Corey P, et al. Sources of variance in 24-hour dietary recall data: implications for nutrition study design and interpretation. Am J Clin Nutr 1979;32:2546-59.

10. gingham SA. The dietary assessment of individuals: methods, accuracy, new techniques and recommendations. Nutr Abst Rev (series A) 1987;57:705-42.

11. Sempos CT, Johnson NE, Smith EL, Gilligan C. Effects of intraindividual and interindividual variation in repeated dietary records. Am J Epidemiol 1985;121:12030.

12. Nelson M, Black AK, Morris JA, Cole TJ. Between- and within-subject variation in nutrient intake from infancy to old age: estimating the number of days required to rank dietary intakes with desired precision. Am J Clin Nutr 1989;50:155-167.

13. Quan de Serrano J, de Gonzales L. The pattern of contemporary dietary vitamin A intake in various geographic regions of Guatemala as determined by a 7-day food frequency questionnaire. (Abstract) J Am Coll Nutr 1987;6(5):438.

14. Wu-Lueng WT, Flores M. Tabla de composición de alimentos pare uso en America Latina. Interdepartmental Committee of Nutrition, and Institute of Nutrition of Central America and Panama (INCAP). Washington, DC: US Government Printing Office, 1961.

15. Flores M, Menchu MT, Lara MY. Valor nutritivo de los alimentos pare Centro America y Panama. Guatemala City: INCAP, 1971.

16. World Health Organization, and Food and Agricultural Organization. Requirements of vitamin A, thiamine, riboflavin and niacin. WHO Technical Report Series, no. 362. Geneva: World Health Organization, 1967.

17. Hankin JH, Reynolds WR, Margen S. A short dietary method for epidemiologic studies: 11. Variability of measured nutrient intakes. Am J Clin Nutr 1967;20:935-45.

18. Suitor CJ, Gardner J, Willet WC. A comparison of food frequency and diet recall methods in studies of nutrient intake of low-income pregnant women. J Am Diet Assoc 1989;89(12):1786-94.

19. Adelson SF. Some problems in collecting dietary data from individuals. J Am Diet Assoc 1960;36:453-61.

20. Anderson GH, Peterson RD, Beaton GH. Estimating nutrient deficiencies in a population from dietary records: the use of probability analyses. Nutr Res 1982;2:409-15.

21. Tangney CC, Shekelle RB, Raynor W, Gale M, Betz EP. Intra- and interindividual variation in measurements of b-carotene, retinol and tocopherols in diet and plasma. Am J Clin Nutr 1987; 45:764-69.

Contents - Next