Contents - Previous - Next

This is the old United Nations University website. Visit the new site at http://unu.edu

Guideline for reporting methods used in dietary surveys

Dietary assessment methods given the same name by different investigators may have the same general approach but differ in detail. It is therefore essential that the methods used should be fully described in published papers.

The following checklist, developed by the UK Nutritional Epidemiology Group, has been adopted by the Food and Nutrition Bulletin as a guideline for all papers that incorporate dietary survey information. The authors of such papers are requested to review this checklist carefully and to be sure to provide all the relevant information called for.

—The Editor

I. Sample characteristics

1. Sample (and control) recruitment

• How subjects were recruited —Sampling framework
• Numbers contacted, recruited, and completing study
• Reasons for non-completion —Use of incentives

2. Sample (and control) characteristics

• Age; sex —Height; weight —Social class —Other demographic/clinical information
• Whether sample represents the population studied
• Geographic coverage

3. Other information relevant to response or interpretation of results, e.g.

• Timing in relation to disease processes
• Timing in relation to interventions
• Timing in relation to season

II. Method of dietary assessment

A. Information required for ail methods

4. Method of dietary assessment; See definitions in section III below.
5. Validity of method

• Rationale for choice of method
• Whether instruments used have been pretested on a similar population
• Whether method has been validated against another dietary method or external markers of intake
• Whether the repeatability has been assessed

6. Method used for quantifying portions; See definitions in section IV below. Specify:

• Source of "average" portions
• Details of aids used to help in quantifying portions
• Scales used for weighing
• Method of quantifying unweighed foods in a weighed record

7. Food compositon database used for the analysis

• What database was used
• How foods that were not in the database were dealt with
• Any supplementary analytical work

8. Interviewers or fieldworkers

• Whether qualified (dieticians/nutritionists)
• Training given to unqualified fieldworkers
• Whether the same workers both collected and coded the data

9. Data collection procedures

Where and how data were collected (home / clinic / interview, face-to-face or telephone / self-completed, by post or computer)

• Number of interviews per subject
• Duration of interviews

10. Checking procedures

• When and how often records were checked with respondents
• Any checks for coding errors
• Any checks on the consistency of fieldworkers

B. Information required specific to different methods

11. Recall method

• How many and what days were recalled
• Whether all days of the week were included
• If not, whether results were weighted

12. Diet history

• Attempted time scale (current / recent past / distant past I season / whole year)
• Open-ended questions, or fully structured interview
• Structure of interview (Did it start with a 24-hour recall? Did it take each meal or each day of the week in turn to build up a picture of the diet? Did it include any cross checks for types or frequency of foods consumed? Were the subjects given any prompt lists?)

13. Food frequency (and amount) questionnaire

• Whether interviewer-administered or self-completed
• Whether instrument was pretested in a similar population
• Foods covered and options for frequency
• Rationale for the choice of foods

14. Study-specific questionnaire

• Whether interviewer-administered or self-completed
• Rationale for the form of the questionnaire
• Whether the instrument was pretested in a similar population
• Include the questionnaire as an appendix (see following note)

» General note on questionnaires

• It is desirable for the questionnaire to be included as an appendix, even if much reduced in size. This best describes the methods since it shows the questions asked and the foods and frequencies chosen. For the instrument to be "available from the authors" is unsatisfactory since it does not permit immediate evaluation of the study and in later years is unobtainable. At the very least, a copy of the questionnaire should be made available for review purposes.

15. All record methods

• How many and what days were studied
• Whether all days of the week were included
• If not, whether any adjustment or weighting was used
• How food eaten away from home was quantified
• What instructions and equipment were given to the respondent

III. Definitions: Dietary assessment methods

Dietary assessment. A blanket term for any method. Past intake may be assessed by interview or questionnaire, and present intake by records at the time of eating. Either approach may be qualitative or quantitative.

A. Interview techniques

Dietary questionnaire. This term has no precise meaning and is not an adequate description.

1. Diet recall. The respondent is asked to recall the actual food and drink consumed on specified days, usually the immediate past 24 hours (24-hour recall) but sometimes longer periods.
2. Diet history. The respondent is questioned about "typical" or "usual" food intake in a 12 hour interview. The aim is to construct a typical seven days' eating pattern. The interview may discuss each meal and intermeal period in turn or each day of the week in turn. Questions are usually open-ended, although a fully structured interview may be used. The diet history may be preceded by a 24-hour recall and/or supplemented by a checklist of foods usually consumed.
3. Food frequency (and amount) questionnaire. The respondent is presented a list of foods and is asked how often each is eaten in broad terms, such as x times per day, per week, per month, etc. The foods listed are usually chosen for the specific purposes of a study and may not assess the total diet. The questionnaire may be interviewer-administered or self-completed. Assessment of the quantities of food consumed on each eating occasion or day may also be included.
4. Study-specific dietary questionnaire. A term covering all dietary assessments using a set of predetermined questions but not conforming to any of the classic techniques defined above. The method is defined only by the questionnaire itself. The questionnaire may be interviewer-administered or self-completed.

B. Record techniques

Diet record. A blanket term for all record methods. In American literature it is often used without qualification but with "quantified in household measures" understood; since there are other forms of record, this is an inadequate description. A record is of actual food and drink consumed on specified days after the first contact by the investigator. The number of days recorded classically is seven but may be fewer or more.

5. Menu record, or food frequency record (the first term is preferable to avoid confusion with "food frequency questionnaire"). Record obtained without quantifying the portions. It may subsequently be analysed in terms of frequencies of consumption, or the investigator may assign "average" weights to portions. Because the respondent does not indicate quantities, there can be no attempt to identify the true weight of individual portions (cf. "estimated record").
6. Estimated record. A record of portions described in household measures (cups, spoons, etc.) with or without the aid of diagrams or photographs. This method aims to estimate the actual quantity eaten.
7. Weighed record (weighed inventory technique). Record with weights of portions as served and plate waste. (Weighed records are rarely fully weighed: estimated portions are usual for foods eaten away from home. )
8. Precise weighed record. A record kept by the respondent of all ingredients used in the preparation of meals, also inedible waste, total cooked weight of meal items, cooked weight of individual portions, and plate waste.
9. Cardiff photographic record. The respondent photographs food on the plate at the time of consumption. Portions are quantified by comparison with reference photographs of portions of known weight projected alongside the survey photographs.
10. Semi-weighed method for measuring family food intake. Method of Nelson and Nettleton. The total quantity of food served to a family is weighed, and quantities served to individuals are given in household measures. The term is sometimes mistakenly used for a weighed diet record where the authors acknowledge that not all food is in fact weighed.

C. Direct analysis techniques

11. Duplicate diets. Respondent keeps a weighed record and also weighs out and puts aside a duplicate portion of each food as consumed for later analysis by the investigator.
12. Aliquot sampling technique. Respondent keeps a weighed record and puts aside aliquot samples of food as consumed for later analysis.
13. Equivalent composite technique. Respondent keeps a weighed record. Subsequently a combined sample of raw foods equivalent to the mean daily amounts of food eaten is made up by the investigator for analysis.

IV. Definitions: Quantifying portions

• Qualitative (or unquantified) assessment. An assessment made only in terms of foods eaten, usually by counting frequency of consumption.
• Quantitative assessment. A dietary assessment that quantifies the portions of foods eaten in order to calculate nutrient consumption.

1. Average portions. Investigator assigns "average" portion weights derived from previous studies or experience. "Small," "medium," and "large" may also be used to indicate portion size in relation to the "average."
2. Household measures. Respondent describes portions in terms of household measures, e.g. cups, spoons, etc. "Standard" weights are assigned to the descriptions.
3. Photographic measures. Respondent is shown photographs of portions of known weight and asked how his/her own portion relates to the portion pictured. (Not to be confused with the Cardiff photographic record—see III(9) above.)
4. Food models/replicas. Respondent is shown three-dimensional models representing foods and asked how his/her own portion relates to the models. Models may be realistic replica foods or a variety of neutral shapes and sizes.
5. Weighed. The subject weighs and records each food item as it is consumed.

V. Computerized assessments

The term computer assessment does not define a method. Assessments conducted by computer should be described in the terms defined above.

Computer-conducted assessments differ from person-conducted assessments in the mechanics used. The computer may substitute for the paper and pencil of a self-completion questionnaire, or it may substitute for the interviewer in a diet history by fully structured interview.

Computerized interviewing may be combined with nutrient analysis to provide "instant" information on nutrient intake. Here the assumptions necessary to code foods and quantify portions are built into the program; the computer substitutes for the investigator in performing the post-interview coding tasks.

Contents - Previous - Next