Dietary assessment is the process of evaluating what people eat by using one or several intake indicators. It is the best approach for identifying nutrients that are likely to either be under-or overconsumed by the individual or groups of interest. It also can be used to identify food patterns and preferences.
Dietary Status Versus Nutritional Status
Dietary status is related to but not necessarily reflective of nutritional status. Nutritional status is a more comprehensive term, referring to health status as it is affected by nutrition. It is measured not only by assessing dietary status, but also by anthropometric, biochemical, and clinical measures. Because dietary methods are less invasive, somewhat easier to obtain than other physiological measures, and do not require medical training, they often are used initially for assessing nutritional inadequacy or excess. Physiological measurements are then used to confirm and corroborate dietary intake evaluation and to arrive at definitive assessments of nutritional status.
Tools and Standards for Assessment
To assess dietary intake, food composition tables for translating foods consumed into nutrients, and a reference against which dietary intakes may be compared, are needed. These tools have been updated and refined periodically and appropriate ways for applying them to assessment tasks are steadily clarified.
Overcoming Imperfections in Assessing Dietary Intake
All dietary assessment methods are imperfect, regardless of how well they are designed. Their major shortcomings and measures for dealing with the imperfections are described briefly below.
Capture Actual Intakes
The various methods for assessing dietary intake are summarized in Table 1 and elsewhere in detail (Dwyer, 1999). All assessment methods fail to capture actual energy intakes precisely and probably intakes of nutrients as well. Some of the errors are inevitable because human beings tend to misreport their food intakes, but the method used also influences assessment outcomes.
Dietary intake is sometimes assessed by an objective observer rather than by the eaters themselves. For example, the intake of a hospitalized patient often is assessed from measured differences of the food served to a patient less any unconsumed amounts. Such objective methods have the advantage of being less subject to reporting biases than those that rely solely on recall. However, more objective methods are time-consuming, costly, cannot usually be employed to assess typical intake, and fail to record all intake. Moreover, they may not reflect what people really eat, since people may eat differently when they know that they are being observed. For these reasons, most commonly used dietary assessment methods rely on eaters' self-reported intakes.
Most methods such as twenty-four-hour recalls, food records, and diaries underreport actual energy intake by at least 20 percent. Underreporting errors are even higher (30 percent or more) in certain groups, such as the obese, women, and the elderly. However, they also vary among individuals in ways that are not always easily identified by demographic or other distinguishing characteristics. The causes of underreporting include forgetting, unconscious alterations in recalling foods eaten (for example, when the individual knows that he or she is being watched), attempts to please the questioner, and occasionally lack of cooperation by the subject. Non-random biases are difficult to deal with statistically.
Intakes obtained using semiquantitative food frequency questionnaires have other shortcomings. This method presents the respondent with a food list. These prompts may decrease forgetting, but insertions and "false memories" of foods consumed or of the consumption of socially desirable foods may be reported rather than true intakes. Semiquantitative food frequency questionnaires are too imprecise to estimate individual intakes quantitatively. Nutrient intakes from semiquantitative food frequency questionnaires usually are overestimated. They usually are adjusted statistically to obtain more accurate estimates of usual intakes. Measures of usual energy intakes for accurate groups specified by sex and age obtained by other methods or from estimates of energy outputs are used to adjust them. They are often derived by "food frequency" approaches and may be accurate enough to provide reasonable group estimates, although such measures are not sufficiently accurate for individuals. Also, precise quantification of absolute amounts (as opposed to levels of intake ranked into quartiles or quintiles) is not possible. The biases involved in food frequency questionnaires are complex, and statistical methods for obtaining valid estimates of intakes are unavailable.
Understandably, retrospective methods that rely on memory are subject to "forgetting bias." Prospective methods, which rely on reporting food intake immediately or shortly after eating, are more subject to alterations in intake due to the individual's awareness that his or her intake is being recorded. The extent to which social desirability and reporting biases intrude in the various methods is unknown, but is probably considerable.
Not all of the problems associated with misreporting can be overcome by the method of choice, but some can be minimized by selecting the appropriate tool for the task at hand.
Obtain Representative Intakes
Dietary assessments must be done frequently and randomly to reflect usual intake faithfully. This is an important shortcoming because only usual intake is correlated with nutritional status. A representative sample of randomly chosen days that includes both weekdays and weekends is best for obtaining accurate twenty-four-hour recalls or records. Semiquantitative or other food frequency questionnaires also may assist in providing information on usual food intake patterns.
Table 1
| Dietary assessment methods | |
| Method | Description, advantages, and limitations of method |
| Retrospective Methods 24-hour recall | Respondent recalls all foods and beverages consumed in a given 24-hour period and reports them to a trained interviewer, who probes to get additional details on portion sizes, frequency, and forgotten items. Positive aspects include low respondent burden, ease in administration, and minimization of biases associated with altering food intake because of knowledge that one is being observed. Negative aspects of the method include forgetting, deliberate misreporting, need for a trained observer to administer, need for several days of intakes to obtain estimate of usual diet, and costs associated with computerized analysis of records |
| Telephone recall | The respondent is contacted or instructed in advance and given instructions about estimating portion sizes and other details. Then the respondent is called by telephone and asked to report dietary intake over the past 24 hours. Probes and techniques are usually standardized to minimize reporting error. Positive aspects of the method include those listed above plus ability to obtain representative random days of intake, and decreased cost of administration. Negative aspects include inability to obtain interviews from those without telephones, and for those who find telephones difficult to use, and errors in reporting portion sizes. |
| Food frequency and semiquantitative food frequency questionnaire | Respondent chooses from a list of different foods or food groups usually eaten over the past month or year. The number and type of foods, and whether portion sizes are specified, varies from one questionnaire to another. Positive aspects of the method include ease of administration, low expense, less forgetting because of prompts furnished by food lists, somewhat more of an estimate of usual intake (perhaps equivalent to 2–3 days), and low costs of data analysis. Negative aspects of method include incomplete reporting of items not included in food lists, overreporting, incomplete or inaccurate response, inaccurate translation of food and food groups to nutrients, and imprecise estimates of nutrient intake |
| Dietary history | Respondent reports all foods and beverages consumed on a usual day to a trained interviewer. The interviewer then probes further on the frequency amount and portion size consumed. Diet diaries are sometimes used to assist respondents in recalling their intakes. Positive aspects of the method are that respondent burden is low and complete intakes are provided. Negatives include high cost, need for trained interviewers, and lack of standardization |
| Prospective Methods Weighed food record | After being instructed, respondent weighs all food and drink consumed on a small weighing scale and reports it on a record that is kept as close to the time of consumption as possible. If observers are available, they can carry out the weighing themselves. Positive aspects of the method are lack of forgetting bias, and ability to obtain random days of intake. Negatives include high respondent burden, refusal to record intakes, need for an expert observer to review and clarify intakes reported, tendency of respondents to alter food intake when they know they are under observation, and costs of data analysis. |
| Food diary | The respondent records all foods consumed in household measures, usually without measuring them, or only measuring foods that are particularly difficult to estimate. Positive aspects are same as food records but respondent burden is less. Negative aspects are that more errors in estimation of portion size may occur |
| Duplicate portion analysis | An observer takes duplicate portions of all foods consumed by the individual and weighs or measures them; in some cases, these may also be chemically analyzed. Positive aspects are similar to food records. Negative aspects are lack of respondent cooperation, need for trained observers, cost of food analysis, and inability to obtain estimate of usual intake. |
| Other: Direct observation by trained observers or by videotaping subjects | Observer records or watches food intake in a controlled or highly supervised environment in which it is possible to videotape or directly observe food intakes. Positive aspects of the methods are that they do not rely on respondent burden. Negative aspects are that the methods are usually too imprecise for obtaining valid estimates of individual intakes. |
Obtain Total Intakes
Many foods and beverages are fortified with nutrients, and a substantial proportion of the population takes nutrient supplements on a regular basis. For some individuals, these nutrient sources contribute a substantial amount of vitamins and minerals. Nutrient intakes from all sources, including foods and beverages, fortified foods, and nutrient supplements must be included in all dietary assessments. If only food sources are queried, this fact should be noted.
Use Complete Food Composition Tables
Once food intakes are obtained, these must be translated into nutrients using food, beverage, and supplement composition tables. Accurate nutrient intakes can be obtained if up-to-date and complete food composition tables are available; that is, the composition of fortified foods, nutrient supplements, and beverages must be included and tables must be complete for all nutrients and other bioactive substances of interest.
Appropriate References
Estimated nutrient intakes must be compared with appropriate references; in the United States and Canada, these are the Dietary Reference Intakes, or DRIs. Their use in dietary assessments is the subject of a recent report (Dietary Reference Intakes, 2000).
Inadequacies, Excesses, and Imbalances May Coexist
In the past, dietary assessments focused on dietary inadequacies. Although these are still relevant, nutrient excesses and imbalances of nutrients also are of concern in most Western countries, and therefore also must be considered. Several of the DRIs are helpful in these respects. DRIs for macronutrients will be published in the near future.
Appropriate Interpretation of Assessment Results
The estimated average requirement, or EAR, is the nutrient intake estimated to meet the requirement of half the healthy individuals in a particular life stage or gender group. The recommended dietary allowance, or RDA, is the average daily dietary intake that suffices to meet the nutrient requirement of nearly all (97–98 percent) healthy individuals in a particular life stage and gender group. The adequate intake, or AI, is a recommended intake based on observed or experimentally determined approximations or estimates of nutrient intake by a group (or groups) of healthy people whose intakes are assumed to be adequate. The AI is used when an RDA cannot be determined. When the AI's are not based on mean intakes of healthy populations, these values are likely to be less accurate. The tolerable upper intake level (UL) is the highest usual daily nutrient intake likely to pose no risk of adverse health effects to almost all individuals in the general population. As intakes increase above the UL, the risks of adverse effects also increase. The assessment of dietary adequacy is imprecise. A specific individual's actual requirement for a specific nutrient generally is never known. Second, often the number of days that intakes are measured are likely to be insufficient to overcome errors in measuring intake and normal day-to-day variation. Although dietary data alone are not sufficient to assess nutritional status, intakes of individuals can be compared to certain of the DRIs. A usual intake based on a large number of days that is at or above the RDA or AI has a low probability of inadequacy. An intake above the UL places an individual at risk of adverse effects from excessive nutrient intakes. When observed intakes are habitually below the EAR, increased intakes usually are needed because the probability of adequacy is 50 percent or less. Habitual intakes between the EAR and the RDA also probably need to be improved because the probability of adequacy is less than 97 to 98 percent. Quantitative estimates of risk of inadequacy are more difficult to obtain. However, they can be calculated using methods described in a recent report (Dietary Reference Intakes, 2000).
The DRIs also are used to assess the dietary intake of groups. These assessments determine the percentage of individuals whose intakes are estimated to be inadequate. The EAR is used to estimate the prevalence of inadequate intakes within a group. A mean usual group intake at or above the AI implies a low prevalence of inadequate intakes. The UL is used to estimate the percentage of the population at risk of adverse effects from excessive intakes consumed on a chronic basis. Thus, the RDA is not used to assess nutrient intakes of groups.
Conclusions
Dietary assessment is a necessary component of nutritional status assessment of individuals, and also is useful for other purposes. It can be done using a variety of methods, each of which has advantages and limitations. However, regardless of which method is chosen, it is important that certain criteria be met. Intake from all sources (food, fortified food, beverages, and nutrient supplements) must be included. Sufficient numbers of days to represent usual intakes must be obtained. Complete food and supplement composition tables must be employed. Appropriate reference standards and statistical procedures for assessing intakes must be used. Dietary assessment methods work best in combination with other methods for the assessment of nutritional status.
Bibliography
Dwyer, J. T. (1997). "Assessment of Dietary Intake." In Modern Nutrition in Health and Disease, edited by M. Shils, J. A. Olson, M. Shike and A. C. Ross, 8th ed., pp. 887–904. Baltimore: Williams and Wilkins, 1997.
Dwyer, J. T. "Dietary Assessment." In Modern Nutrition inHealth and Disease, edited by M. Shils, J. A. Olson, M. Shike, and A. C. Ross, 9th ed., pp. 937–962. Baltimore: Williams and Wilkins, 1999.
Nusser, S., A. L. Carriquiry, K. W. Dodd, and W. A. Fuller. "A Semiparametric Transformation Approach to Estimating Usual Daily Intake Distributions." Journal of the American Statistical Association 91 (1996): 1440–1449.
Poehlman, E. T. "Energy Needs: Assessment and Requirements in Humans." In Modern Nutrition in Health and Disease, edited by M. Shils, J. A. Olson, M. Shike, and A. C. Ross, 9th ed., pp. 95–104. Philadelphia: Williams and Wilkins, 1999.
Subcommittee on Interpretation and Uses of Dietary Reference Intakes and Upper Reference Levels of Nuturients, Food and Nutrition Board, Institute of Medicine. Dietary Reference Intakes: Applications in Dietary Assessment. Washington, D.C.: National Academy Press, 2000.
—Johanna Dwyer
