2705_frame_C19 Page 477 Wednesday, September 19, 2001 1:23 PM

19
Dietary Intake Assessment: Methods for Adults

Helen Smiciklas-Wright, Diane C. Mitchell, and Jenny H. Ledikwe

Introduction
There has been longstanding interest in assessing diets of individuals.1,2 Early in the 20th
century, nutritionists studied food and nutrient intakes in order to provide guidance in
food selection,3,4 to interpret clinical and laboratory findings,5 and to establish dietary
requirements.6,7 Interest in dietary assessment was stimulated in the latter part of the
century with the increasing evidence for the role of diet in promoting health and reducing
chronic disease risk.8-11
Early investigators were concerned with many of the issues that continue to be important
for dietary assessment:

1. Selecting appropriate methods for collecting dietary data5,12-15

2. Assessing the day-to-day variability of intakes by individuals12,16,17
3. Establishing procedures for data analysis18-20
4. Estimating food/food group and nutrient intake5,21

The following statement appears in the National Research Council’s report on diet and
chronic disease: “One of the most difficult tasks in nutrition research is documenting the
actual or habitual food and nutrient intake of individuals or groups.”8 We should not be
surprised that obtaining information about what individuals consume and analyzing for
dietary components is a challenging undertaking. Food intake can be a complex behav-
ior.22 In any day, an individual may consume many different foods, at several eating
occasions, both at home and away from home. Willingness and ability to report what is
consumed can be influenced by social and environmental events and cognitive abili-
ties.23-25 Furthermore, food composition databases must continuously be updated to reflect
an expanding food marketplace and an increasing number of dietary components asso-
ciated with health.26
This section is organized to review the methods most commonly used to assess intakes
by individuals. Attention is given to methodologic validity as well as to the current
emphasis on food pattern analysis, dietary supplements, and functional foods.

© 2002 by CRC Press LLC

2705_frame_C19 Page 478 Wednesday, September 19, 2001 1:23 PM

478 Handbook of Nutrition and Food

Methods of Dietary Assessment

The common methods for assessing intakes by individuals are food records, recalls, and
food frequencies/diet histories. There have been several reviews of dietary assessment
methods, their appropriate uses, modes of administration, and sources of error.1,27-31 Meth-
ods are generally characterized either by the reference period in which respondents are
asked to provide information (i.e., retrospective and prospective methods)31 or by the time
frame for which data are collected (i.e., quantitative daily and food frequency methods).27,28,30
There is no single optimal dietary assessment method. The objectives of an assessment
should be used as a guide in selecting the most appropriate method. Some 30 years ago,
Christakis advised that the assessment method selected should be no more detailed, no
more cumbersome, and no more expensive than necessary.32 This advice is still sound.31
Assessment protocols, regardless of method, may need to provide highly quantitative and
detailed data on food consumption. This would be the case for research studies such as
clinical trials.33 More qualitative data is likely to be appropriate when food intake infor-
mation is used for dietary guidance and counseling.34

Food Record
Food records, also known as food diaries, provide a prospective account of foods and
beverages consumed in a defined period of time. Generally, records are kept for brief time
periods (one to seven days),28 but they have been kept for up to a month35 and even a
year.36 To be representative of usual intake, multiple days of records are needed.27,37
Food records may be used to meet a variety of objectives. Records are useful for detecting
imbalances in food intake and making dietary change recommendations.38 They are used as
self-management tools in weight loss interventions and may be valuable in predicting suc-
cessful weight loss.39 Food records have been used extensively to calibrate other dietary
assessment methods.40-45 Records are also useful for documenting compliance of an individ-
ual’s food intake with a feeding protocol in studies where adherence to a specific diet regimen
is important.46 Intervention studies may use food records to document effectiveness.47-49
Food portions may be either weighed or estimated depending on the subjects and the
purpose(s) of the assessment.50 While weighing foods will increase the accuracy of the
portion size recorded, it can also increase respondent burden. Sophisticated scales that do
not disclose food weights are available, decreasing respondent recording burden51 but
increasing cost. A variety of portion size aids, listed in Table 19.1, are available when
portion sizes are to be estimated.
While records are usually kept by respondents, they may also be kept by observers.
When food intake is recorded by observation, trained personnel visually estimate dietary
intake.52,53 Observation is particularly useful when circumstances preclude self-reporting
of food intake. Thus, observation has been used in assessing intake of nursing home
residents. The Omnibus Budget Reconciliation Act54 requires that all Medicare and Med-
icaid certified nursing facilities implement a standardized comprehensive assessment,
including a measure of nutrient intake, for all residents. Observers visually estimate the
portion of served items consumed (i.e., from all to none) by a resident.55
When using food records, consideration should be given to the record forms to be used
as well as instructions and training for subjects, particularly regarding portion size. Instruc-
tions should include guidance on completing the record form as well as directives encour-
aging subjects to record foods at the time of consumption and not to alter normal eating
patterns. Table 19.2 provides sample instructions for the administration of a food record.

© 2002 by CRC Press LLC

2705_frame_C19 Page 479 Wednesday, September 19, 2001 1:23 PM

Dietary Intake Assessment: Methods for Adults 479

TABLE 19.1
Tools for Portion Size Estimation
Type Examples
Household measures Measuring cups and spoons161
Rulers161
Food models Food replicas165
Graduated food models166
Thickness sticks161
Pictures 2-Dimensional portion shape drawings124,161
Portion photos of popular foods167
Portion drawings of popular foods161
Food labels Nutrition facts label
Food package weights

TABLE 19.2
Sample Instructions for the Administration of a Food Record
To help us do the best analysis of your food intake, please follow these instructions.

Maintain Your Usual Eating Pattern. Try not to modify your food intake because you are keeping a record.
Record Everything You Eat or Drink. Be sure to include all snacks and drinks. Also include any vitamin or
mineral supplements and the dosage for each day.
Write Foods Down As Soon As You Eat Them. Three daily record pages are provided for each day; however,
you may not need to use all three. Try to write clearly.

Details are Important!

Completing the food record form

Date. Please record the date at the top of each form.

Name. Please write your name in the space at the top of the form.
Time of Day. Record the time of the day you ate each meal, including AM or PM.
Meal/Where Prepared? Record the name of the meal eaten (i.e., breakfast, lunch, dinner, supper, or snack) and
where the meal was prepared (i.e., at home, at a restaurant).
Food Item. Write the name of each food item eaten.
Description/Preparation. Include information on how each food was prepared.
Amount. Record the amount of each food either by using the poster provided or common household measures.

After records have been completed, they should be reviewed to ensure completion. If
reviewed with the subject, probing questions may be used to clear up ambiguities and
ensure the completeness of the record. This is termed interviewer-assisted food records.
When data from food records are compiled and analyzed, records will need to be coded
using a standard method.
As subject burden can be high for food records, participant willingness and abilities are
considerations when using this method. Literacy is required for completion of records;
therefore, the method may not be appropriate for all individuals. The act of keeping food
records can affect dietary intake,56,57 which may be critical for estimates of usual intake.
Cost is an additional consideration, as reviewing records for completeness, data entry, and
data analysis can be expensive.31

Recall
Dietary recall provides a retrospective record of intake over a defined time period. While
dietary recall may be for any length of time, this method is almost always administered

© 2002 by CRC Press LLC

2705_frame_C19 Page 480 Wednesday, September 19, 2001 1:23 PM

480 Handbook of Nutrition and Food

to cover a 24-hour time period and is generally termed the 24-hour recall.30 Data can be
collected either for the previous day or for the 24 hours preceding the interview. To
estimate the usual intake of individuals, multiple recalls are needed, preferably on random,
nonconsecutive days.36,58 Typically, an individual is asked to recall all foods eaten during
the reference time period, describe the foods, and estimate the quantities consumed.
The 24-hour recall has become a favored way of collecting dietary data,33,47,59 as recalls
can be administered easily and quickly with low respondent burden. Depending on the
objectives of the recall, the amount and depth of information collected will vary. This
method is becoming the gold standard, particularly as methodological improvements60-62
and technological capabilities63,64 increase validity. With the emergence of technological
aids in dietary assessment, it is becoming more common for interviewers to collect intake
data using interactive software, entering intake data directly into a computer as it is
collected.
Recalls may be obtained either in person or by telephone-administered interview.
Because recalls by telephone interview have been shown to be practical, valid, and cost-
effective,63-68 they are becoming an increasingly popular mode of data collection, especially
for research purposes.
Prior to conducting recalls, training of interviewers is important. This is particularly
relevant when more quantitative data are required, increasing the need to use multiple
pass and probing techniques. Figure 19.1 provides a sample probing sequence to elicit
detailed information regarding one specific food (i.e., macaroni and cheese). The complex-
ity of this probing sequence exemplifies the potentially complex nature of probing ques-
tions and the need for good interviewer training. More qualitative food intake data can
be achieved with more limited questions.
The 24-hour recall has been criticized because of accuracy related to portion size esti-
mation and subject memory. Portion size estimation aids are available to facilitate quantity
estimation (Table 19.1). While 24-hour recalls are not designed to affect the “encoding” of
food information, they can incorporate strategies to facilitate memory retrieval. Those
strategies include standardized data collection protocols, structured probes to ensure
standardized collection, and interactive interview systems.29,47 The multiple pass tech-
nique, which will be discussed later in this section, has also been designed to facilitate
memory retrieval.

Food Frequency Questionnaire

Food frequency questionnaires (FFQs) are designed to obtain information about usual
food consumption patterns. They provide estimates on intake over a specified time period,
ranging from as little as a week69 to as much as one year.70 FFQs consist of a list of foods
and frequency-of-use response categories. Questionnaires may also include portion size
response categories. Food lists may be extensive in order to provide estimates of total
intake or they may be focused on foods, groups of foods, or nutrients. Several nutrient
specific FFQs have been developed which allow for the examination of selected nutrients
such as fat,71 vitamin A,72 and vitamin B6.73 While not necessarily appropriate for identi-
fying precise nutrient intake, these instruments can provide a rapid, cost-effective way to
estimate an individual’s usual intake.74
Questionnaires may be abbreviated when used to screen for nutritional risk. Screening
instruments are typically brief, self-administered, and can be scored quickly, providing an
efficient way to monitor eating patterns of individuals. Examples of screening tools include
the instruments developed as part of the Nutrition Screening Initiative, which were
designed to identify older adults who may need nutrition services and to provide diag-

© 2002 by CRC Press LLC

2705_frame_C19 Page 481 Wednesday, September 19, 2001 1:23 PM

Dietary Intake Assessment: Methods for Adults 481

Macaroni & Cheese

What type of macaroni and cheese was it?

Plain
With egg
With frankfurters
How was it prepared?

Canned or frozen
Prepared from a mix
Prepared from a recipe

Dried cheese sauce

What type of fat was used Prepared cheese sauce What type of cheese sauce
in the mix? Unknown type of cheese was in the mix?

What kind of butter Was the butter

was used in the mix? salted or unsalted?
Butter Regular Salted
Margarine Whipped Unsalted
Oil Light Unknown
Shortening Butter/margarine
Lard Unknown
Animal Clarified or butter oil
No fat used
Unknown if fat used
Unknown type of fat used What type of milk
was used in the mix?

How much did

Whole
you eat?
2% fat
AMOUNT
11 2% fat
CONSUMED
1% fat
1 2 % fat

skim, nonfat, or fat-free

lactose reduced (Lactaid)
Unknown

FIGURE 19.1
A sample probing scheme. This scheme could be used with recalls to elicit more information from a respondent
who consumed macaroni and cheese. Bold print indicates respondent’s reply. Probing questions, which are
specific for each response, are italicized. (Adapted from Nutrition Data System for Research [NDS-R] software,
developed by the Nutrition Coordinating Center [NCC], University of Minnesota, Minneapolis, MN.)

© 2002 by CRC Press LLC

2705_frame_C19 Page 482 Wednesday, September 19, 2001 1:23 PM

482 Handbook of Nutrition and Food

nostic information on nutrition status.75,76 Very abbreviated instruments may not be valid
representations of true intake.77
Creating a food frequency questionnaire is an intensive process,78 thus, there is heavy
reliance on instruments that have already been developed. Validity is a critical issue and
is generally determined by calibration with other assessment methods.73,79-84 When using
an FFQ, it is important to ensure that the questionnaire is valid for use in the population
of interest, as the performance of an instrument may vary between subgroups. Question-
naires have been validated in specific populations such as adolescents,85 pregnant
women,42 and low-income black women.86
Several cognitive issues could compromise the validity of an FFQ. Subjects may have
difficulty recalling foods consumed over a lengthy time period. Additionally, participants
may need to perform arithmetic computations to average usual consumption of foods to
fit into the response categories for consumption frequency. The cognitive demands of the
FFQ may be reduced by a new variation in questionnaire administration, termed the
picture sort approach, in which participants sort food cards into categories.87,88
While FFQs generally provide qualitative data, if portion size estimation is included on
the questionnaire, semi-quantitative information can be deduced from these instruments.
In some cases, inclusion of portion size may yield only small differences in data as
compared to FFQs analyzed using only medium portions.89

Diet History
Food frequency questionnaires are sometimes referred to as diet histories. The classic diet
history method was designed to estimate an individual’s usual intake over a relatively
long period of time. Originally developed by Burke,90 the method consists of several
components, including a 24-hour recall and questions about usual eating patterns, a cross-
check questionnaire with a list of foods and questions about likes, dislikes, and usages,
and a three-day food record. This method is not used commonly today. Administration
requires a highly trained dietitian and can be time consuming.

Summary
Several methods have been listed which can be used to assess the usual dietary intake of
adults. Well-designed quality control procedures are particularly important in research
studies to ensure consistency of data across time and subjects.33 Additionally, the advan-
tages and disadvantages of data collection modes (i.e., self- or interviewer-completed)
should be considered (See Table 19.3). The success of any assessment method is based
upon a partnership between the individual respondent and the assessment staff. Care
should be taken to ensure the appropriateness of the method and the level of detail
collected. Respecting participants’ abilities and ensuring that their dignity is not compro-
mised is salient in establishing a successful partnership.

Issues Affecting Validity

In his address at the First International Conference on Dietary Assessment Methods,
Beaton stated, “In the past decade there has been a great deal published about the errors

© 2002 by CRC Press LLC

2705_frame_C19 Page 483 Wednesday, September 19, 2001 1:23 PM

Dietary Intake Assessment: Methods for Adults 483

TABLE 19.3
Self-Completed and Interviewer-Completed Data Collection
Collection Mode Advantages Disadvantages
Self Interviewer training not needed Response rate may be low
Sense of privacy Respondent burden may be high
Data collection time may be reduced Tasks may be misinterpreted
Respondent training needed for more
complete data
Data preparation and entry time may
be high
Interviewer by phone Good Response rate Contact times may be inconvenient
Opportunities for probing Hearing problems for some subjects
Low respondent burden Availability of portion aids
Relatively quick Data collection may be more
Interviewer anonymity expensive for toll calls
Potential for interviewer bias
Interviewer in person Good response rate Contact times may be inconvenient
Opportunities for probing Potential for interviewer bias
Low respondent burden
Respondent interviewer rapport

in dietary data…this is understandable, but unfortunate because it can easily leave the
impression that dietary data are worthless.”91 He reminded his audience that, while dietary
intake data cannot and never will be estimated without error, a serious limitation is not
the errors themselves, but failure to understand the nature of the errors and the consequent
impact on data analysis and interpretation. Several recent reviews have delineated poten-
tial sources of error for different assessment methods.28,31,51 Consideration of strategies to
minimize error is pertinent in yielding accurate intake data, regardless of assessment
method (see Table 19.4).
Recent attention to the accuracy of dietary intake data has focused on the underreport-
ing of energy by 10 to almost 40%.59,92-97 These findings are based on an extensive literature
comparing intakes to energy needs estimated using doubly-labeled water,98 weight main-
tenance data,59 and applying age- and sex-specific equations to estimate energy require-
ments.99 Underreporting has been found to be more common among women99-101 and
older persons99,100,102 as well as overweight,99-101 post-obese,92 and weight-conscious indi-
viduals.99,100 Literacy103 and depression104 have been associated with underreporting. Selec-
tive underreporting has also been associated with certain food types such as fats95,99,105
and sweets.105
While underreporting of energy is common in groups of individuals,59,94,95,99,100,106-108 both
underreporting and overreporting can occur.44,59,100 Individuals who possess characteristics

TABLE 19.4
Benefits Derived from Minimizing Assessment Error
Clinical Setting Research Setting
Improve ability to detect inadequate, imbalanced, or Improve accuracy of nutrient intake estimations
excessive dietary intake
Provide a better basis for nutrition counseling and Decrease attenuation between intake data and
interventions biomarkers
Improve ability to monitor dietary changes Provide a better basis for nutrition education program
Provide a better basis for elucidation of diet-disease
relationships

© 2002 by CRC Press LLC

2705_frame_C19 Page 484 Wednesday, September 19, 2001 1:23 PM

484 Handbook of Nutrition and Food

associated with underreporting may actually report intake accurately or overreport intake.
However, the magnitude of underreporting may be even greater for individuals as error
due to overestimation can reduce underestimation bias in groups.102
Due to the complex nature of intake data and the variability of under and over reporting,
it is unlikely that a single correction factor will be derived that could be applied to self-
reported energy intake.109 The purpose of this section is to review components of assess-
ment that may be modified to reduce sources of error.

Memory
Food recall is a cognitively demanding task. Understanding of dietary recall accuracy is
derived from advances in cognitive psychology.23 Classic work in this area described
memory processes: encoding or learning information, transmission to long-term memory,
and retrieval.110-112 Early studies described strategies for encoding information as well as
strategies for retrieving memories, such as free recall, recognition, and cued recall.
The memory model of cognitive psychology is applicable to dietary recall.25 To accurately
report intake, people must be able to remember what foods were consumed, how the
foods were prepared, and the quantities of foods eaten. This requires the acquisition of
specific food memories and the ability to retrieve these memories. Individuals who pay
little attention to foods consumed, people who have difficulty storing information in
memory, and those who lack the cognitive ability to retrieve food memories may not be
able to accurately recall dietary intake.
Several techniques have been developed to reduce memory-related error in dietary data.
For the 24-hour recall, techniques such as probing (See Figure 19.1),113 encoding strategies,25
memory retrieval cues,25 and a multiple pass system59,62 have been employed to improve
memory. Campbell and Dodd’s113 classic paper showed that probing elicited additional
information with significant impact on total caloric intake. Ervin and Smiciklas-Wright
found that older adults were able to remember more foods when a deeper processing
strategy was used during encoding and a recognition task was used for memory retrieval.25
Record-assisted recalls may be used to help reduce memory-related error in food
records.114,115
Recent work suggests that 24-hour recalls which incorporate a multiple pass technique
into a standardized interview protocol with structured probes can reduce the commonly
observed underestimation of intake for groups of individuals.62 A multiple pass technique
provides respondents several opportunities (i.e., passes) to recall foods eaten using both
free recall and cued (probed recall) strategies.59,62,116,117 As generally administered, the
strategy involves three recall passes: an introductory opening sequence in which a respon-
dent is asked to recall all items eaten, an interactive, structural probe sequence to elicit
food descriptions and amounts, and a final review of the recall. The multiple pass tech-
nique is theoretically sound,117 and when incorporated into a well-structured interactive
interview process may decrease underreporting for groups of individuals.59,62 While these
studies are encouraging for the presentation of group data, there is room for improvement
in assessing individuals’ intakes.59 Little data exists, however, on alternative modes of
administering a multiple pass strategy and the “gains” at each pass.
A multiple pass technique can be facilitated by the use of interactive software.118 This
allows for a greater level of detail and facilitates data collection, but the technology is
generally expensive and is not used commonly in clinical settings. However, written
tools, such as probing guides, may be used to mimic this process when quantitative
analysis is critical.

© 2002 by CRC Press LLC

2705_frame_C19 Page 485 Wednesday, September 19, 2001 1:23 PM

Portion Sizes
It is well documented that individuals have difficulty estimating amounts of foods and
beverages.61,119-121 There is a tendency toward overestimation of smaller portion sizes and
underestimation of larger portion sizes which can lead to the “flat slope syndrome.”121,122
Portion size estimation aids (Table 19.1) have been shown to reduce portion size estimation
error.123 Estimation aids vary in sophistication and cost. Choice of tools is dictated partially
by feasibility. In a clinical setting, aids such as food replicas, real foods, and food picture
books may be appropriate. For interviews conducted by phone, tools that are compact for
mailing, such as a chart with two-dimensional portions,124 would be more appropriate.
A number of investigators have investigated whether training subjects to “judge” por-
tion sizes improves quantity estimates.61,125-130 These studies suggest that training effects
may be retained for some days after training and may have significant impact on some,
but not all foods. For example, amorphous foods (e.g., salads) are more resistant to training
effects.

Variability of Intake
Day-to-day variation of food intake has been well documented in the literature.36,131,132
Accordingly, assessment of an individual’s total dietary intake, particularly by quantitative
daily methods, at any one time may not yield an accurate measure of usual intake.36
Basiotis et al. found that over 100 days of dietary data may be needed to accurately estimate
an individual’s typical intake for certain nutrients, such as vitamin A.133 To lessen the effect
of day-to-day dietary variation when using 24-hour recalls, assessment should be done
on multiple, random, nonconsecutive days36,58,134 that include both weekend and week-
days. For food records and 24-hour recalls, increasing the number of assessment days will
decrease error related to variation in food intake; however, this must be balanced with
subject tolerability and assessment objectives.

Consumption Frequency
Accurate estimation of how often foods are consumed is particularly important for retro-
spective assessment methods. For food frequency questionnaires, frequency of consump-
tion estimates may contribute more error than portion size estimates.135 The cognitive
demands required to mathematically calculate consumption frequency contribute to the
error involved with this measure. It has been suggested that the precision of food frequency
questionnaires can be increased by not using predefined consumption frequency catego-
ries, such as three to four times per week, instead allowing participants to simply enter
a number to reflect intake.135 Ability to accurately recall the frequency of consumption of
foods deteriorates as the amount of time between intake and assessment increases, yet
longer reference time periods yield more accurate results than questionnaires with shorter
reference periods.136

Response Bias
All assessment methods are subject to response bias. Social desirability may lead some
individuals to selectively omit foods that may be regarded as unacceptable (e.g., alcohol,
high fat foods),23 while others may report eating a healthier diet than that which was

© 2002 by CRC Press LLC

2705_frame_C19 Page 486 Wednesday, September 19, 2001 1:23 PM

486 Handbook of Nutrition and Food

actually consumed.137,138 Self-reported assessment data may also be biased by participation

in a dietary intervention.139
For both interviewer-assisted and self-completed assessment, questions should be
reviewed for face validity to help ensure that participant comprehension of the questions
is appropriate. When using interviewers to collect data, training to avoid leading questions
and verbal and nonverbal cues that may appear to be judgmental can decrease response
bias. Quality control procedures can ensure that interviewer questioning is consistent and
nonbiasing.140,141 Conducting interviews by telephone may reduce bias compared to face-
to-face interviews.63
In regard to particular assessment methods, food frequency questionnaires may be
subject to response bias, as current diet may influence recall of dietary intake in the past,142
especially for individuals with diet-related illnesses.143 Response bias can also be induced
by methods with a high participant burden. For example, the burden of keeping food
records may lead subjects to submit incomplete records, introducing a response bias.144
Techniques that reduce respondent burden, such as interviewer-assisted food records, can
reduce this effect and may improve the quality of data collected.

Data Entry
Data entry is the link between the information provided by a respondent and analysis of
the data. Data entry often requires decisions by coders to adjust information provided to
meet the demands of a specific data analysis program.145 If the respondent provides
incomplete data or the database does not include all diet items, coders must decide on
reasonable substitutions for portion size or food items. Thus, the quality of the data
provided by respondents, the quality of the database, and the default assumptions by
coders can all contribute to variability in the final food and nutrient data descriptions.145
Decisions about amounts of foods eaten may be guided by U.S. Department of Agricul-
ture (USDA) publications on portions commonly consumed in the United States. The
USDA has published several reports on foods commonly eaten and the quantities con-
sumed at an eating occasion.146,147 These reports provide data on amounts eaten by par-
ticipants in nationwide food consumption surveys.

Food Composition Tables

Food composition databases provide values for the nutritional content of foods. Errors in
these databases will introduce systematic biases during data analysis. Further research to
improve the validity of the nutrient values within food composition databases will further
increase the accuracy of any dietary assessment methodology. This topic will be covered
in more detail in Section 23.

Summary
Dietary assessment is a dynamic field, with novel approaches being developed, such as
computer-assisted self-interviews.148 Various techniques to improve validity have been
developed (See Table 19.5), but much work still needs to be done to decrease both sys-
tematic and random errors. Refinement of current methods and the development of new
techniques will improve confidence in the accuracy of dietary data.

© 2002 by CRC Press LLC

2705_frame_C19 Page 487 Wednesday, September 19, 2001 1:23 PM

Dietary Intake Assessment: Methods for Adults 487

TABLE 19.5
Considerations to Reduce Error when Collecting Assessment Data
Potential Food Frequency
Error Sources 24-Hour Recalls Food Records Questionnaire
Memory Multiple pass technique Interviewer assisted records Memory retrieval techniques
Probing questions Encouraging adherence to
Encoding strategies appropriate instructions
Portion size Portion size estimation aids Portion size estimation aids Portion size estimation aids
Subject training Weighing scale Subject training
Interviewer training Subject training
Day-to-day Multiple recall days Multiple days of records N/A
variability Nonconsecutive days of data Include weekdays &
Include weekdays & weekends
weekends Collect data in different
Collect data in different seasons
seasons
Response bias Interviewer training Objective instructions Objective responses for
Clearly worded, open-ended Reduce respondent burden interviewer mode
questions Limit days of data collection
Objective interviewer
responses
Recalls on unannounced,
random days
Data entry Documentation of decisions Strict coding and data entry Strict coding and data entry
Interactive software with rules rules
detailed probes and Interviewer assisted records Computer scannable forms
automatic coding Detailed probing guides and for automatic coding
instructions

Current Issues in Assessment and Analysis

Dietary Quality Scores and Food Pattern Analysis
Historically, dietary quality scores have been a method of interpreting nutrient intakes of
an individual by comparison with a dietary standard such as the Recommended Dietary
Allowances149 or, more recently, the Daily Reference Intakes.150 The nutrient adequacy ratio
(NAR), the amount of a particular nutrient in the diet divided by the dietary standard,
has been commonly used for a number of years.30,151 A mean adequacy ratio (MAR) can
also be calculated and represents the mean of the NAR for several nutrients of interest.30
Another way to examine the quality of an individual’s diet is to calculate the number of
servings from each food group and compare this with food grouping standards such as
the USDA Food Guide Pyramid.152 However, depending on the food group scheme used,
serving sizes and placement of foods into groups may vary considerably. Approaches to
food group analyses also differ.153 For individuals, a behavioral approach, in which food
group changes are based on nutrition education strategies, is more appropriate. This
approach usually defines specific nutritional education programs for health promotion,
such as the National Cancer Institute’s 5 A Day Program.
With the introduction of the USDA Food Guide Pyramid,152 other dietary quality scores
have been developed which take into account not just nutrients or food group servings
alone, but an aggregation of these two assessments into one score. The Healthy Eating

© 2002 by CRC Press LLC

2705_frame_C19 Page 488 Wednesday, September 19, 2001 1:23 PM

488 Handbook of Nutrition and Food

Index score, created by the USDA, takes into account servings from the major food groups
of the Food Guide Pyramid as well as a dietary variety and health risk-related nutrient
intakes.154 This more comprehensive approach, as well as others that have been devel-
oped,155 is becoming an increasingly popular approach to interpretation of the dietary
intakes of groups or individuals. According to a recent article, identifying food patterns
in groups and in individuals might be a better determinant of risk for disease as well as
mortality.156 This is based on the premise that diets are not comprised of single nutrients
or foods but combinations of nutrients and nonnutrient components. The interactions of
all these dietary components make it difficult to determine the effects of single dietary
components. Additionally, dietary behaviors are complex and many different patterns of
intake may be occurring simultaneously, such as decreasing fat intake while increasing
fruits and vegetables.
Any method of assessing an individual’s dietary intake is dependent on the methods
of interpretation. More comprehensive methods of interpretation may facilitate the iden-
tification of more specific patterns of intake and their relationship to disease.

Dietary Supplements
Increased nutrient intakes from supplements have been related to certain disease risk such
as cardiovascular disease (vitamin E), neural tube defects (folate), osteoporosis (calcium),
and cancer (antioxidants such as vitamin C and beta-carotene). Approximately 40 to 50%
of the general population over the age of two years in the U.S. takes a dietary supplement
(see Table 19.6).157,158 These numbers continue to rise especially for more non-traditional
or complementary therapies such as herbal or botanical supplements. Supplement use is
higher in non-Hispanic white females, and increases with age in some segments of the
population.157,159 The supplement intake of special populations such as individuals with
cancer diagnosis are much higher, up to 80% in some studies.159,160 Knowledge of dietary
supplements as well as an understanding of assessment methods is critical to the overall
assessment of nutrient and other dietary components.
Supplement intake data can be assessed in a variety of ways and is usually collected
by questionnaire, including food frequency questionnaires, or as part of intake data such

TABLE 19.6
Categories of Supplements
Category Examples
Vitamins Vitamin C, E, D, B6
(single or multiple formulations)
Minerals Iron, calcium, chromium, zinc
(single or multiple formulations)
Vitamin(s) with mineral(s) Calcium with vitamin D; vitamin E with selenium
Herbs and other botanicals St. John’s Wort, ginkgo biloba, ginseng, saw palmetto
Flavonoids Quercetin, rutin, hesperidin, diadzin
Carotenoids Lycopene, zeaxanthin, lutein, dried carrot extract,
other vegetable extracts
Fatty acids/fish oils, other oils Linoleic acid, omega 3 fatty acids, DHA EPA
Amino acids/nucleic acids/proteins including co- L-glutamine, coenzyme Q-10, bromelain, tryptophan
enzymes, enzymes & hormones
Microbial prepartions/probiotics Lactobacillus acidophilus, B. bifidus, L. bulgaris
Glandular and other organ preparations Dessicated glands such thyroid and adrenal
Miscellaneous Shark cartilage, pycogenol, chrondoitin sulfate

© 2002 by CRC Press LLC

2705_frame_C19 Page 489 Wednesday, September 19, 2001 1:23 PM

Dietary Intake Assessment: Methods for Adults 489

as by 24-hour dietary recall or by food records. When collected by the latter methods it
is important to recognize that the intake of supplements for the day of data collection may
not reflect the pattern of intake over an extended period of time. Detailed questionnaires,
which are better for capturing long-term intake and frequency of intake, are used fre-
quently in research studies, clinical practice, and nutrition monitoring and surveys.161,162
Quantifying supplement intake is a difficult and tedious process. When collecting sup-
plement information, it is important to identify what level of detail is needed to describe
or quantify dietary intakes with dietary supplements. Strategies may include having those
individuals bring in their supplement labels, or photocopy the labels. Other strategies
include having the participants respond to questionnaires that provide lists of single
vitamins and minerals as well as common brand names for multiple formulations. For
herbal and botanical ingredients and other components not typically found in common
formulations, it might be necessary to identify the active components and, above all else,
to obtain brand name and label information.

Functional Foods
According to the Institute of Medicine of the National Academy of Sciences (1994) the
definition of functional foods is “any food or food ingredient that may provide a health
benefit beyond the traditional nutrients it contains.”163 Currently, there is a major research
emphasis to identify physiologically active components, or phytochemicals, and their
potential for decreasing disease risk. Functional foods can consist of foods or food ingre-
dients, including fruits and vegetables to more specialized products such as those contain-
ing soy or phytostanols (i.e., margarines with claims of reducing blood cholesterol levels).
Since many of the phytochemicals in foods are still under investigation, it is premature
to emphasize quantification of single functional food components in dietary assessment.
However, improvements in individual assessment methods will make it possible to quan-
tify and link these components in functional foods with their potential health benefits.
Examples of where individual assessment plays a significant role in providing this linkage
can be found in the literature examining fruit and vegetable intakes. There are now
databases that can accurately quantify the carotenoid content of foods.118,164 As cooking
methods, storage, and exposure to air and water are known to affect the carotenoid content
of fruits and vegetables, the development of these databases has been a challenge. In the
case of assessing carotenoids, for example, it is important to distinguish between pink
and white grapefruit (i.e., pink grapefruit has 3740 mg lycopene, whereas white grapefruit
has 0 mg). This one simple observation in assessing an individual’s intake can have a
significant impact on determining the relationship between dietary carotenoids and blood
carotenoids and the potential role they may have in cancer risk in groups of individuals.
In assessing diets for research purposes it is important to consider the level of detail
required for other components in foods as well. The assessment issues outlined in this
section are the same for assessing functional foods. As more components are identified
and quantification is possible, databases need to be developed that make analysis of
functional foods and their components possible. The knowledge gained in the study of
functional foods will drive the type and level of detail in methodology and database
development.

© 2002 by CRC Press LLC

2705_frame_C19 Page 490 Wednesday, September 19, 2001 1:23 PM

490 Handbook of Nutrition and Food

References
1. Bingham S. Nutr Abstr Rev (Series A) 57: 705; 1987.
2. Medlin C, Skinner JD. JADA 88: 1250; 1988.
3. Mudge GG. J Home Ec 15: 181; 1923.
4. Mudge GG. JADA 1: 166; 1926.
5. Turner D. JADA 16: 875; 1940.
6. Widdowson EM. J Hygiene 36: 269; 1936.
7. Widdowson, EM. J Hygiene 36: 293; 1936.
8. National Research Council, Diet and Health: Implications for Reducing Chronic Disease Risk,
National Academy Press 46; 1989.
9. Public Health Service, Healthy people: Surgeon General’s Report on Health Promotion and Disease
Prevention, US Department of Health and Human Services, Washington DC; 1979.
10. Public Health Service, Promoting Health/Preventing Disease: Objectives for the Nation, US Depart-
ment of Health and Human Services, Washington DC; 1980.
11. US Department of Health and Human Services, Healthy People 2010 (Conference Edition), US
Department of Health and Human Services, Washington DC; 2000.
12. Leverton RM, Marsh AG. J Home Ec 31: 111; 1939.
13. Huenemann RL, Turner D. JADA 18: 562; 1942.
14. Young C, et al. JADA 28: 124; 1952.
15. Young C, et al. JADA 28: 218; 1952.
16. Wait B, Roberts LJ. JADA 8: 323; 1932.
17. Yudkin J. Brit J Nutr 5: 177; 1951.
18. Hunt CL. J Home Ec 5: 212; 1918.
19. Burke BS, Stuart HC. J Pediatr 12: 493; 1938.
20. Donelson EG, Leichsenring JM. JADA 18: 429; 1942.
21. Tigerstedt R. Skand Arch Physiol 24: 97; 1910.
22. Blundell JE. Am J Clin Nutr 71: 3; 2000.
23. Dwyer JT, Krall EA, Coleman KA. JADA 87: 1509; 1987.
24. Smith AF. Eur J Clin Nutr 47: S6; 1993.
25. Ervin RB, Smiciklas-Wright H. JADA 98: 989; 1998.
26. Roberfroid MB. J Nutr 129: 1398S; 1999.
27. Life Sciences Research Office, Guidelines for Use of Dietary Intake Data, Federation of American
Societies for Experimental Biology, Bethesda, MD; 1996.
28. Gibson RS. Principles of Nutritional Assessment, Oxford University Press, New York; 1990.
29. Thompson FE, Byers T. Dietary assessment resource manual, J Nutr 124: 2245S; 1994.
30. Smiciklas-Wright H, Guthrie HA. Nutrition Assessment: A Comprehensive Guide for Planning
Intervention, 2nd ed, Simko MD, Cowell C, Gilbride, JA Eds, Aspen Publishers, Gaithersburg,
MD: 165; 1995.
31. Dwyer J. Modern Nutrition in Health and Disease, 9th ed, Shils ME, Olson JA, Shike M, Ross
AC, Eds, Williams & Wilkins, Philadelphia: 937; 1999.
32. Christakis G. Am J Public Health 63: 1S; 1973.
33. Copeland T, et al. JADA in press.
34. Olendzki B, et al. JADA 99: 1433; 1999.
35. St. Jeor ST, Guthrie HA, Jones MB. JADA 83: 155; 1983.
36. Tarasuk V, Beaton GH. Am J Clin Nutr 54: 464; 1991.
37. Craig MR, et al. JADA 100: 421; 2000.
38. Tian HG, et al. Eur J Clin Nutr 49: 26; 1995.
39. Streit KJ, et al. JADA 91: 213; 1991.
40. Achterberg C, et al. J Can Diet Assoc 52: 226; 1991.
41. Block G, Hartman AM, Naughton D. Epidemiology 1: 58; 1990.
42. Brown JE, et al. JADA 96: 262; 1996.
43. Cummings SR, et al. Am J Epidemiol 126: 796; 1987.

© 2002 by CRC Press LLC

2705_frame_C19 Page 491 Wednesday, September 19, 2001 1:23 PM

Dietary Intake Assessment: Methods for Adults 491

44. Domel SB, et al. J Am Coll Nutr 13: 33; 1994.

45. Kristal AR, et al. Am J Epidemiol 146: 856; 1997.
46. Jackson B, et al. JADA 86: 1531; 1986.
47. Buzzard IM, et al. JADA 96: 574; 1996.
48. Gorbach SL, et al. JADA 90: 802; 1990.
49. Kuehl KS, et al. Prev Med 22: 154; 1993.
50. Moulin CC, et al. Am J Clin Nutr 67: 853; 1998.
51. Bingham SA, Ann Nutr Metab 35: 117; 1991.
52. Gittelsohn J, et al. JADA 94: 1273; 1994.
53. Dubois S. JADA 90: 382; 1990.
54. Omnibus Budget Reconciliation Act of 1990, P. L. 101-508; 1990.
55. Morris JN, et al. The Gerontologist 30: 293; 1990.
56. Rebro SM, et al. JADA 98: 1163; 1998.
57. Pekkarinen M. World Rev Nutr Diet 12: 145; 1970.
58. Larkin FA, Metzner HL, Guire KE. JADA 91: 1538; 1991.
59. Jonnalagadda SS, et al. JADA 100: 303; 2000.
60. Lyons GK, et al. JADA 96: 1276; 1996.
61. Howat PM, et al. JADA 94: 169; 1994.
62. Johnson RK, Driscoll P, Goran, MI. JADA 96: 1140; 1996.
63. Fox TA, Heimendinger, J, Block G. JADA 92: 729; 1992.
64. Derr JA, et al. Am J Epidemiol 136: 1386; 1992.
65. Casey PH, et al. JADA, 99: 1406; 1999.
66. Krantzler NJ, et al. Am J Clin Nutr 36: 1234; 1982.
67. Pao EM, Sykes KE, Cypel VS. USDA Methodological Research for Large Scale Dietary Intake Surveys,
1975-88, Home Economics Research Report no. 49, US Department of Agriculture, Human
Nutrition Information Service, US Government Printing Office, Washington, DC: 181; 1989.
68. Morgan KJ, et al. JADA 87: 888; 1987.
69. Cullen KW, et al. J Am Coll Nutr 18: 442; 1999.
70. Hartman AM, et al. Nutr Cancer 25: 305; 1996.
71. Retzlaff BM, et al. Am J Public Health 87: 181; 1997.
72. Sloan NL, et al. Am J Public Health 87: 186; 1997.
73. Brants HA, et al. Eur J Clin Nutr 51: S12; 1997.
74. Briefel RR, et al. JADA 92: 959; 1992.
75. White JV, Dwyer JT, Posner BM, et al. JADA 92: 163; 1992.
76. The Nutrition Screening Initiative, Incorporating Nutrition Screening and Interventions into Med-
ical Practice, The Nutrition Screening Initiative, Washington, DC; 1994.
77. Posner BM, et al. Am J Public Health 83: 972; 1993.
78. Subar AF, et al. FASEB J 14: A559; 2000.
79. Block G, et al. J Clin Epidemiol 43: 1327; 1990.
80. Longnecker MP, et al. Epidemiology 4: 356; 1993.
81. Musgrave KO, et al. JADA 89: 1484; 1989.
82. Willett WC, et al. JADA 87: 43; 1987.
83. Potischman N, et al. Nutr Cancer 34: 70; 1999.
84. Lund SM, Brown J, Harnack L. Eur J Clin Nutr 52: 53S; 1998.
85. Rockett HR, Wolf AM, Colditz GA. JADA 95: 336; 1995.
86. Coates RJ, et al. Am J Epidemiol 134: 658; 1991.
87. Kumanyika SK, et al. Am J Clin Nutr 65: 1123S; 1997.
88. Kumanyika S, et al. JADA 96: 137; 1996.
89. Laus MJ, et al. J Nutr Elder 18: 1; 1999.
90. Burke BS. JADA 23: 1041; 1947.
91. Beaton GH. Am J Clin Nutr 59: 253S; 1994.
92. Black AE, et al. Eur J Clin Nutr 51: 405; 1997.
93. Bandini LG, et al. Am J Clin Nutr 65: 1138S; 1997.
94. Champagne CM, et al. JADA 98: 426; 1998.
95. Goris AH, Westerterp-Plantenga MS, Westerterp KR. Am J Clin Nutr 71: 130; 2000.

© 2002 by CRC Press LLC

2705_frame_C19 Page 492 Wednesday, September 19, 2001 1:23 PM

492 Handbook of Nutrition and Food

96. Martin LJ, et al. Am J Clin Nutr 63: 483; 1996.

97. Kroke A, et al. Am J Clin Nutr 70: 439; 1999.
98. Schoeller DA, Fjeld CR. Annu Rev Nutr 11: 355; 1991.
99. Briefel RR, et al. Am J Clin Nutr 65: 1203S; 1997.
100. Johansson L, et al. Am J Clin Nutr 68: 266; 1998.
101. Stallone DD, et al. Eur J Clin Nutr 51: 815; 1997.
102. Black AE, et al. Eur J Clin Nutr 45: 583; 1991.
103. Johnson RK, Soultanakis RP, Matthews DE. J Am Diet Assoc 98: 1136; 1998.
104. Smiciklas-Wright H, et al. FASEB J 13: A263; 1999.
105. Bingham S. Am J Clin Nutr 59: 227S; 1994.
106. Champagne CM, et al. JADA 96: 707; 1996.
107. Kortzinger I, et al. Ann Nutr Metab 41: 37; 1997.
108. Mertz W, et al. Am J Clin Nutr 54: 291; 1991.
109. Schoeller DA. Metabolism 44: 18; 1995.
110. Wessells MG. Cognitive Psychology, Harper & Row, New York; 1982.
111. Craik FIM. Philos Trans R Soc Lond B Biol Sci 302: 341; 1993.
112. Schaie KW, Willis SL. Adult Development and Aging, 2nd ed, Schaie KW, Willis SL, Eds, Little,
Brown and Company, Boston; 324; 1986.
113. Campbell VA, Dodds ML. JADA 51: 29; 1967.
114. Eldridge AL, et al. JADA 98: 777; 1998.
115. Lytle LA, et al. JADA 93: 1431; 1993.
116. De Maio TJ, Ciochetto T, Davis W. American Statistical Association: Survey Methods: 1021; 1993.
117. Wright JD, Ervin RB, Briefel RR, Eds. Consensus Workshop on Dietary Assessment: Nutrition
Monitoring and Tracking the Year 2000 Objectives, Department of Health and Human Services,
National Center for Health Statistics, Hyattsville, MD, 1993.
118. Nutrition Coordinating Center, Nutrient Data System for Research (NDS-R) software, University
of Minnesota, Minneapolis, MN.
119. Blake AJ, Guthrie HA, Smiciklas-Wright H. JADA 89: 962; 1989.
120. Smiciklas-Wright H, et al. Progress Report 390, Northeastern Cooperative Regional Research
Publication, Pennsylvania State University, Agriculture Experiment Station: 1; 1988.
121. Young LRN. Nutr Rev 53: 149; 1995.
122. Faggiano F, et al. Epidemiology 3: 379; 1992.
123. Cypel YS, Guenther PM, Petot GJ. JADA 97: 289; 1997.
124. Nutrition Consulting Enterprises, Food Portion Visual, Nutrition Consulting Enterprises,
Framingham, MA, 1981.
125. Rapp SR, et al. JADA 86: 249; 1986.
126. Bolland JE, Yuhas JA, Bolland TW. JADA 88: 817; 1988.
127. Yuhas JA, Bolland JE, Bolland TW. JADA 89: 1473; 1989.
128. Bolland JE, Ward JY, Bolland TW. JADA 90: 1402; 1990.
129. Weber JL, et al. JADA 97: 176; 1997.
130. Slawson DL, Eck LH. JADA 97: 295; 1997.
131. Guthrie HA, Crocetti AF. JADA 85: 325; 1985.
132. McAvay G, Rodin J. Appetite 11: 97; 1988.
133. Basiotis PP, et al. J Nutr 117: 1638; 1987.
134. Hartman AM, et al. Am J Epidemiol 132: 999; 1990.
135. Flegal KM, et al. Am J Epidemiol 128: 749; 1988.
136. Smith A. Cognitive Processes in Long-Term Dietary Recall, DHHS Publication No. 92-1079, Series
6, No. 44, Department of Health and Human Services, National Center for Health Statistics,
Hyattsville, MD; 1991.
137. Hebert JR, et al. Int J Epidemiol 24: 389; 1995.
138. Hebert JR, et al. Am J Epidemiol 146: 1046; 1997.
139. Kristal AR, et al. J Am Diet Assoc 98: 40; 1998.
140. Edwards S, et al. Am J Epidemiol 140: 1020; 1994.
141. Smiciklas-Wright H, et al. JADA 81: 28S; 1991.
142. Dwyer JT, Coleman KA. Am J Clin Nutr 65: 1153S; 1997.

© 2002 by CRC Press LLC

2705_frame_C19 Page 493 Wednesday, September 19, 2001 1:23 PM

Dietary Intake Assessment: Methods for Adults 493

143. Malila N, et al. Nutr Cancer 32: 146; 1998.

144. Gersovitz M, Madden JP, Smiciklas-Wright H. JADA 73: 48; 1978.
145. Lacey JM, et al. Coder Variability in Computerized Dietary Analysis, Research Bulletin Number
729, Massachusetts Agricultural Experiment Station, Massachusetts; 1990.
146. Pao EM, et al. Foods Commonly Eaten by Individuals: Amounts Per Day and Per Eating Occasion,
Home Economics Research Report Number 44, Consumer Nutrition Center, Human Informa-
tion Service, Hyattsville, MD; 1982.
147. Krebs-Smith SM, et al. Foods Commonly Eaten in the United States: Quantities Consumed Per Eating
Occasion and in a Day, 1989-91, NFS Report No. 91-3, US Department of Agriculture, Agriculture
Research Service, Washington, DC; 1997.
148. Kohlmeier L, et al. Am J Clin Nutr 65: 1275S; 1997.
149. Food and Nutrition Board, Recommended Dietary Allowances, 10th ed, National Academy Press,
Washington, DC; 1989.
150. Food and Nutrition Board, Dietary Reference Intakes for Thiamin, Riboflavin, Niacin, Vitamin B6,
Folate, Vitamin B12, Pantothenic Acid, Biotin, and Choline, National Academy Press, Washington,
DC; 1998.
151. Guthrie HA, Scheer JC. JADA 78: 240; 1981.
152. US Department of Agriculture, The Food Guide Pyramid: A Guide to Daily Food Choices, US
Department of Agriculture, Nutrition Information Service; 1992.
153. Cullen KW, et al. JADA 99: 849; 1999.
154. Bowman SA, et al. The Healthy Eating Index: 1994-96, CNPP-5, US Department of Agriculture,
Center for Nutrition Policy and Promotion; 1998.
155. Haines PS, Siega-Riz AM, Popkin BM. JADA 99: 697; 1999.
156. Kant AK, et al. JAMA 283: 2109, 2000.
157. Ervin RB, Wright JD, Kennedy-Stephenson J. Use of Dietary Supplements in the United States,
1988-94, Series 11, No. 244, Department of Health and Human Services, National Center for
Health Statistics, Hyattsville, MD; 1999.
158. Slesinski MJ, Subar AF, Kahle LL. J Nutr 126: 3001; 1996.
159. Newman V, et al. JADA 98: 285; 1998.
160. Winters BL, et al. FASEB J 13: A253; 1999.
161. Tippett KS, Cypel YS, Eds. Design and Operation: the Continuing Survey of Food Intakes by
Individuals and the Diet and Health Knowledge Survey, 1994-96, Nationwide Food Survey Report
No. 91-1, US Department of Agriculture, Agricultural Research Service; 1998.
162. Rock CL, et al. Nutr Cancer 29: 133; 1997.
163. Milner JA. J Nutr 129: 1395S; 1999.
164. US Department of Agriculture, USDA-NCC Carotenoid Database for U.S. foods, 1998-1999.
165. NASCO, Nasco nutrition teaching aids, 1999-2000 catalog Fort Atkinson, WI: 437; 1999.
166. National Center for Health Statistics, Dietary Intake Source Data: United States, 1976-80. (DHHS
publication no. PHS 83-1681), Series 11, no. 231, US Department of Health and Human Services,
Washington, DC; 1983.
167. Hess MA, Ed. Portion Photos of Popular Foods, American Dietetic Association: 128; 1997.

© 2002 by CRC Press LLC