coding large volumes of data for analysis. Also, written reporting requires literacy
(including reading of food labels), and there is a need for multiple days of intake to
assess“usual”intake. Most importantly, this approach commonly leads to reporting
bias in that the act of recording intake leads to changes in eating behavior over time.
In fact, food records are a commonly employed behavioral strategy for the pro-
motion of weight control. Further, those willing to accurately complete records tend
to be more educated, female, and of normal body mass index. In addition to the
concerns associated with data collection, this method is only as accurate as the
nutrient database supporting thefinal analysis of intake. Most investigators use the
Nutrient Database for Research (NDS-R) developed by the University of Minnesota
and based on the USDA Nutrient Database of Foods (Center 2014 ). While accurate
to a large degree, for select nutrients, the database may not be fully populated for all
foods. Thus, some measurement error is introduced at the analysis phase of mea-
surement as well.
Twenty-Four-Hour Recalls
A second approach for self-report of diet is the 24-h recall. The most common
method for 24-h recall collection is the multi-pass method (Conway et al. 2003 ).
This approach has been validated by investigators at the USDA and has an
acceptable level of accuracy if the method is consistently applied by trained
interviewers. Briefly, the respondent is called by telephone to answer a series of
questions regarding their intake over the prior 24-h period. A single recall is not
sufficient to capture usual intake. General practice is to assess intake on 3 weekdays
and one weekend day over a 2-week period. Ideally, the calls are not scheduled in
advance, but in recent times with caller identification technology, many investi-
gators have had to resort to scheduling the calls in advance. The trained interviewer
then asks a series of questions regarding intake as described in Fig.9.1. In addition
to listing food consumed, respondents are probed regarding“missing”foods, such
as beverages, condiments, and added fats/flavorings. Twenty-four-hour recalls also
require the respondent to have knowledge of portion size, and training on portion
size estimation is advised. Recalls are commonly collected using the NDS-R
computer-based software system during the telephone call. The NIH has a
Web-based 24-h recall system (ASA24) that can be used (http://appliedresearch.
cancer.gov/asa24/). The participant burden is increased for the ASA24, but the need
for trained interviewers is markedly reduced except to support respondents that may
need assistance. On occasion, 24-h recalls are taken on-site using pen/paper, par-
ticularly for home-based assessments, and then entered into the computer-based
analytical system at a later time.
Strengths of the 24-h recall include detail of data, lower participant burden and
report bias, and no requirement for literacy as compared to traditional diet
assessment methods such as food diaries. Weaknesses include time needed to
complete repeat recalls, need for trained interviewers, and potential recall bias,
9 Biomarkers of Diet and Nutritional Health 173