Automated Self-administered 24-hour Dietary Recall (ASA24)
Background
Self-reported dietary assessment methods are commonly used to measure food intakes for
dietary surveillance, nutritional epidemiology, clinical and intervention research.
Different methods are used for different purposes, but each has advantages and
disadvantages. For example, dietary food records or diaries have been used to collect
food intake data in real time, but this method requires a highly motivated and literate
respondent and is prone to biases from changing their diets on recording days. This often
leads to underreporting and/or undereating. Further, if not automated, this method
requires labor-intensive and expensive data entry and coding that is not standardized
given that respondents' attention to detail and accuracy is highly variable.
A second common dietary assessment method is the food frequency questionnaire (FFQ),
which is used in large-scale epidemiologic research. FFQs ask about usual frequency of
intake and portion size of a long list of food items consumed over some period of time in
the past (generally the previous year). The FFQ is typically self-administered by paper
and pencil and scanned so that data on intakes are easily obtained. The appeal of the FFQ
is that it can be administered quickly and cheaply, and that it queries food intake over a
relatively long period of time. FFQs' disadvantages are that they have a high degree of
measurement error and bias for some nutrients, and that they are not very detailed with respect to
food types, preparation methods, and portion sizes.
The interviewer-administered 24-hour recall has long been regarded as the best
methodology because it provides the highest-quality and least biased dietary data for a
single day. This method allows collection of detailed intake and portion sizes, and
because the data collection occurs after the consumption, this methodology does not affect
what an individual chooses to eat on a given day. The close proximity in time to the
intake day minimizes memory and cognitive issues that afflict other methodologies.
However, the disadvantages of the 24-hour recall are that many administrations are
required to assess usual dietary intakes, and, because they are administered by trained
professionals, they are costly and therefore not practical in research settings with large
sample sizes and/or a need to collect data for multiple days. The development of a
24-hour recall that could be unannounced, automated, and self-administered would make
feasible the administration of multiple days of recalls in large-scale epidemiologic
studies, surveillance studies, behavioral trials, or clinical research, thus enhancing
investigators' ability to assess usual dietary intakes. Coding of food items and
calculation of nutrient intake would be computed automatically. This instrument could
either be sent to participants over the Internet or administered in a clinic/office
setting at low cost. Therefore, an automated self-administered recall could potentially
change the state of the art for data collection in large-scale population nutrition
research.
Development of Web-based ASA24
In the summer of 2004, the NCI began to assess the possibility of creating an automated
and self-administered 24-hour dietary recall. After some brainstorming and investigational work,
the project began in earnest in 2005. The ultimate goal was to create software for an
engaging and easy-to-use instrument, and to make this software available on a public Web
site so that investigators could easily use the system at minimal or no cost.
The intent is that the ASA24 draw its format and design from the current
interviewer-administered Automated Multiple Pass Method (AMPM) 24-hour recall developed by
the US Department of Agriculture (USDA). This requires multi-level food probes to
accurately assess food types and amounts. However, adaptation of the AMPM is necessary
given that ASA24 will take place in a self-administered computer-based environment.
ASA24 will use state-of-the-art automated computer technology, including graphic
enhancements, animated characters to guide participants, and audio language/cues to
enhance use in low-literacy populations. The food list from which respondents select
their intakes for the previous day will include all foods available from USDA's most
current Food and Nutrient Database for Dietary Studies (FNDDS) database. In addition, the
software will include pictures of foods in multiple portion sizes to aid portion size
estimation by the respondent. The software will have the capacity to immediately compute
nutrient and food group estimates for each recall day. If funding allows, ASA24 will be
translated into Spanish. If not, the software will have the capacity to be translated
into Spanish and other languages in the future.
This work builds on the work of an existing instrument, Food Intake Recording Software
System (FIRSSt), which began in late 1997 by Dr. Tom Baranowski and later in conjunction
with Richard Buday of Archimage Associates, in Houston, Texas. FIRSSt has approximately
8,000 pictures of individual foods depicting up to six portion sizes per food, which users
can select to estimate their individual intakes. The ASA24 will use these pictures as
well as others to depict portion sizes.
Current Status & Availability
Dr. Amy F. Subar is leading the NCI effort to develop the software. This work is being
conducted collaboratively with other investigators at NCI and USDA and under contract with
Archimage in Houston, Texas, and Westat in Rockville, MD. An External Working Group has
met twice to discuss the needs and interests of potential users and to provide advice to
the project.
The goal is to complete the project and have the software publicly available by January 15, 2009.
What the Software Will Do
The software will:
- provide tutorials for respondents on how to complete the interview;
- provide respondents with an animated audio character to guide them through the interview with an option to turn off the audio portion;
- ask respondents to report eating occasion with time of consumption;
- ask respondents to provide a "quick list" of foods consumed the previous day;
- allow respondents to find foods to report on the quick list either by browsing through food groups or by typing and searching;
- use the USDA's AMPM interview as a basis for asking detailed probes about each food reported in the quick list;
- use photographs to assist respondents in reporting portion size;
- allow the respondent to add or modify food choices at any time during the interview;
- include a final review of the day's intake and offer one more opportunity for the respondent to add or delete foods;
- automatically produce individual level nutrient and pyramid food group estimates for researchers based on the most current version of USDA's FNDDS; these data can be used by researchers for analysis or to provide reports to respondents;
- provide an optional vitamin and mineral supplement module;
- provide optional modules to assess where meals were consumed, where food was obtained, who the meals were eaten with, and whether or not the TV was on and being watched; and
- allow researchers to add their own opening and closing text scripts and study logo specific to the needs of their project.
|
