Dietary Assessment
Measurement Issues
Although dietary assessment methods have been rigorously
evaluated, there are some measurement issues that deserve to be discussed.
·
Estimation of portion sizes
·
Mode of administration
·
Estimating usual intake
·
Reference nutrient database
·
Validation
First, researchers have shown that untrained individuals have a difficult time estimating portion sizes of foods. This is less so with defined unit foods, such as a slice of bread, can of Coke, or piece of fruit.
To
educate participants about portion sizes, researchers have used photos and
three-dimensional models to help participants correctly measure portion sizes.
Technology has greatly changed the mode of administration
of dietary assessments in the last few years and will continue to evolve the
methods of collecting data. The reading material discusses the issues between
personal interviews, phone interviews, and mail surveys. Phone interviews have
the highest response rate, but with call screening less people are answering
their phones. Additionally researchers worry about proper sampling of subjects
for phone interviews because few young adults have a landline phone and could be
excluded from current sampling techniques.
The reading material was published in 2008 and since then new issues and solutions have emerged in dietary assessments.
Numerous applications have been designed for smartphones and tablets that help guide individuals through the data collection process.
The applications are getting better and better each day.
One particular app that has promise is a combined dietary record with the participant taking a photo of their plate of food.
The
portion sizes can then be estimated by the researcher team for a more accurate
estimation.
Researchers are interested in the usual intakes of nutrients and food to make associations between diet and health or disease risks.
Estimates of consumption of specific foods and food groups are also important for evaluating the progress in meeting key objectives in such national public health initiatives.
The problem is that the true usual intake is not observable. Several statistical models have been developed to estimate usual intake, but without these models 7-14 days of data collection would be required.
If only a few days of data are used to estimate usual intake, the within-person variation is large, leading to a biased estimate.
Population distributions generated from only a few days of data
collection lead to a wider distribution with larger deviations above and below
cut points.
Researchers spend a considerable amount of time making sure the dietary assessment tool is correct, but just as important is the choosing of a nutrient/food database to analyze the information collected.
A nutrient/food database consists of lists of foods with a food code and the nutrient composition per 100 grams of the food.
The USDA National Nutrient Database for Standard Reference has nearly 8,000 different foods in the nutrient database.
The USDA site is one of the most comprehensive sites on the web, but other specialty sites exist for specific nutrients or food additives.
Researchers must know the content of the
database they are utilizing as well as the methods of analysis for the nutrient
levels, such as laboratory analysis or estimates based on knowledge about the
food.
Lastly,
before any new dietary assessment method is used it must be validated against
other more established methods or a gold standard.
Remember reliability and validity are population specific.
Thus, if an instrument has been evaluated in one population but a researcher wants to use it in a different population, the instrument must be validated in the new population.
Validation studies are
expensive and difficult to conduct, so if behooves a researcher to look at
previously validated assessment tools before deciding to design a new assessment
instrument.
You have reached the end of Topic 15, and the end of the course!
Make sure to complete the review quiz for this topic.