Learn More about Combining Instruments

All self-report dietary assessment instruments have strengths and limitations related to the level of detail captured, [glossary term:] measurement error, cost, respondent burden, and other factors. [glossary term:] Combining instruments analytically is a strategy that can capitalize on the strengths and minimize the limitations of each instrument. This approach has been found to be useful for a variety of purposes.

In studies aimed at describing the intakes of a group or groups, a [glossary term:] short-term instrument such as a 24-hour dietary recall (24HR) is generally recommended. In this case, combining the short-term data with data from another instrument, such as a food frequency questionnaire (FFQ) or a [glossary term:] food propensity questionnaire, is generally not thought to be necessary for estimating means and distributions of usual intake. However, frequency data may be of use in estimating the extreme tails of the distribution, particularly for episodically-consumed dietary components. See Choosing a Dietary Assessment Approach for recommendations on dietary assessment for the purpose of describing the intake of groups.

The available evidence suggests that combining short-term instruments (e.g., 24HRs) with [glossary term:] long-term instruments (e.g., FFQs, food propensity questionnaires, screeners) may be a useful strategy for estimating associations between diet and a health [glossary term:] outcome. The emerging research indicates that collecting 4 to 6 recalls and an FFQ from all participants over a period of time, such as a year, may be optimal. In this scenario, the recalls serve as the [glossary term:] main dietary assessment instrument, with estimates of frequency of consumption from the FFQ included as [glossary term:] covariates in statistical models to estimate diet and health associations. Simulation has shown that the use of data from the combination of instruments may mitigate the effects of measurement error, increasing the [glossary term:] precision of estimates and power to detect relationships between diet and an outcome. FFQ data are likely to be particularly useful in improving estimates of associations between [glossary term:] episodically-consumed dietary components and health outcomes.

Combining data from multiple self-report instruments to improve estimates of dietary intake should not be confused with: 1) [glossary term:] validation studies (see Key Concepts about Validation) in which an instrument, such as an FFQ, is compared to less biased measures (i.e., reference measures), such as [glossary term:] recovery biomarkers (Learn More about Biomarkers) or 24HRs, to assess the extent to which the instrument accurately assesses [glossary term:] true intake, or 2) [glossary term:] regression calibration sub-studies (Learn More about Regression Calibration) conducted as part of a diet and health study in which an instrument, such as an FFQ, is administered to all participants and multiple 24HRs or food records are collected from a subsample in order to improve estimates of diet and health outcomes.

It is also possible to combine self-report data with [glossary term:] biomarker data in diet and health studies. For example, [glossary term:] concentration biomarkers that are correlated with dietary intake data can sometimes be used in combination with self-report data to improve the statistical [glossary term:] power to detect relationships. This is a promising area of ongoing research to mitigate the effects of measurement error in dietary intake data.

For More Information

Carroll RJ, Midthune D, Subar AF, Shumakovich M, Freedman LS, Thompson FE, Kipnis V. Taking advantage of the strengths of 2 different dietary assessment instruments to improve intake estimates for nutritional epidemiology. Am J Epidemiol 2012 Feb 15;175(4):340-7. [View Abstract]

Freedman LS, Midthune D, Carroll RJ, Tasevska N, Schatzkin A, Mares J, Tinker L, Potischman N, Kipnis V. Using regression calibration equations that combine self-reported intake and biomarker measures to obtain unbiased estimates and more powerful tests of dietary associations. Am J Epidemiol 2011 Dec 1;174(11):1238-45. [View Abstract]

Tooze JA, Midthune D, Dodd KW, Freedman LS, Krebs-Smith SM, Subar AF, Guenther PM, Carroll RJ, Kipnis V. A new statistical method for estimating the usual intake of episodically consumed foods with application to their distribution. J Am Diet Assoc 2006 Oct;106(10):1575-87. [View Abstract]