November 1, 1989American Journal of Epidemiology
The Reproducibility of Dietary Intake Data Assessed With the Cross-Check Dietary History Method
Authors
Discussion
Loading...
Member takes
Implication
Repeat dietary survey demonstrates strong reproducibility of nutrient intake measurements in elderly men, suggesting minimal correlation attenuation from measurement error.
Key Points
- To evaluate the reproducibility and measurement error of dietary intake data gathered using the cross-check dietary history method over repeated intervals.
- Repeated dietary surveys were administered to elderly men in the Zutphen Study (Zutphen, The Netherlands) at 3 months (n = 115) and 12 months (n = 145) following a baseline assessment in April 1985.
- Calculated ratios of interindividual to intraindividual variance and estimated attenuation factors across multiple nutrients to evaluate error-induced correlation loss.
- Differences in reproducibility estimates between the 3-month and 12-month periods were generally small.
- Ratios of interindividual to intraindividual variance were high for carbohydrates (2.6 at 3 months; 2.7 at 12 months) and low for vitamin A (1.1 at 3 months; 0.4 at 12 months).
- Most nutrients showed an attenuation factor of approximately 0.8, meaning an underlying correlation of 0.40 with a risk factor would attenuate only slightly to 0.32.
Cite This Study
Bloemberg et al. (1989) studied this question.