The validation of dietary assessment measures is a necessary part of any dietary investigation. The use of unvalidated instruments is likely to lead to misinterpretation of diet-disease relationships. In most circumstances, unvalidated measures of diet will lead to the conclusion that there is no diet-disease relationship when in reality one exists (bias toward the null). In circumstances in which differential mis-classification is at play, it may also lead to the conclusion that there is a relationship when in fact none exists. In order not to waste resources, it is vital to ensure either that dietary assessments are valid or that the errors associated with them are clearly identified and taken into account in the interpretation of diet-disease relationships. In summary,
• Never take dietary measurements at face value.
• Always include an analysis of errors of dietary data in epidemiological studies.
• Ensure that the validation study sample is representative of the population in the main study.
• Quantify all significant sources of error.
• Obtain measures of within-subject reliability.
• Adjustment of estimates of intake is preferable to excluding misreporters from the diet-disease analysis.
Finally, it is worth noting that virtually all of the discussion in this article relates to nutrient intakes. There is an urgent need to improve our understanding of the validity of measurements of the consumption of foods. Advice on healthy eating is given primarily in terms of foods, not nutrients.
See also: Dietary Intake Measurement: Methodology. Dietary Surveys. Energy Expenditure: Doubly Labeled Water.
Was this article helpful?