Self-reported measures of dietary intake are prone to measurement error that may obscure the relationship of diet and disease. This review addresses strategies to decrease errors during collection of dietary data and statistical approaches to deal with measurement issues once the data are collected.  Examples from two US studies-- the Women’s Health Initiative (WHI) Dietary Modification Trial and the Hispanic Community Health Study/Study of Latinos (HCHS/SOL) have been used. 

Approaches to manage biases during the collection of data include home visits to assess actual portion sizes of foods consumed, early detection of random and systematic error, and technology-based tools to capture dietary intake. The discussion of the WHI and HCHS/SOL ancillary studies illustrates appropriate use of calibration and biomarkers. Biomarkers of interest include both recovery (doubly labeled water, urinary nitrogen, sodium, and potassium), and predictive biomarkers such as total sugars. Discussion of blood concentration biomarkers such as folate, tocopherols and carotenoids and novel biomarkers will supplement this review.

Once data are collected, biomarker calibration is an approach to deal with measurement error. Here the discussion of statistical methods includes the development of regression calibration equations that use both biomarkers and self-report measures along with pertinent participant characteristics; and hazard ratios (ratio of incidence rates) based on Cox regression for assessing diet-disease associations with biomarkers. This review concludes with approaches applicable to nutrition research in the context of Iran.