Measurement error correction for nutritional exposures with correlated measurement error: Use of the method of triads in a longitudinal setting

• Published in: Statistics in medicine Vol. 27; no. 18; pp. 3466 - 3489
• Main Authors: Rosner, Bernard, Michels, Karin B., Chen, Ya-Hua, Day, Nicholas E.
• Format: Journal Article
• Language: English
• Published: Chichester, UK John Wiley & Sons, Ltd 15.08.2008; Wiley Subscription Services, Inc
• Subjects: Aged; Bias; biomarkers; correlated error; Correlation analysis; Diet; Diet Records; Estimates; Feeding Behavior; Female; Humans; longitudinal data; Longitudinal Studies; Male; measurement error; Measurement errors; Models, Statistical; Nutrition; Nutrition Assessment; Nutritional Status; Regression analysis; Surveys and Questionnaires; Vitamin C
• ISSN: 0277-6715; 1097-0258
• DOI: 10.1002/sim.3238

Nutritional exposures are often measured with considerable error in commonly used surrogate instruments such as the food frequency questionnaire (FFQ) (denoted by Qi for the ith subject). The error can be both systematic and random. The diet record (DR) denoted by Ri for the ith subject is considered an alloyed gold standard. However, some authors have reported both systematic and random errors with this instrument as well. One goal in measurement error research is to estimate the regression coefficient of Ti (true intake for the ith subject) on Qi denoted by λTQ. If the systematic errors in Qi and Ri (denoted by qi and ri) are uncorrelated, then one can obtain an unbiased estimate of λTQ by λRQ obtained by regressing Ri on Qi. However, if Corr(qi, ri)>0, then λRQ>λTQ. In this paper, we propose a method for indirectly estimating λTQ even in the presence of correlated systematic error based on a longitudinal design where Qi (surrogate measure of dietary intake), Ri (a reference measure of dietary intake), and Mi (a biomarker) are available on the same subjects at 2 time points. In addition, between‐person variation in mean levels of Mi among people with the same dietary intake is also accounted for. The methodology is illustrated for dietary vitamin C intake based on longitudinal data from 323 subjects in the European Prospective Investigation of Cancer (EPIC)‐Norfolk study who provided two measures of dietary vitamin C intake from the FFQ (Qi) and a 7‐day DR (Ri) and plasma vitamin C (Mi) 4 years apart. Copyright © 2008 John Wiley & Sons, Ltd.