A toolkit for measurement error correction, with a focus on nutritional epidemiology

Abstract

Exposure measurement error is a problem in many epidemiological studies, including those using biomarkers and measures of dietary intake. Measurement error typically results in biased estimates of exposure-disease associations, the severity and nature of the bias depending on the form of the error. To correct for the effects of measurement error, information additional to the main study data is required. Ideally, this is a validation sample in which the true exposure is observed. However, in many situations, it is not feasible to observe the true exposure, but there may be available one or more repeated exposure measurements, for example, blood pressure or dietary intake recorded at two time points. The aim of this paper is to provide a toolkit for measurement error correction using repeated measurements. We bring together methods covering classical measurement error and several departures from classical error: systematic, heteroscedastic and differential error. The correction methods considered are regression calibration, which is already widely used in the classical error setting, and moment reconstruction and multiple imputation, which are newer approaches with the ability to handle differential error. We emphasize practical application of the methods in nutritional epidemiology and other fields. We primarily consider continuous exposures in the exposure-outcome model, but we also outline methods for use when continuous exposures are categorized. The methods are illustrated using the data from a study of the association between fibre intake and colorectal cancer, where fibre intake is measured using a diet diary and repeated measures are available for a subset. © 2014 The Authors.

Comparing metabolite profiles of habitual diet in serum and urine

“…Cases had a lower intake of fish and higher intake of meat fat. 3 One participant each was missing information on supplemental vitamin E use and supplemental calcium use.…”

“…These correlation magnitudes, which broadly reflect strength of serum and urine multiple-metabolite prediction of diet based on statistical modeling (i.e., prediction based on a training data set and testing with the use of a testing data set), were generally comparable in serum and urine, with evidence that results did not Values may not sum to 100% because of rounding. 3 Comparison by case-control status was significant at P , 0.05 on the basis of a chi-square test for categorical variables and a t test for continuous variables. 4 Calculated by adding the number of times per week engaged in moderate-intensity leisure-time physical activity plus 2 times the number of times per week engaged in vigorous-intensity leisure-time physical activity.…”

“…However, inconsistent findings for some diet-disease associations have contributed to a shifting evidence base, curbing the strong promotion of dietary guidance in disease prevention. These inconsistent findings may partly reflect methodologic limitations in measuring diet with the use of selfreported questionnaires, which are subject to recall bias and measurement errors that may weaken diet-disease associations (2)(3)(4). Dietary biomarkers have the potential to improve the measurement of dietary exposures by validating diet questionnaires (5), by replacing inadequate dietary data (6), or for regression calibration (7).…”

Determination of alkylresorcinols and their metabolites in biological samples by gas chromatography–mass spectrometry

“…Most studies rely on self-reported intakes from food frequency questionnaires (FFQ), which are known to suffer from relatively large systematic and random measurement errors [4]. Accurate whole grain intake data acquisition via FFQ may be hampered by having few questions on cereal foods, difficulties among consumers in recognizing whole grain products, large variations in whole grain content in cereal foods and lack of food composition data [5]. Therefore, dietary biomarkers may provide an independent tool to overcome some of these obstacles and a better way for objective ranking of whole grain intake in epidemiological studies [6,7].…”

The validity of self-reported behaviors: methods for estimating underreporting of risk behaviors

“…regression calibration 19, 20 ) have been used to adjust risk-outcome associations for measurement bias, but also require a biomarker with high sensitivity and specificity. STIs, for example, are imperfectly sensitive as markers for condomless sex since one may engage in the behavior but not become infected with an STI.…”