A Multivariate Method for Measurement Error Correction Using Pairs of Concentration Biomarkers

Department of Adventist Health Study-2, Loma Linda University, Loma Linda, CA 92350, USA.
Annals of Epidemiology (Impact Factor: 2). 02/2007; 17(1):64-73. DOI: 10.1016/j.annepidem.2006.08.002
Source: PubMed


Measurement error is a pervasive problem in behavioral epidemiology, and available methods of correction all have generally untenable assumptions. We propose a multivariate method with more realistic assumptions.
The method uses two concentration biomarkers for each nutritional variable of interest and structural equation modeling. This produces corrected estimates of the effects on an outcome variable of changing the true exposure variables by one standard deviation, a standardized regression calibration. However, hypothesis testing in original units is preserved. The main assumptions are that certain error correlations between dietary estimates and biomarkers or between biomarkers be close to zero.
Two illustrative models used simulated data with the covariance structure of a real data set. The corrections produced often were very substantial. A sensitivity analysis allowed error correlations to depart from zero over a modest range. Root mean square biases show the advantage of the corrected approach. Relatively large calibration studies are needed for adequate precision.
As long as concentration biomarkers are selected carefully, error-corrected multivariate hypothesis testing and standardized effect estimation is possible. With the deviations from assumptions that were tested, the corrected method usually produces much less biased results than an uncorrected analysis.

3 Reads
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Dietary biomarkers are objective measures of food or nutrient intake and can be related to endpoints in epidemiological studies, used in the validation of dietary assessment instruments and used to check of compliance during intervention studies. Alkylresorcinols (AR), phenolic lipids present exclusively in the outer parts of wheat and rye grains, have been suggested as biomarkers of whole grain wheat and rye intake. The overall aim with this thesis was to evaluate AR as specific biomarkers of whole grain wheat and rye intake. This was conducted by developing a rapid GC-MS method for the analysis of AR in plasma and by studying AR pharmacokinetics, dose-response, reproducibility and relative validity in human intervention studies under controlled intake conditions. Factors affecting plasma AR concentrations were investigated in free-living Danish women. The method developed proved suitable for the analysis of relatively small sample volumes (50- 200μL). The results showed that AR in fasting plasma samples can be used as short-term concentration biomarkers, reflecting the intake range normally found in the Nordic countries in a dose-dependent manner. One or two repeated measurements of AR were found to adequately describe a subject’s average plasma AR concentration at regular and constant intake. In free-living Danish women, rye bread was identified as the major factor affecting plasma AR concentration and there was no evidence of non-dietary factors or other foods having an effect. In conclusion, our results support that AR can be used as biomarkers in intervention studies on whole grain wheat and rye and probably also in epidemiological endpoint- and validation studies.
  • International Journal of Epidemiology 05/2008; 37(2):260-5. DOI:10.1093/ije/dym165 · 9.18 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: A number of studies have evaluated the health of vegetarians. Others have studied the health effects of foods that are preferred or avoided by vegetarians. The purpose of this review is to look critically at the evidence on the health effects of vegetarian diets and to seek possible explanations where results appear to conflict. There is convincing evidence that vegetarians have lower rates of coronary heart disease, largely explained by low LDL cholesterol, probable lower rates of hypertension and diabetes mellitus, and lower prevalence of obesity. Overall, their cancer rates appear to be moderately lower than others living in the same communities, and life expectancy appears to be greater. However, results for specific cancers are much less convincing and require more study. There is evidence that risk of colorectal cancer is lower in vegetarians and in those who eat less meat; however, results from British vegetarians presently disagree, and this needs explanation. It is probable that using the label "vegetarian" as a dietary category is too broad and that our understanding will be served well by dividing vegetarians into more descriptive subtypes. Although vegetarian diets are healthful and are associated with lower risk of several chronic diseases, different types of vegetarians may not experience the same effects on health.
    American Journal of Clinical Nutrition 04/2009; 89(5):1607S-1612S. DOI:10.3945/ajcn.2009.26736K · 6.77 Impact Factor
Show more