Professor Laurence Freedman
Director, Biostatistics Unit,
Gertner Institute for Epidemiology and Health Policy Research,
Sheba Medical Center,
Tel Hashomer 52161
Israel
Telephone: 972-3-5305390 (W)
972-2-6793472 (H)
Fax: 972-3-5349607
e-mail: lsf@actcom.co.il
Abstract
Combining self-reports and concentration biomarkers in the analysis of nutritional epidemiologic studies
Identifying diet-disease relationships in nutritional cohort studies is plagued by the measurement error in self-reported intakes. We propose using combined analyses of self-reported dietary intake together with levels of a biomarker known to be modified by the dietary intake, so as to strengthen analyses of diet-disease hypotheses. We consider combining these two measures using a joint statistical test of effects in a bivariate model, the “extreme quantiles” method, Howe’s method, and principal components and compare the statistical power of the methods using computer simulation under different scenarios. The scenarios considered are based on data from the literature on the relationship between lutein intake and age-related macular degeneration and on the relationship between beta-cryptoxanthin and stomach cancer. The results show that principal components performs particularly well under a variety of scenarios, leading to sample sizes between 30-50% of those required for conventional analyses of self-reported intake and disease. We discuss the practical and theoretical considerations that accompany the proposed approach.