Exposure measurement errors in cancer epidemiology poses special methodologic challenges. For example, nutritional and physical activity patterns form the basis for many etiologic hypotheses concerning cancer. However, nutrient intake and physical activity are difficult to measure precisely. It is the role of measurement error correction methods to validly and efficiently estimate the relationship between exposures and cancer outcomes. To accomplish this requires both a main study where disease and the surrogate exposure are measured and validation data to determine the extent of the measurement error. In this proposal, we seek to further expand our group's previous work on methods of corrections for measurement error and misclassification. A major focus is on nutritional studies based on intakes and activity measures reported at multiple questionnaires over time when the target exposure is long-term average diet and/or activity. Here, the dependent variable is time to cancer incidence or mortality;thus, Cox models are considered in which time-varying covariates such as cumulative averages and cumulative exposures are of primary interest. Modeling and variable selection issues in the context of real-life applications of measurement error correction methods will be carefully studied, with an aim to provide guidance for current and future users of the methodology;no such guidance currently exists. Application of methods to correct for measurement error in time-varying exposures such as cumulative averages requires validation studies in which repeated measures within study participants are repeatedly validated. Analysis of the currently ongoing HSPH Lifestyle Validation Study'which will make available validation of up to four repeated dietary and physical activity measurements, using state of the art biomarker technology including doubly labeled water and urinary nitrogen, will be another major focus of our work. Finally, our attention will turn to an entirely new area of research in this cycle: methods which consider measurement error and misclassification in the development and evaluation of models for the prediction of breast, colon and ovarian cancer risk. As previously, user friendly , public use software development will be a focus of all new methods of development.

Public Health Relevance

Measurement error is a major source of bias in epidemiologic research aimed at elucidating the relationship between diet, physical activity, and cancer incidence and mortality. A competing continuation of a methodologic research effort ongoing since 1989 focusing on the development of methods and software to adjust for this bias in point and interval estimates of relative risk, the current proposal aims to develop methods for bias reduction and elimination in prospective cohort studies with repeated measures of exposure over time, for relative risks and in cancer risk prediction models. In addition, we will analyze the newly completed Lifestyle Validation Study, a unique resource with state of the art biomarkers and up to four repeated validations, and apply this information to the newly developed methods to cancer studies arising in the Nurses'Health Study, the Nurses'Health Study II, and the Health Professionals Follow-up Study.

National Institute of Health (NIH)
National Cancer Institute (NCI)
Research Project (R01)
Project #
Application #
Study Section
Epidemiology of Cancer Study Section (EPIC)
Program Officer
Verma, Mukesh
Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
Harvard University
Public Health & Prev Medicine
Schools of Public Health
United States
Zip Code
Aschard, Hugues; Spiegelman, Donna; Laville, Vincent et al. (2018) A test for gene-environment interaction in the presence of measurement error in the environmental variable. Genet Epidemiol 42:250-264
Kim, Sehee; Li, Yi; Spiegelman, Donna (2016) A semiparametric copula method for Cox models with covariate measurement error. Lifetime Data Anal 22:1-16
Rosner, Bernard; Tworoger, Shelley; Qiu, Weiliang (2015) Correcting AUC for Measurement Error. J Biom Biostat 6:
Yi, Grace Y; Ma, Yanyuan; Spiegelman, Donna et al. (2015) Functional and Structural Methods with Mixed Measurement Error and Misclassification in Covariates. J Am Stat Assoc 110:681-696
Rosner, Bernard; Hendrickson, Sara; Willett, Walter (2015) Optimal allocation of resources in a biomarker setting. Stat Med 34:297-306
Khudyakov, Polyna; Gorfine, Malka; Zucker, David et al. (2015) The impact of covariate measurement error on risk prediction. Stat Med 34:2353-67
Wang, Molin; Liao, Xiaomei; Spiegelman, Donna (2013) Can efficiency be gained by correcting for misclassification? J Stat Plan Inference 143:
Zucker, David M; Gorfine, Malka; Li, Yi et al. (2013) A regularization corrected score method for nonlinear regression models with covariate error. Biometrics 69:80-90
Liao, Xiaomei; Spiegelman, Donna; Carroll, Raymond J (2013) Regression calibration is valid when properly applied. Epidemiology 24:466-7
Spiegelman, Donna; Logan, Roger; Grove, Douglas (2011) Regression calibration with heteroscedastic error variance. Int J Biostat 7:4

Showing the most recent 10 out of 51 publications