Our Projects

Measurement Error in Cancer Epidemiology

Sponsored By

Measurement Errors in Cancer Epidemiology Project Abstract: Exposure measurement errors in cancer epidemiology pose 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 correction for measurement error and misclassification. In this new cycle, our 8th continuous cycle of NIH/NCI funding since 1989, we drill down deeper into the multi-leveled themes which arise in cancer research from the population to the individual to the molecular. We branch out in several seminal new directions of critical importance for the translation of results of population-based research to clinical practice and public policy, with one focus on measurement error correction of the population attributable risk percent (Aim 1). Since our last competing continuation, novel technologies for the processing of high-throughput high-dimensional data have emerged and are taking on an increasingly important role in mainstream cancer research. Within the molecular focus of Aim 3, we will develop methods to fully utilize this growing and massive body of big 'omics data, which along with its increased size comes increased measurement error and misclassification. Gene by environment interactions are increasingly the focus of novel hypotheses, but the ability to detect these and quantify them with accuracy is challenged by measurement error in the environmental variable and misclassification of the finely mapped genetic data. We will continue to pursue unsolved methodologic problems arising from measurement error in prospective cohort studies to permit unbiased and efficient estimation of the effects of individual-level risk factors (Aim 2), in particular as motivated by studies of diet and physical activity in relation to cancer incidence and mortality, with a new focus on methods suitable for rarer cancers, mediation analysis, the effects of change in risk factors, for prospective studies of diet and cancer arising from sample survey data, and more stable methods for data with moderate to substantial misclassification error. As previously, user-friendly, public use software development will be a central feature accompanying all new methods development.