This project responds to timely methodology needs and opportunities in chronic disease population science and disease prevention research, as arise in prospective cohort studies and disease prevention randomized trials.
The first aim proposes the development of improved failure time data analysis methods, including continuing work on a nonparametric multivariate survivor function estimator having Kaplan-Meier marginals and its regression generalizations;continuing work on the use of positive and negative attributable fractions for disease risk attribution;and the development of mean process modeling procedures with shared parameters across clinical outcomes, for multivariate longitudinal data.
A second aim i s concerned with covariate measurement error methods development to enhance the reliability of nutritional and physical activity epidemiology research through the use of biomarkers, including nonparametric approaches to hazard ratio association estimation, and methods based on fraction of provided nutrient explained by candidate biomarkers in human feeding study contexts.
A third aim will develop marker panel scoring methods to assess treatment benefits versus risks In relation to high-dimensional genomic markers, as a key step toward a treatment choice objective.
A final aim responds to opportunities for biological network and preventive intervention development, through methods development for the analysis of allele-specific tumor DNA copy number alterations array data, and through the use of high-dimensional blood protein concentration data for the initial evaluation of candidate disease prevention interventions. The general approach involves the specification of pertinent statistical models and the use of both theoretical probabilistic methods, and computer simulations that are informed by application to data from the substantive research contexts in which the six participating biostatistical investigators are engaged.
This project proposes to develop statistical methods to augment the reliability of nutritional and physical epidemiology reports;to classify persons according to prevention intervention benefits and risks based on their genotype;to support the development and initial evaluation of candidate preventive interventions;and to more efficiently analyze cohort study follow-up data on multiple clinical outcomes for individual study subjects.
Showing the most recent 10 out of 319 publications
