Statistical Challenges in Public Health and Medicine

James Rosenberger Chair
Penn State University and NISS
Lingzhou Xue Discussant
Pennsylvania State University
Lingzhou Xue Organizer
Pennsylvania State University
James Rosenberger Organizer
Penn State University and NISS
Tuesday, Aug 8: 8:30 AM - 10:20 AM
Invited Paper Session 
Metro Toronto Convention Centre 
Room: CC-718B 



Main Sponsor

National Institute of Statistical Sciences

Co Sponsors

Biometrics Section


Challenges of Modeling Longitudinal Intensive Care Unit Data

Prediction of health outcomes is an important component for determining how to make recommendations and treat individuals. Regarding treatment, the intensive care unit (ICU) is a place where many such decisions are made. A primary goal for ICU patients is treating them to achieve positive outcomes (e.g., hospital discharge alive, improvement from in-hospital ailments, extended survival). A major analytical issue is the preponderance of information available at ICU entry (e.g., age, sex, co-morbidities, prescriptions, vital signs), and especially longitudinally (e.g., vital sign changes, dynamic renal function, in-ICU treatment). I will present some interesting analytic challenges utilizing longitudinal data for predictive modeling that my collaborators and I have encountered from a large ICU database, and discuss a few remedies that we have investigated. 


Joel Dubin, University of Waterloo

Ensemble methods for testing a global null with application to whole genome sequencing association studies

Testing a global null is a canonical problem in statistics and has a wide range of applications. In view of the fact of no uniformly most powerful test, prior and/or domain knowledge are commonly used to focus on a certain class of alternatives to improve the testing power, e.g., the class of alternatives in the scenario of the same effect sign or signal sparsity. However, it is generally challenging to develop tests that are particularly powerful against a certain class of alternatives. In this paper, motivated by the success of ensemble learning methods for prediction or classification, we propose an ensemble framework for testing that mimics the spirit of random forests to deal with the challenges. Our ensemble testing framework aggregates a collection of weak base tests to form a final ensemble test that maintains strong and robust power. The key component of the framework is to introduce a certain random procedure in the construction of base tests. We then apply the framework to four problems about global testing in different classes of alternatives arising from Whole Genome Sequencing (WGS) association studies. Specific ensemble tests are proposed for each of these problems,  


Xihong Lin, Harvard T.H. Chan School of Public Health

Mediation Analysis with External Summary Data on Total Effect

As modern assaying technologies continue to improve, environmental health studies are increasingly measuring endogenous omics data to study intermediary biological pathways of outcome-exposure associations. Mediation analysis is often carried out when there is a well-established literature showing statistical and practical significance of the association between an exogenous exposure and a health outcome of interest, or the total effect. For example, there are a plethora of studies associating maternal phthalate exposure with preterm delivery, and researchers are now trying to characterize the mechanisms by which phthalate exposure impacts final gestational age. Existing methodology for performing mediation analyses does not leverage the rich external information available on the total effect. We show that incorporating external summary-level information on the total effect improves estimation efficiency of the natural direct and indirect is a function of the partial R-squared comparing the outcome model with and without the mediators. Additionally, we discuss how to handle incongruous external information which can arise from transportability violations or fundamental 


Bhramar Mukherjee, University of Michigan