CLOSE MENU
Join the monthly forum from the Department of Epidemiology and Biostatistics.
This month’s presentation is by Dr. Giorgos Bakoyannis who will discuss; “Estimating optimal individualized treatment rules with multistate disease processes.”
No registration is required. Join via the Microsoft Teams link.
Meeting ID: 296 097 469 797 65
Passcode: aL7Vh6tm
Keynote Speaker: Dr. Giorgos Bakoyannis
Title: Estimating optimal individualized treatment rules with multistate disease processes
Abstract: Modern precision medicine recognizes the substantial heterogeneity of human diseases and seeks to develop and deliver therapies tailored to individual patients. At the heart of these efforts is the development of data-driven individualized treatment assignment rules.
The purpose of such rules is to provide the right treatment to a given patient and, thereby, to improve health outcomes among patients overall. Multistate disease process data are ideal for precision medicine purposes as they can be leveraged to improve more refined health outcomes (compared to, e.g., overall or progression-free survival), as well as incorporate patient preferences regarding quality versus quantity of life.
In this talk, Bakoyannis will present a nonparametric outcome weighted learning approach for estimating optimal individualized treatment rules based on multistate process data from both randomized clinical trials and observational studies. The validity of the proposed approaches is formally established to ensure they reliably identify optimal individualized treatment assignment rules. He will also provide practical tools to quantify uncertainty in estimated outcomes and propose an approach for constructing confidence intervals.
Simulation studies show that the proposed methodology and inference procedures perform well even with small sample sizes and high rates of loss to follow-up. The methodology is illustrated using (i) data from a randomized clinical trial on the treatment of metastatic squamous-cell carcinoma of the head and neck and (ii) observational data on the treatment of multiple myeloma.
Speaker Bio: Dr. Giorgos Bakoyannis is an Associate Professor in the Department of Biostatistics and Health Data Science at Indiana University Indianapolis, where he also serves as Director of Public Health Science Research. His methodological research focuses on precision medicine, particularly the development of statistical learning methods for estimating optimal individualized treatment rules and causal inference. His expertise further includes nonparametric and semiparametric methods for analyzing complex event history data, with particular emphasis on challenges commonly encountered in real-world data, such as missingness, misclassification, and interval censoring.
Dr. Bakoyannis has received awards from the American Statistical Association, the International Biometric Society – Eastern North American Region, and the International Chinese Statistical Association. He has served as Principal Investigator on research funded by the Patient-Centered Outcomes Research Institute (PCORI) and the National Institutes of Health (NIH), supporting his methodological work. In addition, Dr. Bakoyannis applies statistical and data science methods to health science research across multiple domains, including cancer, infectious diseases, and geriatrics. Throughout his career, he has served as the lead statistician on multiple NIH-funded projects in these areas.
