More than 300 participants from across the country participated in the 7th Duke-Industry Statistics Symposium (DISS2021) “Emerging Clinical Initiatives in Pharmaceutical Development: Methodology and Regulatory Perspectives.” While the symposium was delayed last year and had to be held online due to the Covid-19 Pandemic, there was noticeable uptick in enthusiasm and participation by students and newcomers.
The first day of the Symposium was devoted to five short courses with topics such as: Data Monitoring Committees, Multiple Endpoints, Reproducible Data Analysis Pipelines in R, FDA Guidance on Adaptive Clinical Trials and Artificial Intelligence in Health Research.
On Thursday, David Page, Chair of the Duke Department of Biostatistics and Bioinformatics and Organizing Committee members Rakhi Kilaru, PPD Director of Statistical Science and Richard Payne, Eli Lilly Research Scientist gave opening remarks. Dr. Yeh-Fong Chen Organizing Committee member and mathematical statistician from FDA/CDER introduced the first keynote speaker, ShaAvhrée Buckman-Garner, MD, PhD, FAAP. Dr. Buckman-Garner is the Director of the Office of Translational Sciences (OTS), Center for Drug Evaluation and Research (CDER), U.S. Food and Drug Administration. Her keynote was titled “A Perspective on Progress, Challenge and Opportunity." She began with an overview of CDER and an explanation of priorities and achievements during the COVID-19 pandemic. The remainder of her talk addressed strategic objectives of the regulatory program modernization as required by the 21st Century Cures Act, with a focus on the use of complex innovative trial designs (CID). Dr. Buckman-Garner also discussed challenges inherent in the evolving clinical trials environment, including emergence of new technologies that enable integration of non-traditional data sources with clinical trial data, including wearable devices and medical records. Dr. Buckman-Garner concluded her talk by discussing opportunities for clinical trial sponsors to engage with CDER during the drug development process and discussed training opportunities at the FDA available through the Oake Ridge Institute for Science and Education (ORISE) and the Translational Science Interagency Fellowship Program at the National Center for Advancing Translational Sciences (NCATS).
Twenty-five parallel sessions were held on Thursday and Friday. These sessions included diverse topics such as: Pediatric Clinical Development, Simulation in Complex and Innovative Trials, Integrated Analysis of RCTs and Population-based Studies, Considerations for Rare Disease Clinical Trials, Statistical Considerations for RWE/RWD Studies, Statistical Challenges and Advances in HIV Drug Development, Artificial Intelligence in Health Care Research, Use of Companion Diagnostics (CoDx) in Clinical Trials, Statistical Methodology for Cancer Immunology Trial Design, Biomarker-driven Trials with Adaptive Threshold Detection, Combination Therapy in Oncology Clinical Development, Digital / Mobile Health in Clinical Trials, Analytical and Methodological Challenges in the Study of Complex Liver Diseases, Better Evidence Synthesis via Innovative Methods, and Interactive Visualization Tools.
The second keynote was given on Friday morning by Dr. David Page, Chairman of the Duke Department of Biostatistics and Bioinformatics. Dr. Page spoke on “Machine Learning from Electronic Health Records.” Dr. Page began his talk by reviewing how machine learning (ML) has already had positive impact on our daily lives, including movie selection on Netflix and natural language processing. Dr. Page then moved on to discuss the potential for ML in healthcare, especially as it relates to personalized medicine. The majority of his talk focused on the use of deep neural networks for learning from electronic health record (EHR) data. Dr. Page gave examples of such ML models from his work with the Marshfield Clinic where he and his collaborators applied ML to predict atrial fibrillation, stroke, and mortality in an EHR dataset with 1.5 million records that covered 40 years of calendar time. In addition to presenting the technical highlights of this work, Dr. Page also pointed to the potential risks of applying ML in a healthcare setting, including the risk of ML learning from datasets with built in biases, as well as data and privacy issues. Perhaps the most intriguing part of Dr. Page’s presentation was when he discussed the potential of ML for drug re-purposing and discovery of previously unknown adverse reactions to existing drugs using EHR data. Dr. Page concluded his talk with a discussion of the technical issues related to causal inference in ML models and some suggestions as to how deep neural networks could help evaluate causal associations.
The poster session was chaired by Marlina Nasution from Parexel and Qing Yang from Duke. The judges were Sin-Ho Jung from Duke and Michael Sonksen from Eli Lilly. 16 posters were submitted and presented from industry, academia and government. They covered a variety of topics and highlighted the great research and activities of the DISS 2021 participants. Best poster awards went to: Daniel Hatch, Duke University, for “Analysis of a Stepped Wedge Cluster Randomized Implementation Science Trial to Improve CHG Bathing Compliance;” Jennifer Marcello, Parexel, for “Challenges and Solutions for Adaptive Design COVID-19 Treatment studies;” Honghe Zhao, NCSU, for “Outcome-Adjusted Balance Measure for Generalized Propensity Score Model Selection”, and Nan Miles Xi, UCLA, for “Statistical Modeling in Pre-clinical Drug Proarrhythmic Assessment.”
A virtual job fair was organized with the leadership from Shuyen Ho, UCB. It started with the introduction by the participating companies on general information and potential job openings and it ended with a Q&A session.
The Journal of Biopharmaceutical Statistics (JBS) will publish a Special Issue based on the invited talks and posters. Organizing Committee member Herbert Pang, Genentech and Nelson Lu, FDA/CDRH will serve as co-guest editors.
This is the 7th symposium organized by the Department of Biostatistics and Bioinformatics, Duke University School of Medicine and is led by faculty member Xiaofei Wang. It is co-sponsored by Amgen, Boehringer-Ingelheim, Eli Lilly, Genentech, ICSA, IQVIA, Janssen, Merck, NC-ASA Chapter, Otsuka, PAREXEL, PPD, SAS, UCB, and ViiV. This symposium was established to discuss challenging issues and recent advances related to the clinical development of drugs, biologics and devices and to promote research and collaboration among statisticians from industry, academia and regulatory agencies.