DBMI Students/Faculty Lead OHDSI To Successful 5th Annual US Symposium
Columbia University, the coordinating center for the Observational Health Data Sciences and Informatics (OHDSI) network, assisted in producing another successful OHDSI Symposium Sept. 15-17 in Bethesda, Md. Speakers, panelists, poster winners/presenters and book authors from Columbia all helped share both the OHDSI mission and latest research with an enthusiastic room of more than 420 attendees.
While the main symposium was held Sept. 16, the weekend kicked off with the highly anticipated Women of OHDSI Real-World Analytics Leadership Forum. Co-interim Department Chair Noémie Elhadad, PhD, was the first of four speakers on a panel of esteemed female leaders in the field of real-world analytics. Check out the video below for a recap of that night, which includes thoughts from Dr. Elhadad.
Department Chair George Hripcsak, MD, MS, welcomed an enthusiastic audience to open the actual symposium. He provided an overview of OHDSI, an especially beneficial topic for a room that had more than 100 first-time attendees, as well as the mission, scope and tools of the OHDSI community. Much of that information is available on Dr. Hripcsak’s presentation slides.
Both Dr. Hripcsak and Adjunct Assistant Professor Patrick Ryan, PhD, concluded the opening plenary session with a presentation entitled Proving reliable real-world evidence: Replicating RCTs using LEGEND to announce OHDSI’s Large-Scale Evidence Generation and Evaluation across a Network of Databases (LEGEND) project. This study could transform the way real-world evidence is used in healthcare by, as Dr. Ryan says, “reproducibly generating evidence by applying advanced analytics across a network of disparate databases for a wide array of exposures and outcomes.”
Prior to the discussion on LEGEND, the Book of OHDSI was unveiled to the symposium. David Madigan, PhD, a Professor of Statistics at Columbia, joined OHDSI collaborator Martijn Schuemie, PhD, in co-authoring the book, which will assist both veterans and newcomers in doing research within the community. Dr. Schuemie adds insight about the book below.
Several members of Columbia DBMI presented their research during the Collaborator Showcase. The full list of posters and presenters is available on the OHDSI web site; PhD candidate Amelia Averitt (pictured, right) was voted Best Contribution in Methodological Research for the poster The Counterfactual 𝝌-GAN: An Adversarial Method To Support Covariate Balance. Assistant Professor Adler Perotte, MD, MA, and DBMI Masters graduate Natnicha Vanitchanant also collaborated on the work.
The afternoon session highlighted community evidence in action, and the Women of OHDSI closed it with an impressive presentation on Predicting Breast Cancer to improve screening. OHDSI project manager Maura Beaton, MS, opened the talk with an overview of the Women of OHDSI and how it developed this research topic, while DBMI PhD student Anna Ostropolets, MD, shared the results of the study.
Dr. Ryan closed the Symposium with a look back at 10 years of the OMOP/OHDSI Common Data Model, and why the mission of OHDSI is vital to the future of health and health care. (And as you can see in the photo gallery above, cake was heavily involved).