Eugene Washington PCORI Engagement Award – OHDSI Symposium Summary Report Background The Observational Health Data Sciences and Informatics (OHDSI) collaborative is a multi-stakeholder, interdisciplinary collaborative that is striving to bring out the value of observational health data through large-scale analytics. Our research community enables active engagement across multiple disciplines (clinical medicine, informatics, biostatistics, computer science, epidemiology, life sciences) and spans multiple stakeholder groups (researchers, patients, providers, payers, product manufacturers, regulators). Our mission is to transform medical decision making by generating reliable scientific evidence about disease natural history, healthcare delivery and the effects of medical interventions through large-scale analysis of observational health databases for population-level estimation and patient level prediction. While the OHDSI collaborative is a growing community, many potential stakeholders remain unaware of how OHDSI’s standardized analytic tools could support their research and decision making, or how they could get involved and contribute to future research and development within the community. To help raise awareness of observational research and the OHDSI project, the first annual OHDSI symposium was held on October 20 th 2015 in Washington DC. The OHDSI symposium had two primary objectives: increasing community engagement and improving our ability to receive feedback from a broad range of stakeholder groups; the significance of which is cyclical. As our engagement improves, our international community will continue to grow in number and diversity. Bringing together diverse skill sets and perspectives will allow us to optimize our technical infrastructure and streamline collaborative processes. With regular feedback from stakeholder groups, we can ensure future research is conducted to answer relevant health care questions. Through these efforts, our ability to facilitate open science will increase exponentially and greatly accelerate the generation of high quality, meaningful evidence. Conference Summary The symposium was a full day event with morning and afternoon sessions (see Appendix A). The morning comprised of three presentations which gave an overview of the OHDSI program, a presentation of the first OHDSI network research study and the unveiling of OHDSI’s latest open-source software tools. The aim of these presentations was to demonstrate the potential of observational research and to inform participants about OHDSI’s history, mission and future goals. Each presentation included a Q&A session to engage the broader community and ask participants how their own research or policy questions could be answered through collaboration with OHDSI. The afternoon was more hands on. It included demonstrations of OHDSI software tools, a poster session and two panel discussions designed to facilitate audience feedback. The demonstrations aimed for maximum impact by providing tangible examples of how OHDSI conducts collaborative research. The poster session gave OHDSI collaborators a chance to showcase their individual work and highlight the diverse range of research topics covered across the community.
15
Embed
Eugene Washington PCORI Engagement Award – OHDSI …presentation of the first OHDSI network research study and the unveiling of OHDSI’s latest open-source software tools. The aim
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Eugene Washington PCORI Engagement Award – OHDSI Symposium Summary Report
Background The Observational Health Data Sciences and Informatics (OHDSI) collaborative is a multi-stakeholder, interdisciplinary collaborative that is striving to bring out the value of observational health data through large-scale analytics. Our research community enables active engagement across multiple disciplines (clinical medicine, informatics, biostatistics, computer science, epidemiology, life sciences) and spans multiple stakeholder groups (researchers, patients, providers, payers, product manufacturers, regulators). Our mission is to transform medical decision making by generating reliable scientific evidence about disease natural history, healthcare delivery and the effects of medical interventions through large-scale analysis of observational health databases for population-level estimation and patient level prediction. While the OHDSI collaborative is a growing community, many potential stakeholders remain unaware of how OHDSI’s standardized analytic tools could support their research and decision making, or how they could get involved and contribute to future research and development within the community. To help raise awareness of observational research and the OHDSI project, the first annual OHDSI symposium was held on October 20th 2015 in Washington DC. The OHDSI symposium had two primary objectives: increasing community engagement and improving our ability to receive feedback from a broad range of stakeholder groups; the significance of which is cyclical. As our engagement improves, our international community will continue to grow in number and diversity. Bringing together diverse skill sets and perspectives will allow us to optimize our technical infrastructure and streamline collaborative processes. With regular feedback from stakeholder groups, we can ensure future research is conducted to answer relevant health care questions. Through these efforts, our ability to facilitate open science will increase exponentially and greatly accelerate the generation of high quality, meaningful evidence.
Conference Summary The symposium was a full day event with morning and afternoon sessions (see Appendix A). The morning comprised of three presentations which gave an overview of the OHDSI program, a presentation of the first OHDSI network research study and the unveiling of OHDSI’s latest open-source software tools. The aim of these presentations was to demonstrate the potential of observational research and to inform participants about OHDSI’s history, mission and future goals. Each presentation included a Q&A session to engage the broader community and ask participants how their own research or policy questions could be answered through collaboration with OHDSI. The afternoon was more hands on. It included demonstrations of OHDSI software tools, a poster session and two panel discussions designed to facilitate audience feedback. The demonstrations aimed for maximum impact by providing tangible examples of how OHDSI conducts collaborative research. The poster session gave OHDSI collaborators a chance to showcase their individual work and highlight the diverse range of research topics covered across the community.
While all sessions were relevant to PCORI, the last panel, “The Value and Challenges of Evidence from Observational Data: A Multi-Stakeholder Perspective,” included Maryan Zirkle from PCORI. This panel took the broadest view of observational research and OHDSI and discussed the value proposition for a range of stakeholders.
Key Findings The major outcomes resulting from the OHDSI symposium were:
Increased community engagement - Since the symposium, we have seen an increase in the number of new collaborators, higher attendance at community meetings, more activity in the online OHDSI forums, the formation of several new working groups, and an increased demand for OHDSI network research
Improved ability to receive feedback from a broad range of stakeholder groups - The symposium offered a unique opportunity to collect feedback about the value OHDSI’s open source software and network research can offer the healthcare community. Thanks to questions raised during the symposium Q&A sessions and feedback from the evaluation questionnaire, the OHDSI community can ensure our work results in meaningful research reflecting the needs of patients
Stakeholder Engagement The symposium was an excellent opportunity to foster further collaboration between academia, industry, and government stakeholders. An example of this was the formation of the symposium planning committee. To aide with planning, an organizing committee was formed and included representatives from government, academia, and industry. Members of the committee gave valuable insights on how to increase participation from each sector and aided in developing the symposium agenda to ensure each presentation was relevant and valuable to the range of stakeholders present in the audience. With the committee’s help, we were able bring in speakers for our closing panel from the FDA, Takeda Pharmaceuticals, John Hopkins and PCORI. In addition, the symposium attracted almost 300 participants from academia, government, industry, and patient groups. Having this range of stakeholder perspectives allowed the symposium to facilitate cross-sector discussions and provide insights on how OHDSI can achieve its aim of generating reliable evidence to inform medical decision making. Evaluation results and findings Overall, the symposium was well received with participants giving each presentation positive feedback. Based on responses in the evaluation questionnaire, it is clear the broader healthcare community sees OHDSI’s open source tools as a powerful platform to facilitate clinical research. However, some participants felt the symposium was too focused on software development and would have liked to see more OHDSI research. In addition, while many attendees enjoyed the multi-stakeholder perspectives shared throughout the day, some would have liked more time for in-depth discussion on specific focus areas.
Finally, many participants enjoyed the OHDSI software demonstrations and mentioned they would like to learn more about how to use the suite of OHDSI tools. Some even suggested providing how-to tutorials during the next symposium. Lessons learned The 2015 OHDSI symposium provided a wonderful opportunity to showcase OHDSI’s achievements to-date and improve our engagement efforts both across the OHDSI collaborative and broader healthcare community. To build upon this year’s success, the next symposium should include:
More network research – Based on participant feedback, there is a clear demand for OHDSI network research. To fulfill this demand, the OHDSI community should streamline collaborative research processes and create a central infrastructure to facilitate result sharing.
Break-out sessions and tutorials – To allow for more in-depth discussions on specific focus areas, the next symposium should include more structured time for break-out sessions and tutorials on how-to use OHDSI tools. This could be accomplished by offering workshops the day before or after the symposium. Alternatively, the symposium agenda could be broken down to include workshop sessions.
The symposium informed not only the next symposium, but also the actual OHDSI research direction. More emphasis will be placed on carrying out studies, rather than so much focus software development. We believe that this is a natural progression, where early work has to build the infrastructure to get started, but the collaborative must quickly emphasize concrete results to provide value and remain focused on the ultimate goal.
Dissemination Plans
Following the symposium, a thank-you email and evaluation questionnaire was circulated to all participants. Links to the conference materials were circulated as well. This included presentation slides and poster abstracts which had been uploaded onto the OHDSI.org website. In addition, videos of each presentation were recorded during the symposium. In the upcoming months, these videos will be edited and made available on the OHDSI.org website. Once available, these presentations will be an excellent tool for further sharing OHDSI’s vision and highlighting our continued progress towards our goals. The OHDSI collaborative aims to hold a symposium each year. As the time comes to plan the 2016 OHDSI Symposium, links to the 2015 presentations will be circulated to potentially attendees, including patient groups, to encourage their participation both at the symposium, and the OHDSI community. One potential barrier to this dissemination plan is obtaining approval from all presenters to make their presentations publically available. Since each presenter will be asked to review and approve the content uploaded onto OHDSI.org, it is possible some will decline to share their work.
Appendix A- Conference Agenda
2015 OHDSI Symposium
Time Description
7:30 - 8:00am
Registration
8:00 - 8:30am
Introductions
8:30 - 10:00am
Welcome to the journey: Overview of OHDSI : past, present, future Speaker: Patrick Ryan, PhD, Sr. Director and Head, Epidemiology
Analytics, Janssen Research & Development
10:00 - 10:15am
Break
10:15 - 11:15am
OHDSI in action: Real-world evidence for clinical characterization Speaker: George Hripcsak, MD, Chair of the Department of Biomedical
Informatics at Columbia University Medical Center
11:15 - 12:15 am
OHDSI in action: Open-source analytics for patient-centered evidence Speaker: Jon Duke, MD, Senior Scientist, Regenstrief Institute
12:15 - 2:45 OHDSI collaborator showcase Poster session of OHDSI research Software demonstrations of OHDSI open-source tools
During this time, lunch will be provided
2:45 - 3:45pm
Panel Discussion – Experiences from the OHDSI international data network Moderator: Christian Reich, MD, PhD, Vice President of Real World
Evidence Systems, IMS Health, USA Rae Woong Park, MD, PhD, Professor, Ajou University School of
Medicine, South Korea Peter Rijnbeek, PhD Assistant Professor, Erasmus Medical Center,
Netherlands Parsa Mirhaji, MD, PhD, Director of Clinical Research Informatics at
Montefiore Healthcare System, Albert Einstein College of Medicine, USA
Marc Suchard University of California, Los Angeles
Jimeng Sun Georgia Institute of Technology
Helena Sviglin US Food & Drug Administration
Casey Sydnor Commonwealth Informatics, Inc.
Ana Szarfman US Food & Drug Administration
Nicholas Tatonetti Columbia University
Rosenie Thelus US Army
Rosenie Thelus US Army
Zoltan Thinsz Karolinska Institutet
Anto Thomas NantHealth
Tony Thomas Johnson & Johnson
Yuxi Tian University of California, Los Angeles
Jawahar Tiwari US Food & Drug Administration
Joseph Tonning US Food & Drug Administration
Don Torok Ephir Inc
Lisa Torosyan US Food & Drug Administration
Matthew Tyson AstraZeneca
Brandon Ulrich B2i Healthcare
Levon Utidjian Children's Hospital of Philadelphia
Rohm Vashisht Stanford University
Melissa Vaughan
Veronica Sansing US Food & Drug Administration
Erica Voss Johnson & Johnson
Kin Wah Fung National Institute of Health
Magnus Wallberg WHO Uppsala Monitoring Centre
Xia Wang AstraZeneca
jian wang US Food & Drug Administration
Cunlin Wang US Food & Drug Administration
Bruce Weaver US Food & Drug Administration
Rachel Weinstein Johnson & Johnson
Emily Welebob-Yost AstraZeneca
Brian Wilson Tufts University
Lesley Wise Takeda Pharmaceuticals
Jason Woo US Food & Drug Administration
Yonghui Wu Google
Suji Xie US Army
Hua Xu University of Texas
Alexandre Yahi Columbia University
Qin Ye AstraZeneca
Soo Yeon Cho Ajou University
Zhong Yuan Johnson & Johnson
Lilly Yue US Food & Drug Administration
Richard Zhang US Food & Drug Administration
Changgeng zhao Novartis
Yueqin Zhao US Food & Drug Administration
Vivienne Zhu Medical University of South Carolina
Maryan Zirkle PCORI
Appendix C - Evaluation Questions
1. How valuable did you find each component of the OHDSI Symposium? a. Welcome to the journey: Overview of OHDSI : past, present, future (presentation by
Patrick Ryan) b. OHDSI in action: Real-world evidence for clinical characterization (presentation by
George Hripcsak) c. OHDSI in action: Open-source analytics for patient-centered evidence (presentation by
Jon Duke) d. OHDSI collaborator showcase (poster presentations and software demonstrations) e. Panel Discussion – Experiences from the OHDSI international data network (moderated
by Christian Reich) f. Panel Discussion – The Value and Challenges of Evidence from Observational Data: A
Multi-Stakeholder Perspective (moderated by David Madigan)
2. How did you find the breakdown and time distribution between presentations, posters, demos, and panels? What would you like to see differently at future symposiums?
3. At the the OHDSI Symposium, we discussed the type of work we're collaborating on in terms of two dimensions: analytical use case (data management, clinical characterization, population-level estimation, patient-level prediction) and lifecycle (methods research, open-source development, clinical applications). We highlighted work across all use cases throughout the lifecycle, with particular emphasis on the foundational community-building activities to allow all stakeholders to get started in collaborating with the OHDSI community. Moving forward, please help us prioritize where OHDSI should focus across these dimensions for future community events (and offer suggestions for any other areas we may be missing):
a. Observational data management i. Methodological research
ii. Open-source analytics development iii. Clinical applications
b. Clinical characterization i. Methodological research
ii. Open-source analytics development iii. Clinical applications
c. Population-level estimation i. Methodological research
ii. Open-source analytics development iii. Clinical applications
d. Patient-level prediction i. Methodological research
ii. Open-source analytics development iii. Clinical applications
e. Other, please specify.
4. Do you have any feedback on the logistics of the: a. Website content and registration b. Call for participation: posters/demonstrations
c. Conference location and venue
5. Do you have any other comments, questions, or concerns?
6. What stakeholder group(s) do you represent? (Government, Academia, Industry, Health System/provider, Patient, Other)
7. What disciplinary perspective(s) do you represent? (Epidemiology, statistics, informatics, computer science, clinical science, health policy, other)