From last Friday to Sunday, Erasmus MC hosted OHDSI Europe 2019 in Rotterdam, Netherlands. We had 260 collaborators from 27 different countries from across government, academia, pharmaceutical industry, technology, health systems and patient representatives, all come together to participate. While we had strong representation across Europe, it was also very fun to see colleagues from South Korea, Singapore, Taiwan, Saudi Arabia, and Australia all making the trip as well to make it a truly global community. It was a truly inspiring event, so I thought I’d post a quick recap for those who missed it.
The theme of the OHDSI Europe 2019 was ‘The Journey from Data to Evidence’. @hripcsa delivered the keynote address, focusing on the opportunity for an international community to come to get to scale the process of reliable real-world evidence generation, highlighting the recent OHDSI LEGEND hypertension study as a demonstration of what’s possible if everyone joins the journey.
@Rijnbeek presented on EHDEN, the groundbreaking IMI project that is building a EU-wide observational data network using the OMOP CDM and OHDSI tools. He announced EHDEN’s Open Call for Small/Medium Enterprises (SME) Certification: https://www.ehden.eu/. SMEs will be part of the data/technology ecosystem to support health systems and other data holders in applying data standards and enabling analytics.
@Daniel_Prieto , Niklas Noren (Uppsala Monitoring Centre [UMC]), and I highlighted motivating use cases for real-world data. Dr. Prieto focused on clinical characterization, and specifically the value that drug utilization studies (summarizing the prevalence of drug use over time and across populations) can have for regulatory context. Niklas showcased the population-level effect estimation use case of pharmacovigilance, and how observational data could complement spontaneous reporting to help identify unknown issues and also support evaluation of hypotheses arising from other sources. I introduced OHDSI’s 3 analytic pillar: patient-level prediction, and how the OHDSI network could be used to develop and externally validate models for disease interception and precision medicine which could be used in clinical care.
A panel followed which demonstrated the standardized analytics tools that have been developed within the OHDSI community to support these use cases. @anthonysena showcased ATLAS, and how the open-source platform could be used to define cohorts, characterize and compare populations, estimate incidence rates, and explore treatment pathways. @jennareps performed a live demo of the PatientLevelPrediction package, which allows researchers to design a prediction study across multiple outcomes and multiple populations at-risk, execute any number of advanced machine learning algorithms, and produce an interactive web application to explore the predictive model performance. @schuemie illustrated how you can design population-level effect estimation studies in ATLAS and generate large-scale evidence for causal inference, highlighted by a live demonstration of the LEGEND Hypertension results.
@Christian_Reich provided a compelling argument for what we have to come together as a community to agree on data standards for structure, content, and semantics if we want to raise the quality of evidence we deliver and the efficiency with which we deliver it. @ericaVoss showed how the OHDSI community has made it possible for organizations to join the journey in adopting the OMOP common data model through community tools to support the ETL lifecycle, from design through implementation through testing. @mvanzandt presented on the progress of the THEMIS workgroup in establishing community-wide conventions to improve the quality and consistency of data in the OHDSI network.
The collaborator showcase was another success with > 40 posters and software demos offering a great chance for networking and shared learning, and was highlighted by the collaborator lightning talk session that showcased the tremendous progress being made around the world on the journey from data to evidence, with talks from collaborators in Belarus (great job @Alexdavv) , Saudi Arabia (well done Fatemah), Spain (thank you Leonardo), France (bravo @nthurin) and Denmark (terrific Ismail!).
Rebecca Chandler (UMC) provided a valuable clinical perspective about the potential role of observational data in pharmacovigilance, highlighting a real example of a safety signal recently uncovered during UMC’s signal detection sprint process. It was clear that while no one individual or organization had all the necessary knowledge, skills, tools, and data access to go it alone, together as an OHDSI community, we have everything we need to generate reliable real-world evidence from across our global OHDSI network in a way that could meaningfully impact decisions that matter to patients.
And of course, nobody will ever forget Peter’s AMAZING send-off, handing the reigns of the OHDSI Europe Symposium to Dani by singing an inspired rendition of ‘You’ll Never Walk Alone’! Only at an OHDSI event can you see such talent from around the world come together on their own free will for a common good, working together to improve public health and having a great time doing it! I was truly inspired and come home energized and eager to collaborate with the community to realize the tremendous potential ahead of us.
The materials and video recordings from the event will be posted at ohdsi-europe.org. I tweeted out some photos from the festivities @OHDSI, but if anyone got some good shots of the day, let’s use this forum thread to capture them. And if anyone else has thoughts or perspectives about the event, I’d love to hear them!
I want to personally thank Peter for his incredible leadership and thank the entire Erasmus MC team for hosting us for such a wonderful event. A special thanks to the OHDSI Europe sponsors for their support, and to the tutorial faculty and the AWS team who delivered 5 full-day courses to ~100 collaborators as a hands-on education on how to join the journey. I’m proud to be on this journey with all of you.