The AMIA 2023 annual Symposium was held on Nov 11 – 15 in New Orleans. It was a record-breaking event with more than 2,300 attendees, making it the highest-attending AMIA Annual Symposium of all time. Out of the 2,300 attendees, 68% were AMIA members, 32% were non-members and 23% were students. 49 out of 50 states plus Washington DC and Puerto Rico as well as 28 Countries were represented (8% of the attendees were international).
I have attended AMIA in the past but haven’t for many years and my experience at the conference felt a bit like when you go back to your old school and things are somewhat the same but the people are all new. Overall though, the content, the presenters, the organization and the exhibit hall all impressed me and it was fun to re-engage with the AMIA community again.
Since the conference is so big and has so much content I focused on two topic areas; Clinical Decision Support (CDS) and Population/Public Health (HLN staff are active in both the CDS and Public Health workgroups). In the public health area it’s worth to mention a very good overview of the Data Modernization Initiative (one of the better summaries of the drivers of DMI and all its components I have seen so far) and a presentation of evaluation of data quality in the Immunization Information System in Missouri, where both the methods used for the study and the result is applicable to other public health datasets as well. Not surprisingly the two major issues found were 1) Duplicated records of patients, immunizations, providers and clinics and 2) incorrect mapping of race categories and erroneous mapping of jurisdiction.
The CDS topic had a fairly large presence at the conference, and there were a lot of interesting presentations on how to best implement CDS that will assist the providers in their decision making while still allowing for flexibility in their actions. One presentation that stuck with me was: “Dynamic Tools and Metrics for Success”, by John D. McGreevey III, MD. Dr McGreevey’s overview of the core principles of an effective CDS and examples of how some very simple changes could lead to a big impact was enlightening. For example, a change in the opioid prescription dispense default from 28 to 15 created an overnight shift of approximately 150 prescriptions/week to a safer amount.
I think public health has much to learn about CDS and how we could more effectively use data at public health agencies to provide information back to EHRs to support effective CDS.
During the conference I also staffed a poster submitted by the UofM and the Minnesota Department of Health, talking about how MDH started setting up governance for its data lake environment. This is work we did at MDH before I joined HLN, and I was very excited the poster got accepted in this competitive space.
As a final note, I must say I enjoyed the conference very much, the only negative was how hard it was to find the public health content as it was embedded into sessions with more clinical focus. That said, I think it’s beneficial to learn and share across healthcare and public health, so I didn’t really mind. I was also a bit sad to see how few of my public health informatics colleagues were in attendance, although I very much appreciated the public health workgroup meeting, and the opportunity to connect with the ones that were there.
Here are a couple of interesting reads related to CDS:
An overview of clinical decision support systems: benefits, risks, and strategies for success