The September 2018 Health Level 7 (HL7) international working group meeting was held in Baltimore, Maryland and was packed full of user tutorials and learning sessions. Because this was my first HL7 working group meeting, I checked out my colleague’s report (“What to expect at the HL7 working group meeting: A first timer’s experience”), and I’d like to share my experience of the event.

Overall, the working group meeting focused on driving interoperability and advancing healthcare IT. HL7 did a great job of providing resources and keynote speakers from within the informatics community. The keynote discussions were robust and mostly a high-level overview of the direction HL7 is working towards. However, the speakers were at times overly technical for someone with a clinical rather than a technical background (like me), as HL7 has a large interest in developing applications. Don’t let this discourage you from attending HL7—the presentations were mostly understandable, and the speakers gave a good overview of the hot button issues that healthcare IT needs to address:

• Prohibiting information/data blocking: The ONC is incentivizing data transparency, issuing fines for data blocking and working towards national implementation of interoperable data across platforms (for a quick overview, read “TEFCA: A beginning for nationwide interoperability”).
• Value-based care and the Patient Triple Aim: Patients are suffering due to the slow pace of healthcare IT improvement; there’s a need for workflow integration beyond data aggregation. Healthcare needs to enhance the patient experience, focus on health populations, and reduce per capita spending.
• Clinical Information Model Initiative (CIMI) and Fast Health Interoperability Resources (FHIR): There is potential for achieving data interoperability through using FHIR profiles derived from CIMI (for a quick, historical overview, see “On convergence of open standards – CIMI and FHIR”). Applications should be developed by utilizing standardized information and creating machine readable definitions that are SMART (specific, measurable, attainable, realistic, timely).

One of the presentations I found particularly interesting was given by Dr. Stan Huff (co-chair of the CIMI working group and chief medical informatics officer at Intermountain Healthcare). During his talk “What is CIMI doing and why interoperability,” he said “Healthcare was once simple and ineffective, now healthcare is complex and dangerous,” indicating that 256,000 deaths occur each year due to medical errors, roughly 600 deaths per day. This was a reminder of why I decided on a career in medical informatics.

Leading up to this year’s event, I started participating in the HL7 CIMI working group, collaborating to develop information models to enhance health information technology. CIMI meets twice a week for two hours and is focused on improving interoperability of healthcare systems through shared implementable clinical information models. My experience with CIMI thus far has included a steep learning curve and understanding there’s no “how to learn HL7 for dummies” guidebook. Meetings are mostly agenda based, focused on community submitted use cases that quickly dive deep into real world interoperability challenges. Initially, CIMI was difficult to follow, requiring me to study previous discussions and familiarize myself with the terminology and references used during meetings. The CIMI working group has a reference webpage, providing information on the CIMI mission, meeting minutes, leadership contact information, and the listserv, your source for following scheduled CIMI meetings. Following the CIMI listserv was crucial for getting updates on the most active discussion topics and immersing myself into discussions, whether listening and learning or actively engaging in group discussion.

For this event, the HL7 working group CIMI discussion centered around a presentation on modeling glucose tolerance testing. The model was built with pre-coordinated terms and reconstructed to accommodate post-coordinated relationships. I see how this model will accommodate more ambiguous data to improve interoperability, but I’m curious to see what the unintended consequences of creating this type of model are, if any. The meaning or definition of data between two model structures could differ in the level of granularity and I look forward to future CIMI discussions that follow the evolution of this project.

In addition to the many available working group meetings, there are also multiple tutorials and learning sessions that participants may find value in, tutorials such as “HAPI on FHIR” or “Introduction to FHIR” – my learning session of choice for first time HL7 attendees (for another perspective on learning FHIR, see “How I learned HL7 FHIR as a clinician without losing my mind”). This is just one of the many working groups our 3M Healthcare Data Dictionary (HDD) team participates in and I encourage those who are new or wanting to learn more about HL7 to read the insights reported in their blogs.