Log in Register

How I learned HL7 FHIR as a clinician without losing my mind

  • Thursday, 07 December 2017 02:00
  • Written by 

The Health Level 7 (HL7) FHIR is a new interoperability standard for exchanging healthcare data. HL7 monitors implementations of FHIR and continually works to improve the specifications, in addition to providing support to those implementing FHIR at their organization. FHIR stands for Fast Healthcare Interoperability Resources and uses RESTful web services to share resources, which are how data is represented or organized. RESTful services (the acronym stands for Representational State Transfer) is an architectural style that specifies constraints and are used in FHIR to exchange data using the structure it defines. In FHIR, there are approximately 100 resources covering data such as problems and medication that are reused across FHIR “profiles.” Profiles provide general agreement on how different data should be shared and represented.

FHIR goes for the 80/20 rule, meaning resources that exist in FHIR cover 80 percent of data elements used in existing healthcare systems currently. The remaining 20 percent are specific use cases that can be dealt with as FHIR extensions. Extensions are additional resources that are created at individual organizations to cover data not handled by existing core FHIR profiles from HL7. Extensions only need to be shared outside your organization if you want to send those data to another organization. Because FHIR is a new interoperability standard that is being implemented in various healthcare organizations, it is important for clinicians with an interest in medical informatics to have at least a high-level understanding of FHIR methodologies and processes.

I have attended the various tutorials at HL7 on FHIR. There were multiple tracks to choose from such as FHIR for specifiers, FHIR for developers and FHIR for architects. All these tutorials were detailed and informative; however, as a clinician, I found a lot of the material too technical for me. It was easy to get lost or distracted while I was attending these tutorials because many times the material was beyond my level of understanding. This November, at the annual conference of the American Medical Informatics Association (AMIA), I noticed that there was a tutorial called “Clinicians on FHIR.” I decided to attend hoping that the tutorial would improve my knowledge of FHIR.

Clinicians on FHIR is an HL7 working group. In addition to the AMIA tutorial they provided this year, they have more extensive tutorials at HL7 meetings. The working group consists of Clinicians and FHIR experts devoted to testing FHIR resources of interest to clinicians, working on interoperability issues that arise from clinical use of FHIR resources, recommending enhancements to FHIR, improving clinician involvement in the development of FHIR methodology and processes, and educating clinicians about FHIR.

The tutorial was an all-day event at AMIA and I found it informative, useful and interesting. It started with a great introduction to FHIR that was high-level, yet captivating and kept me focused on the material. Next, a deep dive into FHIR was provided, using clinical examples that I could relate to, rather than technical examples that made me lose interest quickly. This gave me a more in depth understanding of the goals and use of FHIR. The FHIR website was explained, and we were shown how to navigate the specifications and what they mean. Also, resources were explained in a way that I could relate to using narrative examples, such as what a clinician would read in the chart and how this would be translated and shared with FHIR resources. Profiles, data types, terminology use, value sets, and REST services were taught and illustrated in the context of how they are used in FHIR.

After the deep dive, there was a presentation given to “tie it all together.” This was useful as it explained the way data is being exchanged today, and how FHIR is intended to be used to improve this process. Projects and organizations like CIMI (Clinical Information Modeling Initiative), SMART (Substitutable Medical Apps, Reusable Technology) on FHIR, and HSPC (Healthcare Services Platform Consortium) were discussed regarding how they support the implementation of FHIR. These different platforms are developing profiles and information models to help improve the consistency of FHIR implementation. There is currently no governing body for how FHIR is implemented. This could potentially create issues when sharing data outside of your organization. Although you may be using the resources from FHIR, your profiles and the way you are representing your data may not be easily translated by other organizations that have represented their data in a different way. The goal of these diverse groups is to provide consistent profiles and resources that you can start with, so your FHIR implementation can support interoperability outside of your organization.

After this presentation, there was a hands-on session where we went into ClinFHIR, an open source tool, to create resources. This part of the tutorial really helped someone like me who learns and remembers by “doing.” Going into the tool, creating a patient and resources, and seeing how it all looks really made everything I learned at the tutorial stick. In ClinFHIR, after you have created resources, you can also see these resources in a graph representation. This helps to see the data in a nice “picture” view. When you create the resources, you also have to pick the correct terminology codes for representing the resource. The tool is helpful in providing the list of available codes to use from the correct standard for the resource. For me, whether it is because I am a clinician or just my learning style, this hands-on session was the most valuable piece of the tutorial to make me feel like I really learned and understood the principles of FHIR.

In conclusion, as a clinician who has attended the different tutorials for FHIR at HL7, I recommend that clinicians start with the Clinicians on FHIR tutorial, perhaps at AMIA as an introduction, then maybe following up at the 2-day long tutorial at HL7, which I am planning to attend in January 2018. Clinicians on FHIR was a great experience, informative, relatable, and most importantly FUN!

Flames1920

  • 10485
  • Last modified on Monday, 11 December 2017 19:27
Rachael Howe

Rachael Howe, RN, MS has been a member of the 3M HDD team since 2010. She is the Clinical Informatics manager co-leading the Healthcare Data Dictionary (HDD) team. She directs the team in process flow, operating procedures, and information and terminology modeling, in addition to designing for standardization and interoperability in various projects. Rachael leads projects that integrate standard terminologies such as SNOMED CT, LOINC, ICD-10-CM, ICD-10-PCS, CPT and RxNorm into the HDD, as well as developing content for customer specific use cases and integration of the HDD into other products within 3M. Her interests include using terminology models and principles to improve interoperability in EHRs, implementation of value sets and map sets, in addition to content support for 3M clinical NLP solutions. Rachael obtained her Masters of Science in Healthcare Informatics from the University of Colorado Denver.