Review of Tim Benson and Grahame Grieve, Principles of Health Interoperability with SNOMED CT, HL7 and FHIR
Posted on August 6, 2016
[Author’s note: This is the same review that I wrote for this book at Amazon.com.]
This is third edition of Tim Benson’s now standard text on healthcare interoperability. It does focus on the specific technologies listed in the title, but it’s more than just a handbook. Instead, it begins with a careful discussion of interoperability, and why building interoperable solutions in the healthcare domain has proven so difficult. The analysis is insightful, and well worth reading for any healthcare professional, informatics, or developer. The book offers no magic bullets, but understanding the nature of the problem is key developing solutions.
In the second part of the book, Benson moves on to a surprisingly detailed discussion of SNOMED CT, first considering its history and guiding principles, and then delving into the details. It may seem odd that the title of the book only mentions SNOMED. Why not ICD-10, LOINC, or any of the other vocabularies that play such a basic role in EHR systems today? In fact, Benson does discuss ICD-10, LOINC and UMLS. But SNOMED CT is presented as an example of how a vocabulary (really, an ontology) ought to be designed, and is compared in some detail with position dependent coding schemes. In my opinion, this was useful as practical example of the theory discussed in the first part of the book. But it also made me think about software design.
The remaining sections of the book focus on HL7, XDS, and FHIR. Unfortunately, the treatment of HL7v2 was brief, almost too brief. But that’s also inevitable. It’s simply not possible for a book such as this one to cover the protocol in detail. Instead, the book covers enough detail to understand the essentials of message structure, encoding, and the event model. Examples include patient identification (PID) and visit (PV1) But the main focus here is providing a reference point.
The next chapters consider HL7 V3, including the Clinical Document Architecture (CDA) and information model (RIM) . The coverage of HL7 V3 was more detailed than I expected, including a discussion of constrained models and the dynamic model. But again, much of the focus was on what has proved successful in HL7 V3, what has not, and why. In this part of the book, as elsewhere, there is a focus on common themes, and we are able to see how different technologies are used to address the same problems.
At this point, I’ll skip over the section on XDS and move on to FHIR. This shouldn’t be taken as a criticism of the chapter on IHE XDS. Rather, I feel less qualified to comment here (and this review is already getting a bit long!)
The final part of the book, co-authored with Grahams Grieve, the creator of FHIR, presents Fast Healthcare Interoperability Resources (pronounced “fire”) This is a technology based on the principles of REST and agile computing, and which incorporates many of the best ideas of earlier generations of HL7. This book does a good job of presenting the main ideas from scratch, and does not fall prey to th e temptation to leave out important topics such as extensions, conformance and profiles, and implementing FHIR. Of course, there is much about the details of the specification that is not covered, but the book does a good job of presenting the core principles and underlying concepts.
In the end, we see enough of these technologies to understand how they can be used to address problems of interoperability, and that is the point.