Principles of Health Interoperability: FHIR, HL7 and SNOMED CT

Citation

Benson, T., & Grieve, G. (2020). Principles of Health Interoperability: FHIR, HL7 and SNOMED CT (4th ed.). Springer.

Chapter Summary

The fourth edition has twenty-three chapters in about 475 pages, in the Health Information Technology Standards series. The publisher says it was completely reorganised into five sections: core principles, an extensive review of FHIR, older HL7 standards that are still widely used, clinical terminology including SNOMED CT and LOINC, and a final section on privacy, models, XML and JSON, standards development organizations and HL7 v3. The page numbering shows breaks between chapters 4 and 5, 11 and 12, 14 and 15 and 18 and 19, which match the boundaries of the five sections.

Chapters 1–4: The Health Information Revolution; Why Interoperability Is Hard; Clinical Terminology; Safety

The opening section sets out the principles. It begins with the information revolution in health care and asks why exchanging information between systems is difficult. Chapter 3 introduces clinical terminology early, ahead of the standards chapters, and Chapter 4 addresses safety. Together these 70 or so pages frame the problem before any standard is described.

Chapters 5–6: Principles of FHIR; The FHIR API

The second section starts the extended treatment of FHIR (Fast Health Interoperability Resources), which the publisher calls the most important health interoperability standard globally. Chapter 5 gives the principles of FHIR and Chapter 6 its API. The publisher says FHIR combines the best features of HL7's v2, v3 and CDA while using current web standards and clinical terminologies, with a tight focus on implementation.

Chapters 7–8: Resource Basics and Datatypes; Selected Resources

These chapters cover the building blocks of FHIR: the basics of resources and their datatypes, followed by a selection of individual resources.

Chapters 9–11: Conformance, Terminology and Profiles; Implementing FHIR; Security & Integrity in FHIR

The rest of the FHIR section deals with conformance, how terminology is handled in FHIR and the use of profiles; with implementation; and with security and integrity.

Chapters 12–14: HL7 Version 2; CDA—Clinical Document Architecture; IHE XDS

The third section covers older standards that, according to the publisher, are still widely used: HL7 version 2 messaging, the Clinical Document Architecture (CDA) and IHE XDS.

Chapters 15–18: Classifications; SNOMED CT; LOINC; Using Terminology

The terminology section covers classifications, SNOMED CT and LOINC, then how terminology is used in practice. At about 32 pages, SNOMED CT is the longest chapter in the book.

Chapters 19–23: Privacy and Consent; Models; UML, XML and JSON; Standards Development Organizations; HL7 v3

The final section covers privacy and consent, models, the UML, XML and JSON, the standards development organizations and, last, HL7 v3.

Key Concepts

Chapters 1–4: Principles

• Interoperability and why it is hard: exchanging health information between systems with different structures and meanings.
• Clinical terminology: coded vocabularies that give data a shared meaning.
• Safety: a topic placed among the opening principles.

Chapters 5–11: FHIR

• FHIR resources and datatypes: the building blocks of FHIR.
• The FHIR API: how FHIR systems exchange resources.
• Profiles and conformance: constraining and checking use of the standard.
• Security and integrity: protecting FHIR exchanges.

Chapters 12–23: Older standards, terminology and supporting topics

• HL7 v2, CDA and IHE XDS: the older standards that remain in widespread use.
• SNOMED CT and LOINC: major clinical terminologies, with classifications and their practical use.
• Privacy and consent: treated in the final section.
• Standards development organizations: how the standards are produced.

Critical Analysis

Strengths

• Authors close to the standards. Grahame Grieve is described as HL7's FHIR Product Director and the founder of FHIR; Tim Benson helped develop the Read Codes, one of the two sources of SNOMED CT, and co-chaired the HL7 Education Committee. The book is written from within the standards community.
• FHIR given the largest share. Seven chapters (5–11) occupy about 130 of the 475 pages. The publisher says this edition expands the discussion of FHIR, and the section moves from principles and the API through resources, profiles and implementation to security.
• Legacy standards and terminology included. Chapters on HL7 v2, CDA, IHE XDS and HL7 v3, and four chapters on classifications, SNOMED CT, LOINC and the use of terminology, make the book relevant to environments where older standards remain in use.
• A clear internal structure. The five-section reorganisation lets a reader with one task, such as FHIR implementation or terminology work, go directly to the relevant section.

Limitations

• Notations come late. UML, XML and JSON are covered in Chapter 21, near the end, although a reader meeting FHIR for the first time may need them earlier. Readers unfamiliar with these formats should consult Chapter 21 before the FHIR chapters.
• A moving target. The publisher describes FHIR as having rapidly become the most important standard, so the specification continues to change after a 2020 book; readers should check the current published specification.
• Little independent evaluation. No independent review of the fourth edition was found, and the publisher's description does not mention exercises, so the book looks suited to self-directed study rather than problem-driven courses.

Where It Fits in the Course

This is the standards-focused specialist text among the medical informatics books. Coiera's Guide to Health Informatics is the broad conceptual foundation, covering clinical skills, health records, decision support and terminologies without a chapter on messaging standards, and it is the natural book to read first. Hersh's Health Informatics: Practical Guide is another introductory survey, and Wachter's The Digital Doctor is a journalistic account of what happened when medicine adopted electronic records, useful for the context that makes interoperability matter.

Shortliffe and colleagues' Medical Informatics is a systems-oriented text on clinical data, decision support and records, and Hersh's Information Retrieval: A Biomedical and Health Perspective addresses searching biomedical information. Benson and Grieve go deeper than these on the standards that let systems exchange data, and are the book to read after an introduction when the work involves building or procuring interoperable systems.

How to Use This Book

The book is aimed at readers who need to understand or implement health interoperability: informatics students, software developers and analysts, and clinicians or managers who commission integration projects. Basic knowledge of health-care processes and of data formats such as XML or JSON will help; for the latter, Chapter 21 can be read first.

Read Chapters 1 to 4 for the principles. Then choose by purpose: Chapters 5 to 11 in order for FHIR; Chapters 12 to 14 for legacy interfaces and document exchange; Chapters 3 and 15 to 18 for terminology; Chapters 19 and 22 for privacy and governance. Chapter 23 on HL7 v3 is last and will matter mainly to readers who meet it in existing systems. Keep the current FHIR specification and, where possible, a test server at hand while reading the FHIR chapters.