Understanding FHIR: The Critical Pathway to the Future of Healthcare — The Evolution of HL7 and the Significance of FHIR

This article is reprinted from e-Healthcare.

FHIR, or Fast Healthcare Interoperability Resources, can be literally translated as “Fast Healthcare Interoperability Resources.” FHIR supports concise, rapid, and effective clinical data sharing across different systems, aiming to facilitate broader healthcare data exchange and ultimately advance population health management. For this reason, FHIR is regarded as representing the future of interoperability.

In June 2019, Dr. Russell Leftwich, Senior Clinical Advisor for Interoperability at InterSystems, visited Beijing to exchange insights on “FHIR and Interoperability.” This report is based on the content of his lecture. Dr. Leftwich currently serves as Senior Clinical Advisor for Interoperability at InterSystems and concurrently holds the position of Assistant Professor of Biomedical Informatics at Vanderbilt University School of Medicine. With an engineering background and over two decades of experience in healthcare, he is dedicated to promoting and guiding healthcare institutions in enhancing medical health services through the adoption of Health Information Technology (HIT) and Health Information Exchange (HIE) concepts. He actively participates in international and domestic efforts to develop standards for clinical data exchange interoperability, considering it his professional mission. Dr. Leftwich is a member of the HL7 International Board of Directors and also serves on the HL7 Steering Committee.

How to Correctly Understand Two Key Concepts in FHIR: “Resources” and “Profiles”? What Are the Driving Forces Behind FHIR? Starting with these questions, Russell Leftwich provides a detailed interpretation of the original intent, development history, and value of FHIR.

Scan the QR code below to register and watch the full video of Dr. Russell Leftwich’s speech.

The healthcare industry is data-intensive. In its evolution from informatization to intelligence and ultimately to smart healthcare, data accumulation has experienced explosive growth. Data generated internally by medical institutions and acquired through models such as “Internet + Healthcare” will interact within the information platforms of these institutions via various information systems. Through such interactions, new valuable data are formed, providing evidence and support for clinical care, management, and doctor-patient services.

According to IDC Digital’s forecast, the volume of healthcare data will reach 40 trillion gigabytes by 2020, representing a 30-fold increase from 2010. Meanwhile, IDC has also provided industry projections for the next two years: as healthcare institutions become increasingly digitalized, patients are now not only willing to retrieve health-related information for their own benefit but are also capable of independently measuring, collecting, and sharing health data. With advancements in nanotechnology, an growing number of Internet of Things (IoT)-enabled sensors capable of measuring and automatically transmitting data have emerged, making patient-generated data a valuable source of information for clinical decision-making.

Meanwhile, driven by policy directives, the implementation of models such as medical consortia, regional medical centers, and hospital groups has led to a significant increase in data exchange among healthcare institutions. Data sharing and utilization within specific regions have become an important trend. Faced with diverse data sources and interactions, the challenge of enabling seamless “communication and dialogue” among the more than 100 information systems within healthcare institutions, and ensuring smooth and efficient information transmission and utilization across different healthcare institutions, has become imperative to address. As circumstances demand, the topic of “interoperability” has garnered unprecedented attention.

When discussing interoperability, a key organization that must be mentioned is HL7 (Health Level 7, hereinafter referred to as “HL7”). For over three decades, HL7 has been deeply engaged in the field of healthcare data exchange and information modeling standards, continuously evolving to adapt to the digital transformation of the healthcare sector. Health Information Exchange Standards—standardized health information transmission protocols—serve as the protocols for electronic transmission between different applications in the medical field. HL7 consolidates standard formats used by various vendors to design interfaces between software applications, enabling data interaction among heterogeneous systems across healthcare institutions. Since its establishment in May 2006 (formerly known as the HL7 China Committee), HL7 China has relentlessly promoted the development and adoption of HL7 in China, gaining widespread recognition in the process.

Tip: About HL7

Health Level Seven (HL7) was founded in 1987, when Dr. Sam Schultz facilitated the establishment of the HL7 organization and its communication standards at a conference hosted by the Hospital of the University of Pennsylvania. With the participation of numerous users, vendors, and consulting organizations, the HL7 community has continued to grow. HL7 International is a non-profit, volunteer-based organization. It comprises thousands of members, including both corporate and individual members. HL7 has more than 40 work groups, each dedicated to specific domains such as medications, genomics, and emergency care. The objective of these work groups is to develop or maintain standards relevant to their respective areas of responsibility. HL7 International holds three conferences annually, all of which are free and open to the public, allowing anyone to attend, contribute ideas, and discuss relevant issues. HL7 has 35 national affiliates, including HL7 China. These affiliates aim to address localization needs, accommodate local customization requirements, and resolve issues related to local implementation.

*Content edited and compiled from the HL7 China official website

HL7 has evolved through various versions over time, with each new release designed to better align with industry needs and development. Well-known examples include HL7 Version 2.x, HL7 Version 3, and the CDA (Clinical Document Architecture) standard. However, these standards are no longer sufficient to support interoperability among the complex and diverse information systems in healthcare institutions, prompting the emergence of a new interoperability standard—FHIR (Fast Healthcare Interoperability Resources).

Tip: Major Version

HL7 Version 2 (HL7 V2) is the first information exchange standard developed by HL7. Although HL7 V2 is used in other contexts, it is currently one of the most widely adopted prominent standards in inpatient settings around the world. HL7 V2 employs messages composed of various reusable segments. These messages are used to transmit healthcare-related information between sending and receiving systems, as well as to trigger related actions (such as patient transfers, laboratory test orders, etc.). Furthermore, HL7 V2 supports notification-based unidirectional communication, as well as queries and other workflows. HL7 V2 is currently the most widely used interoperability standard in the healthcare industry worldwide, with approximately 80% of global systems utilizing HL7 V2 to achieve interoperability. Although HL7 Version 2 lacks certain capabilities found in FHIR, such as REST APIs and other features specific to FHIR, the HL7 V2 standard will continue to deliver value and play a significant role in the future.

HL7 Version 3 (HL7 V3) is the next-generation messaging standard of HL7. It introduces, for the first time, a common Reference Information Model (RIM), a data types model, a set of vocabularies, and a formal methodology for standards development. Furthermore, as an alternative architecture for messaging to share healthcare information, HL7 V3 also introduces the use of “documents” (see the comparison with CDA below). Although the term “V3” nominally covers both messaging and documents, it usually refers to “V3 messaging.” Currently, ISO has adopted the data types that form the foundation of HL7 V3 as part of ISO 21090. The HL7 RIM has also been adopted as an ISO standard. The HL7 V3 messaging standard has been adopted by many large-scale projects, particularly in the area of electronic health records (EHRs), although it has not yet achieved the market penetration of HL7 V2. Meanwhile, other standards development organizations (SDOs) and projects that have not fully adopted the HL7 V3 methodology have also adopted the HL7 RIM and ISO 21090 data types. Most of the comments and guidelines provided here will also apply to such solutions.

Clinical Document Architecture (CDA) is the most widely adopted HL7 V3 standard. CDA not only provides a standardized header containing metadata about the corresponding document (clinical document), but also supports a wide variety of clinical content organized into sections. The content of a CDA document can range from unencoded files, such as PDFs, to fully encoded V3 structured document instances.

*Content edited and compiled from the official HL7 China website

Russell Leftwich, a member of the HL7 International Board of Directors and the HL7 Executive Committee, as well as Senior Clinical Advisor for Interoperability at InterSystems, stated: “The need to transfer vast amounts of data between systems was a primary driver behind the creation of FHIR. The FHIR standard was proposed in 2011, with development commencing that same year under the leadership of HL7 International. FHIR facilitates translation between standards, enabling legacy systems that rely on older standards to continue exchanging information within the health information data ecosystem. Its development has drawn upon the expertise and insights of numerous industry veterans with ten to twenty years of experience, effectively addressing the complex challenges of interoperability among heterogeneous data.”

InterSystems was among the first companies globally to support the FHIR standard and remains a staunch advocate for it. Currently, InterSystems HealthShare Unified Care Record fully supports the FHIR standard, enabling hospitals and clinics to better analyze and present data, while helping partners significantly shorten application development cycles.

At a sharing session in June 2019, Russell Leftwich interpreted “FHIR and Interoperability,” tracing the original intent, development history, and value of FHIR, thereby providing valuable insights for the industry to correctly understand FHIR!

In the view of this HL7 veteran, the primary concept of FHIR is “resources.” Russell Leftwich believes that the foundation for building FHIR solutions lies in a set of modular components known as “Resources.” By investing only a small fraction of the cost of existing solutions/systems, these resources can be easily assembled into effective systems capable of addressing practical clinical and administrative challenges.

“Resources” serve as the foundation for constructing the entire data framework. FHIR endpoints are accessed via URLs, which provide the storage location of FHIR resources, thereby enabling precise locating of a specific FHIR resource. Through the URL, one can determine whether the resources within a given scope reside on a local server, in the cloud, or at other healthcare institutions, allowing for direct resource localization. Furthermore, “Resources” represent data with well-defined semantics, specifically referring to healthcare data; this involves interpreting their content and categorizing them into distinct resource types.

Russell Leftwich emphasizes that “Resources” is an independent concept, broader in scope than data sources but narrower than a patient’s entire medical record. Taking a patient as an example, the patient constitutes a FHIR resource, which contains attributes such as data elements including name, date of birth, gender, and national ID number. Medication represents another FHIR resource, encompassing attributes such as drug name, code, and dosage form. Family history is yet another resource, listing potential disease histories among family members. All such data ultimately form lists, which constitute a particularly interesting type of resource. A list is not a concrete clinical concept but rather a collection of related data or content—for instance, a set of medications for a patient or a group of patients under the care of a specific physician. These are all represented using lists; therefore, lists serve as foundational infrastructure resources.

The core philosophy of FHIR is to reuse all such data and structures.

In Russell Leftwich’s view, “Profiles” are the second most important concept in FHIR.

“Application Profiles” are collections of all information required for specific business scenarios and use cases, forming the foundation of FHIR implementation and containing constraints for those use cases. “In other words, merely having FHIR resources does not enable interoperability; rather, interoperability is achieved by possessing and sharing these profiles,” stated Russell Leftwich.

“Referencing”—the ability to reuse concepts and data across different paradigms—is a crucial concept in FHIR. For instance, certain clinical decision support services can leverage FHIR data without requiring any data transformation; the data can be directly copied and pasted into other paradigms to achieve the desired functionality. Currently, no other interoperability standards offer this capability. In other words, when using other standards, data must undergo transformation when being transferred between different interoperability paradigms.

Russell Leftwich stated, “True interoperability means enabling the interoperability of the actual semantic meaning of healthcare data—that is, allowing systems to recognize and understand the true meaning of such data. Just as we need a unified standard to distinguish between ‘a white horse with black stripes’ and ‘a black horse with white stripes,’ we require standardized frameworks to ensure accurate interpretation.”

Although FHIR is only a draft standard, compared to five years ago, an increasing number of users have begun to adopt FHIR, using it with greater frequency and across a broader range of applications. Currently, FHIR has been updated to its fourth version, and as more users implement it, FHIR will undergo a continuous process of maturation.

REST API is a critical driver of the FHIR standard, aiming to share and exchange data and data models across systems. As vast amounts of data have migrated from offline to online environments, REST APIs have been widely adopted in various industries during their digital transformation, including the Internet of Things (IoT) and mobile devices. This same technological approach is being applied in the healthcare sector. The core rationale for its adoption lies in enhancing healthcare data interoperability—enabling seamless data exchange among internal systems within healthcare institutions or between different institutions—and leveraging the insights derived from such data to improve clinical decision-making.

“Furthermore, as we know, human decision-making is limited to five factors, necessitating computer-aided decision support. This intensifies the demand for interoperability, namely, data sharing and exchange,” said Russell Leftwich, discussing its real-world applications.

“The FHIR API enables easy access to and retrieval of small subsets of patient data, such as current medication lists or historical trends in blood glucose or blood pressure levels. Thus, it can efficiently handle small data fragments or data of a specific type. ‘When the results are returned, they are presented in a machine-readable format, allowing decision support systems and other applications to easily interpret and utilize this data,’ further explained Russell Leftwich.”

FHIR standards are available for download from the internet, and the downloaded versions are machine-readable. A key feature of FHIR is its ease of learning and understanding. “IT engineers can quickly grasp it and develop simple applications based on it in a short period,” said Russell Leftwich.

Compared with previous versions, FHIR was entrusted with high hopes of being “simple, easy to learn, and easy to use” from the outset. “HL7’s idea is to model the most commonly used 80% of data (rather than all data), that is, to identify the most common data exchanged between systems and develop corresponding models,” explained Russell Leftwich. By incorporating a relatively small set of data directly into the core standard model of FHIR, the FHIR standard becomes simple, easy to learn, and easy to use. What about the remaining 20% of data?

“For the remaining 20% of data, although use cases are limited, it is highly likely that certain scenarios may require it. To address this, FHIR introduces a new mechanism—Extensions. Through Extensions, a specific resource can incorporate this 20% of data, thereby enabling expansion beyond FHIR’s core data elements,” added Russell Leftwich.

When discussing the advantages of FHIR, Wang Lihua, Director of the Information Center at Beijing Friendship Hospital, Capital Medical University, noted that all FHIR standards are open-source and free, allowing anyone to use them at no cost to develop their own applications and achieve interoperability. In light of the hospital’s current development status and trends, the advantages of FHIR are particularly evident, such as its strong support for mobile platforms, robust security and ease of customization, excellent compatibility with emerging technologies, and alignment with relevant new standards.