Interoperability is critical for the future of healthcare and the desired continuum of care, yet it continues to be a challenge faced by healthcare providers worldwide. Advances in technology have made it possible to move healthcare data efficiently and securely from one system to the next. To achieve this, conventional healthcare software vendors are opting to layer open standards APIs on top of their existing monolithic platform to comply with regulations and provide interoperability whereas new healthcare software vendors are architecting their platform ground up using an open standards framework. However, in this context, the architecture and choice of interoperability technology and standards to use in healthcare IT systems are still largely subjective and driven by human and engineering preferences.
Over the course of three blogs, I explore two well-established interoperability standards, Fast Healthcare Interoperability Resource (FHIR) and open Electronic Health Record (openEHR) and provides a comparison of these standards in an open healthcare platform setting. These standards are still evolving and as described, each approach has its own strengths and weaknesses.
Introduction
In the last two decades, digitalisation of healthcare information has advanced at pace and many hospitals in the developed world have adopted an electronic health record (EHR) to manage clinical and administrative data. These healthcare systems such as Electronic Medical/Patient Record (EMR/EPR), Patient Administration System (PAS) and Laboratory Information System (LIS) generate large amounts of data continuously. In addition to the sheer volume and importance of this data, the data sets are stored in a variety of formats consisting of structured data (vital signs, diagnostic results, medication administration, etc.) and unstructured data (images, radiology reports, progress notes, etc.).
Adoption of different standards and technologies to manage this data can improve the availability of information, efficiency, effectiveness and quality of patient care. It has also demonstrated improvement of the financial performance, operational capacity and administrative workflows. To realise these benefits, it has been necessary for healthcare providers to connect these disparate systems and share the siloed data. Despite recognising the existence of these benefits, the interoperability goal remains elusive in the healthcare industry and is a significant issue that is impacting the care delivery system.
There have been various attempts to standardise the format and structure of healthcare data, however, many of these standards have met with technological, regional and functional barriers to achieve the ultimate success. Two popular standards namely, FHIR and openEHR, along with their corresponding implementations highlight a promising concept that could influence the future of healthcare systems development and remove the barriers faced by legacy standards. These new standards have gained momentum due to technological advancements such as introduction of the serialised lightweight data format JSON as well as engagement from the community of adopters, developers, clinical data modellers and implementors.
FHIR and openEHR both achieve their intended purpose, that is, to create an open healthcare ecosystem by defining standards and allowing open exchange of data. They also provide additional advantages such as security, consent, reusability and terminology binding. They share several similarities while at the same time differ in their philosophy and approach:
- The FHIR standard has been recognised and accepted as the “go-to” interoperability standard by many healthcare organisations and industry partners. In addition to being the “gold standard” for interoperability, FHIR is also proving to be successful as a standalone clinical data repository format that can be effectively used to store data and build an open healthcare platform.
- openEHR’s vision, architecture, standards and specifications have generated strong community interest and present it as a strong contender in the open healthcare platform movement. The openEHR models are publicly available and have shown a great amount of reusability for various models of care. Due to the design complexities and steep learning curve, the adoption for a large scale EHR design has been somewhat limited.
Benefits
More than most, the healthcare environment is complex due to the diversity of tasks involved in the delivery of patient care. This complexity has inadvertently led to the creation of a variety of closed specialist systems to capture, manage and analyse health information. Due to this approach, any new entrant in healthcare information technology faces a significant barrier to succeed and innovate.
As the healthcare setting moves towards value-based and patient-centric models, the ability to capture, analyse and share information will determine the success of such programs. Using a common data standard is a foundation that allows different systems to conform to an agreed data model. This means any application built on an open and common data standard will exchange data seamlessly and interoperate in a cohesive environment.
The general philosophy of an open platform approach is that it must be vendor neutral and technology independent. These characteristics eliminate system lock-in and promote innovation in sync with the ever-evolving technology.
Other benefits include:
- Continuum of care
Most patients transition through different care settings, providers, clinics, hospitals, etc. during an episode of care. At each step of the healthcare value chain, important data about the patient is produced, creating a longitudinal medical history. This data is often siloed across disparate systems and many care providers. With an open standards and platforms approach, it is easier to gain access to all this data in a timely manner thus improving the quality, continuity and access to information for clinical teams caring for the patient. Full visibility and accessibility to consolidated patient data for both the healthcare institution, care providers and the patient is the primary benefit of interoperability.
- Regulatory compliance
In recent times, several countries have highlighted the importance of providing healthcare data to statutory bodies, patients and care providers and have enacted policies to enforce this requirement. Such regulations dictate that healthcare IT vendors must ensure that the data is accessible in a machine-readable format.
- Operational efficiency
Timely access to data leads to informed decision making, better planning, controlled resource management and advanced treatment plans. Integrating information from disparate systems can also drastically reduce the duplication of data and repetitive tasks in an organisation. The outcome is an increase in the overall throughput and quality of patient care. Applying technologies such as artificial intelligence and machine learning to the available data further streamlines the administrative and clinical functions.
- Financial performance
Interoperability and open standards-based platforms will continue to play a major role in the development of new models of care such as value-based care and patient-centric models. They will also provide other direct financial benefits such as simplified installation and maintenance, reduced costs to integrate, etc.
- Innovation
There is no debate about the advantages of open standards to achieve interoperability on all levels in healthcare IT. For healthcare systems to reduce human error and risk to quality of care while improving access to medical history and supporting innovation, the systems must adhere to certain data interchange standards. Adoption of standards not only enables easy construction of information from disparate sources but also allows complex reviews of such information to gather knowledge that can fuel clinical decisions.
Healthcare standards exists on many levels, for example, data formats, terminologies, concepts, communication and messaging specifications, functional specifications, etc. The focal point of the healthcare industry has been the data format and elements to exchange data consistently, efficiently and accurately. The two frontrunners, FHIR and openEHR, aim to address these initiatives to support an open healthcare ecosystem.
FHIR
FHIR is a set of specifications and standards developed by HL7 International for the electronic exchange of healthcare information. FHIR was developed as new specification building on industry approaches such as REST, JSON and search and informed by previous standards such as HL7v2 and CDA. The design intent of FHIR is to enable applications and systems to extract data from healthcare systems and expose them via modern RESTful web APIs in a machine-readable format (json/xml).
FHIR, being a system-based approach, is specifically designed for fast and easy integration of dissimilar systems. It is built using discrete categories of information called “Resources” and represents generic clinical data models and templates containing data elements for different types of clinical and administrative functions in a healthcare setting. Currently there are approximately 150 FHIR resources supporting concepts such as Patients, Encounters and Observations defined across the healthcare domain and care settings. These resources are further extended and customised through profiles to add additional attributes or enforce constraints.
Key design principles of FHIR are:
- Prioritising implementation: This is the fundamental principle of FHIR aimed at wider adoption of the standard. FHIR Resources are designed to be easily understood using key clinical concepts and terminologies and are readily exchanged using industry standards, common programming languages and established data exchange technologies.
- Reuse: FHIR resources are designed to cover 80 percent of the data elements that are used in existing healthcare systems while the remaining 20 percent remain for specific use cases that can be handled by profiles.
- Forward and backward compatible: This capability allows version transparency and safe interoperability.
- Performance: FHIR resources are serialised and simpler in construction making them easier to implement and transport over existing networking protocols. FHIR also uses modern web-based technologies such as HTTP, REST, OAuth, etc. to facilitate rapid deployment and ease of development.
openEHR
openEHR is a ground-up open healthcare platform standard that describes the architecture, data model, management, storage, querying and exchange of data in EMR/EPR systems.
The openEHR standards comes in three categories:
- Abstract specifications: Platform-independent UML information models, service interfaces, language recognition grammars, etc. The abstract specification includes the reference model, service model and archetype model aimed at standardising information, computational and knowledge viewpoints.
- Implementation technology specifications: Platform-specific specifications including XSDs, JSON-schema, REST APIs.
- Conformation specifications: Defining the basis of conformance testing and product profiling.
The above principles form an ontological separation of data models, terminologies, human language and underlying technology. One of the key architectural principles behind openEHR is two-level modelling which enforces a stable reference information model and a flexible, changing domain conceptual model. Like FHIR, the information model represents how data is structured and presented in a patient record. It provides the specifications including data types, data structures, design patterns, etc. for representing information in an EMR/EPR.
The knowledge and content model incorporates clinical community input to provide a formal definition of concepts through artefacts, mainly Archetypes and Templates. Archetypes are aimed at creating a maximal set of clinical data to be reused across multiple systems and care settings. openEHR templates take data points and concepts from one or more archetypes aimed at serving a particular use case.
Unlike FHIR, openEHR is a data driven approach aimed at organisations that have capacity and desire to develop a long-term stable knowledge representation and manage their own healthcare system from the ground up. Such an initiative has its own strengths and weaknesses. For example, adhering to an openEHR model assumes everyone adopts the same reference models therefore making it challenging to make localised changes. Additionally, it requires a strong commitment from the healthcare organisation and ongoing contribution from the clinical community to maintain and extend the underlying model. At the same time, adapting an openEHR model opens the possibility for disparate systems to operate on a predetermined standard thus achieving ultimate interoperability.
Comparison
FHIR and openEHR both aim to achieve their intended purpose, that is, creating an open healthcare ecosystem by defining standards for interoperability. Both standards support exchange as well as persistence of healthcare data.
For the added capabilities, openEHR is based on hierarchies and implements a two-level modelling approach, for example, Reference Model and Archetype Model for its data structures and concepts; while FHIR embraces a generic and flattened representation of concepts through the usage of resources, bundles and profiles. FHIR has been designed on an 80/20 rule, meaning it includes the data elements expected to be used in 80 percent of the system implementations. The remaining 20 percent are categorised as custom and specific to use cases that can be extended through FHIR profiles. Conversely, openEHR reference and archetype models are designed to contain a maximal set of data elements that represent all the use cases in the medical world.
The following table provides a comparison of these standards.
With the initial release of FHIR in 2014, it created a strong interest amongst the technology community and healthcare providers. For example, the below charts from Google trends shows a steady increase in search interest in FHIR topics compared to openEHR for the given region and time.
Figure 1: Comparative Interest – FHIR & OpenEHR
Figure 2: Interest by Region – FHIR
Figure 3: Interest by Region – OpenEHR
Open Platform Architecture
Defining an open platform architecture is a non-trivial task and is often influenced by human motivations, engineering preferences and regional interpretations of healthcare practice. Regardless, the primary focus of an open platform should be to design and implement a system that is built on open, universally agreed standards using cohesive technologies.
In that context, an open platform should adhere to the following principles:
- Data models and message exchange specification based on open standards
- Vendor neutral, portable, shareable and accessible data repository
- Resource discovery supported
- Secured access to the platform through open APIs and services
- Shareable resources for reusability and de-duplication available
- Modular and interoperable architecture adopted with independently developed applications
The specific product architecture of an open platform will differ from vendor to vendor. However, at a high level the architecture will take the form as illustrated in the diagram below.
There will be a core clinical data repository (based on open standards) and an enterprise service bus to securely serve and route system events (publish and subscribe model) such as application requests, data requests, authorisation requests, etc. to the underlying clinical data repository. Such architecture not only allows the underlying data to be independent and vendor neutral but also accelerates onboarding and portability of external applications in an open healthcare ecosystem.
Conclusion
The primary goal of interoperability is for healthcare care systems to communicate efficiently and this can be achieved in many ways, including embracing open architecture and communication standards while designing healthcare systems. FHIR and openEHR both achieve their intended purpose. That is, to create an open healthcare ecosystem by specifying standards and allowing open exchange of data. They share other similarities such as security, reusability, resource discovery, querying, community engagement and terminology binding, etc. However, they differ in their philosophy and approach around definition of data elements as well as structure and governance for extensions and maintenance.
As technology is ever-changing, it is imperative that the philosophy of open standards does not influence innovation and product engineering. The aim should be to surpass the hype of which standard to select and to design a system that follows the open platform guidelines to ease innovation and interoperability. FHIR, though aimed at standardising the exchange of data, can also be used as a data model and storage format for a modular EMR/EPR. Storing clinical data in a native FHIR format directly into an independent data repository is a novel engineering effort to achieve interoperability and a modern real-time EHR. Using such an approach not only removes the barriers to entry for new entrants but also drastically improves the overall performance and usability of the application by reducing the data retrieval and processing time, thus realising the ambition of a true longitudinal patient record.
Vivek Krishnan is the Chief Technology Officer at Alcidion