Sage Journals: Discover world-class research

Abstract

To achieve useful interoperability between electronic health record (EHR) systems, many approaches have been proposed. To date, none has prevailed as a clear solution. This scoping review studies 24 publications from 2014 to 2023. The aim is to streamline the understanding of current EHR interoperability expectations, practices, and problems, highlight learnings from the Levels of Conceptual Interoperability Model (LCIM), and suggest means for expediting EHR interoperability. Four interoperability levels are visible in EHR compared to seven in the LCIM: technical/foundational, syntactic/structural, semantic, and process/organization. Semantic interoperability—preserving meaning of exchanged data—is the main focus and the problem to solve. Its many expectations cause implementation difficulty. Standardization of data structures, transfer protocols, terminologies, vocabularies, and ontologies are the most common approach, but there is a lack of consensus on standards. Emerging approaches include fuzzy ontologies, natural language processing, and bidirectional transformation. Standardized data structure is not a prerequisite to useful EHR interoperability. Focusing on the state of health records rather than full system integration can expedite interoperability. Different use cases can benefit from various approaches. Artificial intelligence shows promise for handling semi-structured or unstructured data. Stronger regulations may be necessary to guide ongoing integrations.

Keywords

interoperability semantic interoperability meaningful use electronic health records electronic medical records EHR EMR healthcare standards ontology health records artificial intelligence

Introduction

Interoperability of Electronic Health Records (EHR) systems has been a growing interest for academic research over the past two decades. It is a shared need among medical academics and practitioners, but a usable form of interoperability in healthcare has not been reached. Interoperability in healthcare is not a priority of technological developments.¹ Some studies analyze and project information exchange cost-benefit,^2,3 but its benefits on the quality and safety of care are unclear.^4,5 That reinforces private ownership of health data⁶ and limits information sharing, which in turn has a negative impact on healthcare effectiveness, efficiency, and equity.⁵

Interoperability is “the ability of two or more systems or components to exchange information and to use the information that has been exchanged”⁷ (p. 1). In healthcare, those systems can be EHRs, medical devices, scheduling applications, billing applications, and more.

The Levels of Conceptual Interoperability Model (LCIM) outlines seven levels ranging from 0 (no interoperability) to 6 (conceptual interoperability).⁸ The LCIM was first developed for military simulation and modeling (S&M).^9,10 While we are not aware of its application in EHR specifically, it has been applied to Health Information Technology (HIT) in general and medical device interoperability in particular.⁹ The principles of LCIM can provide valuable insights for understanding the variance between conceptual and practical interoperability in EHR.

The seven levels of the LCIM serve as stepping stones for implementing interoperability between digital systems.⁸ Level 1 (technical) is simply a pre-requisite infrastructure integratability step. Level 2 (syntactic) introduces the need for a common data format and communication structure. Level 3 (semantic) is for ensuring there is common meaning and understanding to the exchanged data because of a common reference model. At level 4 (pragmatic), the interoperable systems take into consideration the context at which data is exchanged. Interoperability is represented by levels 2, 3, and 4.⁷ Levels 5 and 6 (dynamic and conceptual) represent composability where assumptions and constraints are understood, and purposeful abstractions of reality are possible for conceptualization.⁷

According to the LCIM also, implementation should be guided by an ontology—a formal specification of the conceptualization.⁷ It is a set of models that details the entities, their relationships, and the rules for their interoperability. Federation of data and ontologies—creating layers of each local system data structures and models—helps to keep local systems autonomous, avoids disrupting current states, and ensures only needed components are mapped with clear processes of exchange.⁸

Solving for semantic interoperability is critical for enabling logical interpretation and understandability.⁸ Seamless semantic interoperability between EHR systems remains a major challenge,^11–14 even with the existing standards.⁷ The implementation of multiple standards can sometimes lead to complexity.⁶ The LCIM is designed to cope with the obstacles expected from interoperability. It provides a good reference, rather than a measuring stick, for the progress of interoperability between EHR systems. We use it in our review and discussion as such.

To help bridge the gap between conceptual and practical interoperability in EHR, we carried out a scoping review of peer reviewed publications. Our research questions are: what learnings are there from current EHR interoperability levels, expectations, and approaches? What problems and gaps hold progression of EHR interoperability? What learnings from the LCIM are relevant to EHR interoperability? And what recommendations are necessary for a practical approach?

This scoping review applies the LCIM framework to EHR interoperability literature to connect theoretical models and practical implementation challenges. Our research questions explore EHR interoperability conceptual forms, expectations, implementation problems, and proposed solutions.

Methods

To carry out our scoping literature review, we followed the Arksey and O'Malley framework^15,16 and defined a protocol (Appendix A). We applied the extension of PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) for Scoping Reviews (PRISMA-ScR)^17,18 and adhered to the guidelines suggested by 19. PRISMA is a research review flow for identifying an evidence-based minimum set of items for reporting. A summary of the PRISMA process is in Figure 1, and the PRISMA-ScR checklist is in Appendix B.

Figure 1.

PRISMA Flow Diagram for Identifying Studies: A summary of the PRISMA steps of identifying studies, screening studies, deciding on studies’ eligibility, and finalizing the list of 24 studies to include in the review.

Search

We searched PubMed, EBSCO, Scopus, and Google Scholar databases. We expected duplicates records since a significant portion of Scopus content is indexed in Google Scholar.²⁰ We searched titles only and screened abstracts because they retrieve 88% of relevant references for reviews in medical topics.²¹

The search query included three criteria. The first was the inclusion of the term “interoperability” and its alternatives “interoperable” and “integration”. The second criterion was the inclusion of terms “health”, “medical”, “clinical”, “ehr”, and “emr” to capture healthcare-related articles using either the full terms “electronic health records” and “electronic medical records” or their abbreviations. The third criterion was for narrowing down the work to focus on approaches by including the terms “framework”, “structure”, “standard”, “ontolog” (covering ontology and ontologies), “approach”, “model”, “type”, “meaningful use”, “outcome”, and “direction”. The search syntax looked like this:

(Interoperability OR Interoperable OR Integration) AND (Health OR Medical OR Clinical OR EMR OR EHR) AND (Framework OR Structure OR Standard OR Ontolog OR Approach OR Model OR Type OR “Meaningful Use” OR Outcome OR Direction)

Each of the above criteria alone can generate a broad set of results. But it is their combination, especially the third, that allowed us to narrow down the results to those addressing overall solutions and approaches for the healthcare interoperability domain specifically. We also limited the publication years to be between 2014 and 2023, since the analysis was done early 2024. Given the rapid evolution of health IT, we were careful to contextualize findings from older studies within more recent developments in our analysis and discussion.

Study selection

We limited the search to articles, book chapters, and conference proceedings written in English. Our initial search resulted in 1651 records (299 in PubMed, 488 in EBSCO, 332 in Scopus, and 532 in Google Scholar). After ineligibility assessment, duplicate removal, and preliminary screening, we ended up with 340 records. The 340 records were screened by abstract by both authors leaving 149 records. 143 records of the 149 were retrieved and full text was reviewed (6 were irretrievable due to being withdrawn or inaccessible). From the 143 records, we identified 24 that met the eligibility criteria of discussing interoperability types, levels, expectations, standards, or approaches. Figure 1 summarizes the PRISMA selection process, and the list of the 24 publications is in Appendix C. The inclusion and exclusion criteria are part of the protocol statement in Appendix A.

Information extraction

Zotero was used for managing selected entries and extracting their metadata, which was verified. During full text reviews, Microsoft Excel was used to chart and document additional information on study’s type, main approach, aim/purpose, interoperability definition and expectations, main problems, proposed solutions/approaches, and findings/outcome. The content was reviewed by both authors and assessed for quality.

Quality assessment

Both authors ensured that the selected entries were from sources that practice peer reviews. We adopted the scoring approach of Azarm-Daigle et al.'s 2015 review²² and fitted it to the specific focus of our study on EHR interoperability. After full text reviews, both authors independently scored the entries on four criteria.

• Reference to differentiated types or levels of interoperability

• Reference to a definition of meaningful EHR interoperability

• Reference to EHR interoperability solution or approach

• Reference to standardization

Each criterion was scored as follows: 0 (no mention or inadequate discussion), 1 (brief mention of partial discussion), or 2 (comprehensive discussion and analysis). An entry could get a maximum score of 16 (4 criteria x 2 max score x 2 reviewers). Studies that scored 10 or more met the inclusion criteria. When major discrepancies in scoring occurred, the two authors discussed the differences and reached a consensus.

Information synthesis

Extracted information from the publications were grouped into the following EHR interoperability areas: levels and types, expectations, standards, ontologies, other approaches, and problems. These topics make the sections of our Findings. The consolidated learnings from all publications across the topics is in the final section of the Findings.

Findings

This paper provides an overview of some of the literature on interoperability of EHR systems from 2014 to 2023, then it deduces a consolidated view of the EHR interoperability landscape to help with expediting practical solutions. The details in the following subsections are summarized and consolidated in Table 1.

Table 1.

Summary of the scoping review of EHR interoperability levels, expectations, approaches, and problems.

Level	Expectations	Approaches	Problems	Suggested focus
1) Technical/Foundational	• Secure machine-to-machine connectivity and exchange	• Internet, cloud computing, web services	• Security	Shareability of health record from sender
1) Technical/Foundational	• Secure machine-to-machine connectivity and exchange	• Blockchain	• Privacy	Shareability of health record from sender
2) Syntactic/Structural	• Common data organization	• IMs and CIMs	• Heterogeneous systems	Integrability of shared health record into receiver
	• Common data type and format	• Transfer/messaging standards	• Lack of consensus on standards
	• Common communication protocols	• HL7 RIM	• Not practical and unscalable
		• HL7 CDA	• Legacy data
		• openEHR RM	• Semi-structured data
		• ISO 13606 architecture	• Unstructured data
		• DICOM
		• Data federation
3) Semantic	• Effective usage	• Terminologies and vocabularies (SNOMED-CT, LOINC, ICD)	• Lack of consensus on terminologies	Interpretability and usability of shared and integrated health record within receiver environment
	• Understandable content	• HL7 FHIR resources	• Lack of consensus on approach
	• Preserved meaning	• openEHR archetypes	• Lack of unified CIM
	• Shared clinical coding	• ISO 13606 archetypes	• Ontology contradictions
	• Consistent logical interpretation	• Rule-based ontologies	• Unsatisfiable results
	• Unambiguous inferences	• Fuzzy ontologies
	• Machine-readable	• Federated ontologies
	• Human-readable	• Natural language processing
	• Unstructured data interpretation	• Bidirectional transfer (BX)
	• Linguistic and cultural interpretation	• Knowledge graphs
4) Process/Organizational	• Coherence between systems	• Laws and regulations	• Conflicting needs between patient care process, clinical process, administrative process	Maintainability of received health record
	• Integrity of work procedures	• Common institutional interests	• Lack of cost benefit
	• Workflow integration
	• Ongoing alignment

Consolidated view and learnings from the scoping review of EHR systems interoperability. Column 1 includes the names of the 4 levels of EHR interoperability. Column 2 includes the expectations from each level. Column 3 includes the approaches pursued for each level. Column 4 includes the emerged problems faced at each level. Column 5 includes the authors’ suggested focus for each level that prioritizes the health record transfer over the holistic system integration

Levels of EHR interoperability

Basic levels of interoperability

Only 11,12,23 acknowledge no interoperability as a formal level of interoperability. Few authors recognize a basic level of interoperability that includes system connectivity and machine-to-machine secure exchange of data without any type of organization or interpretation of the data shared. One study (12) classify this type of interoperability as technical, in alignment with the LCIM. 14 and 24 refer to this level as foundational interoperability. The key concern at this level is the security, privacy, and confidentiality of exchanged data. Two studies (25,26) suggest blockchain as an approach for secure sharing of EHR data across providers while preserving patient privacy.

The majority of authors agree on a level of interoperability with defined data organization, format, syntax, and communication protocols. At this level, the systems accept each other’s data and share a common structure, but they may not be able to interpret the content of the data. Few studies (14,24,26) call this level structural interoperability. Others (11,12,23,27) refer to it as syntactic interoperability, consistent with Tolk et al.’s LCIM.

The technical/foundational level has direct dependency on the presence of Internet, cloud computing, and common web services. Solutions and approaches for syntactic/structural interoperability are dependent on, and should be unique to, EHRs. That’s where the focus on interoperability solutions in healthcare starts.

Semantic interoperability

Semantic interoperability is widely seen as a crucial level to reach in healthcare.²⁶ But understanding what it entails can be confusing. The European Commission’s lengthy definition can be summarized as the use of technology to exchange, understand, and use health information across linguistically and culturally disparate systems and organizations.²⁸ It is general focusing on expectations and including the need for language translation. The Electrical and Electronics Engineers (IEEE) Standard 1073, on the other hand, calls for “shared data types, shared terminologies, and shared coding”²⁹ (p. 299). The definition of semantic interoperability by Health Level Seven (HL7) introduces more variables: “the ability of two parties, either human or machine, to exchange data or information where this deterministic exchange preserves shared meaning”¹³ (p. 1060).

According to HL7, which offers some of the most commonly used transport standards,⁶ not only systems, but humans should be able to understand exchanged information. In addition to the conditions of correctness, accuracy, consistency, uniformity, and standardization, 13 state that semantic interoperability should include the ability to exchange clinical documents with both human-readable and machine-readable data. At the same time, they add that it should include the ability for documents to be correctly interpreted by automated tools without human intervention.¹³ de Mello et al.¹⁴ state that semantic interoperability aims for shared data to be understood and interpreted regardless of who is involved.

Achieving meaningful exchange of data seems to require the resolution of structural conflicts between EHR systems.³⁰ Common structuring, standard definitions, data codification, terminology, and vocabulary are required conditions.^14,26,29,31 At the same time, Dridi et al.³² call on semantic interoperability to achieve the integration of not only structured health data, but also unstructured data—no common format, syntax, or organization. This conflicts with the need for common data structures and definitions.

Adding to the confusion, there are two studies (11,26) that define semantic interoperability itself as containing two levels: partial and full. For Sonkamble et al.,²⁶ partial semantic interoperability is when the sender and receiver systems use an intermediate standard, and full semantic interoperability is when health records get reproduced in the local standards of the receiver. There is practicality to this distinction of capabilities. For Adel et al.,¹¹ partial semantic interoperability is recognizing the meaning of exchanged information. At full semantic interoperability, there is coherence between the various systems and organizations.¹¹ It may imply understanding the context of exchange or adjusting to assumptions, which puts it closer to pragmatic and dynamic interoperability defined by the LCIM.

Higher levels of interoperability

23 introduce process interoperability as a level above semantic interoperability where the integrity of work processes and procedures—relating to people, workflows, and organizational factors—can be kept up. Full semantic interoperability by 11 may align more with process interoperability. Zhang and Saltman²⁴ define organizational interoperability—above semantic interoperability—as a level where there is ongoing data sharing, use, and communication within and between organizations, entities, and individuals. Sreenivasan and Chacko²⁷ see organizational interoperability as the level when social, political, and legal entities work together for a common interest or exchange of information or both. This is a challenging and demanding state for successful HIT implementation.³³

None of the three top levels of the LCIM (pragmatic, dynamic, and conceptual)⁸ is explicitly mentioned in our reviewed publications. Pragmatic is implied with the many expectations from semantic interoperability. Both process and organization interoperability emphasize the condition of cross-organizational alignment over and above the interoperability of their EHR systems. They may align with the dynamic level, implying an ongoing integration rather than a situational exchange. Azarm-Daigle et al.²² find that laws and regulations are needed for cross-organizational sharing of healthcare data, but the existing methods are not sufficient.

Expectations from EHR interoperability

Interoperability is crucial to gain maximum benefits of health IT.¹¹ It allows the aggregation of patient history and improves decision-making support, workflow management, and evidence-based healthcare;¹¹^,34 and it contributes to reducing healthcare costs, enhancing quality of care, simplifying care access, increasing patient satisfaction, and improving clinician experience.^11,24

The majority of our reviewed publications see semantic interoperability as the highest level to achieve these expectations and benefits. Everyone associates it with the exchange of clinical and medical data between different systems or organizations while preserving its meaning and attaining usability. It requires a comprehensive approach of data integration, data querying, and a consistent interpretation.^12,30,34,35 The interpretation has to be logical, meaningful, precise, and accurate.^{14,27,36–38} It should result in an unambiguous inferences and effective usage of the data.^24,29,31,33

The role of EHR interoperability standards

Adoption of common standards across different organizations and providers is one way of achieving interoperability that preserves meaning in EHR systems. 11 and 12 identify 22 different Information and Communication Technologies (ICT) standards in healthcare by seven independent international organizations. They cover the syntax (structure) and semantics (meaning).¹¹ This is not a review of available standards, but understanding the role they play along the interoperability levels is necessary.

In reference to the International Organization of Standardization (ISO), interoperability of EHR systems requires: information and reference models, data structure definitions, standardized interface models, a standardized set of concepts and their relationships (ontology),²⁶ and standardized terminology—a common labeling dictionary of the data content—related to controlled vocabularies.^28,35 Terminologies can be a subsets of ontologies.^26,35

Information models and transfer standards

An Information Model (IM) or Clinical Information Model (CIM) is a structural standard for health data representation;^31,39 it encompasses the technical specifications that define how clinical information is organized inside an EHR system or for communication between systems,³⁹ also known as transfer, interface, or messaging standard.^11,26

Standards like HL7 Reference Information Model (RIM), openEHR Information Reference (RM), and ISO 13,606 standard information architecture are references to CIMs.^35,40 They are open standards that health institutions or EHR vendors can adopt and adapt as needed. HL7 Clinical Document Architecture (CDA) is a document standard²⁶ that specifies the exchange structure.^35,41 The Digital Imaging and Communications in Medicine (DICOM) standard is most used for radiology, cardiology, and radiotherapy.⁴² The Integrating the Healthcare Enterprise (IHE), an initiative by healthcare professionals and industry, promotes the coordinated use of established standards such as DICOM and HL7.⁴³

CIMs are recognized as essential for achieving semantic interoperability and creating standardized interoperable EHR systems.^38,39 Currently, each system may have its own CIM. It is ideal to develop a unified best practice or common approach for clinical information modeling, to promote correctness, reliability and quality of CIMs.³⁹ When combined with terminologies and vocabularies, information models become knowledge models forming multi-level frameworks with a set of modular components,⁴⁰ like patient profile, doctor visit, or medication. HL7’s Fast Healthcare Interoperability Resources (FHIR) call these components resources,⁴⁴ while openEHR and ISO 13,606 refer to them as archetypes.^40,45,46

Terminology and vocabulary standards

The following are some of the most popular global medical terminology and vocabulary standards used for labeling conditions, observations, and diseases: Systematized Nomenclature of Medicine Clinical Terms (SNOMED-CT) by the International Health Terminology Standards Development Organization (IHTSDO),⁴⁷ Logical Observation Identifiers, Names, and Codes (LOINC),⁴⁸ International Classification of Diseases (ICD) by the World Health Organization (WHO),⁴⁹ and Medical Dictionary for Regulatory Activities (MedDRA) by the International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH).⁵⁰ ICD has more than one active version; the latest is version 11 (ICD-11). In late 2022, a new initiative was started to minimize duplications between LOINC and SNOMED-CT and to enhance the ability of stakeholders to use the two terminologies together.⁵¹

EHR ontologies

An ontology is a formal specification of a conceptualization aimed at describing the information exchange requirements to enable independently developed systems to interoperate.⁸ In practice, an ontology represents the mapping across different CIMs to create a common, shareable, and reusable view of the medical practice, inclusive of components, relationships, and rules.^12,36,52 Coordination between standardized data structures, messaging, and terminologies are needed for ontology-based solutions.^12,14,23 Ontologies signify a mature state of interoperability and a bridge toward composability.⁸

There is a growing adoption of ontologies and linked data combined with international standards to solve interoperability problems.¹⁴ But efforts are not coordinated. Many of our reviewed studies propose different rule-based ontologies mapping EHR systems to standards like HL7 FHIR, openEHR, and SNOMED-CT,^{28–30,35,36,38} including dynamic feedback and semantic/syntactic validation.²⁷

11,12,23,52 support and explore fuzzy ontology approaches. Fuzzy ontology can handle different standards, custom databases, and vague/imprecise data while preserving semantics of the integrated data.²³

Other interoperability approaches

Ontology mapping is quite subjective and relies heavily on concept definitions and explicit semantics.³⁴ It is hard to achieve. Other approaches include information retrieval, artificial intelligence (AI), and Natural Language Processing (NLP)—a type of AI.^31,34 If it is not an alternative, AI should be part of the solution for healthcare interoperability. 31,32,34,52 propose approaches combining NLP, ontologies, and open standards. They demonstrate promising results in identifying semantic differences, establishing logical similarities, and accommodating linguistic variables and vagueness in medical concepts.

53 demonstrate a novel application of bidirectional transformation (BX) techniques. BX focuses on one-to-one interactions between two systems rather than the one-to-many or many-to-many interactions that ontologies and open standards are trying to solve. BX is practical when there is a need to synchronize information across two or few separate medical record systems using a push model, such as in e-referral and e-consultation workflows.⁵³

Finally, 40 analyze the components of knowledge graphs (KG) and propose using knowledge graph structures as a virtual intermediate layer to facilitate semantic interoperability across different EHR standards. While an ontology is a conceptual model, a knowledge graph is a model of reality.

Problems with EHR interoperability

Problems with heterogeneity

Current EHR data is dispersed across heterogeneous systems and use different local technology and content standards.^{11,12,30,35,40,52,53} Achieving interoperability across EHR systems is problematic due to issues like lack of unified model, varied data sources, partial mappings, need for user intervention to identify meaning, existence of semantic differences, imprecise terminologies.^11,34 There are barriers and open questions around structuring legacy data, normalizing semi- and unstructured data, ambiguity in medical terms, lack of consensus on standards, and identifying and updating concepts over time.^14,32

Operational processes across healthcare institutions differ: patient care process, clinical process (day-to-day work), and administrative process (management and reporting) interoperability.³³ These three process types have their own interoperability needs and they can be at odds with each other.³³

Problems with standards

Standards can help with interoperability but have limitations. The reliability and precision resulting from using these standards are questionable.¹¹ Using proprietary or open standards leads to issues like naming disputes and access control dependencies that need resolution.²⁷ Despite standardization efforts, implemented interoperability interfaces and tools often fail to function correctly.⁵³ For example,13 analyze the government required CDA standard and find numerous real-world issues with the semantic correctness and consistency across vendors. Standards play a necessary function in presenting the required step toward reliability and quality.²³ However, additional factors, such as implementation practices, governance, and operational processes, are also crucial for achieving comprehensive interoperability.

Problems with ontologies

Consistency and coherency across multiple ontologies are necessary,⁵⁴ but that is not an easy task. They have contradictions and unsatisfiable results.⁵⁴ Ontologies are complex; a truly simple definition might not capture the full scope of what ontologies represent in EHR systems.³⁵

Learnings

Table 1 summarizes the review on EHR interoperability levels, expectations, approaches, and problems. Our review suggests that a simplified four levels may be more practical for addressing current EHR interoperability challenges. This four-level model aligns with the LCIM but focuses on the levels most relevant to EHR systems: technical/foundational, syntactic/structural, semantic, and process/organizational.

We deduce that better focus area for every level is needed to avoid conflicting definitions and expectations. These focus areas should be around the health records themselves rather than the holistic system to system integration. They are in column 5 of Table 1.

At level 1, the focus is shareability of the health record including exporting, packaging, and forwarding, regardless of format, structure, and source. At level 2, the focus is integrability—the ability of the receiver system to import, store, and cross-reference the record. This should be independent on whether the sender and receiver share structures, or whether data transfer is multi-directional, bidirectional, or even unidirectional. The receiving system can be another EHR system or an AI application. At level 3, the focus is interpretability and usability of the record. Interpretability can occur via rule-based approach, syndication of terminologies, or natural language processing from a trained AI model. Meaning of the record is assumed in this level, and usability can be different by different receiver systems or entities. The focus at level 4 is maintainability of the health record—the need for cross-organizational alignment should that record have a long-life span and role in future analysis.

Discussion

To better understand the levels, expectations, approaches, and problems of interoperability between EHR systems, we carried out a scoping review of studies over the recent 10 years (2014-2023), and we drew learnings from the LCIM. Semantic interoperability promises meaningful exchange of information if there are common standards and terminologies adopted by the exchanging systems. Large efforts have been put forward by many independent organizations and researchers to propose those standards and approaches. To date, none have gained consensus. We set to explore four research questions: (1) current EHR interoperability levels, expectations, and approaches, (2) problems and gaps holding progressing of EHR interoperability, (3) learnings from the LCIM, and (4) recommendations for a practical approach. The following four sections answer and discuss each.

Levels, expectations, and approaches

Literature on EHR interoperability supports the linear advancement through levels, in alignment with the LCIM but without explicit connection to it. The advancement is along four levels instead of seven: technical/foundational, syntactic/structural, semantic, and process/organizational. Interoperability of EHRs is a needed advancement for better patient care and reduced medical costs. Useful interoperability is when the meaning of shared data is preserved—commonly referred to as semantic interoperability.

Standard-based and rule-based interoperability—requiring common data representation, document structures, terminologies, and ontologies—have been the most common approach. Different types of data within EHR systems may require varied approaches to achieve effective interoperability. Open standards (HL7 FHIR, openEHR), ontology-based approaches, standardized terminologies are best for structured data, like lab results and vital signs. NLP, machine learning, are fuzzy ontology approaches are necessary for unstructured data with free-text notes. Semi-structured data may need a combination of both approaches. Continuous monitoring and image analysis may need standards for sharing but AI and NLP for pattern recognition. This demonstrates that achieving comprehensive interoperability often requires combining multiple techniques.

Problems and gaps

There is a lot expected of semantic interoperability in healthcare and some definitions can be contradictory: machine-readable and human-readable, clinical interpretation and linguistic interpretation, structured data and unstructured data, intermediary standards or local standards. Executing on it can be difficult.

The field of health IT continues to evolve year after year. Since 2014, there has been growing acceptance around certain standards, particularly HL7 FHIR.⁵⁵ The United States Core Data for Interoperability (USCDI) has also emerged as a key standard for promoting interoperability. However, with the inconsistency in adopting standards and the difficulty in mapping the growing number of systems, implementation has been hard.

Ontologies are complex and hard to scale. Even if organizations are using the same EHR system vendor or open standards, their setups may have undergone a number of customizations affecting how data is represented. Recent explorations are discovering value in AI, NLP, and bidirectional transformation. These approaches may bypass the need for strict syntactic alignment but will continue to need terminological coherence. Many EHR systems have legacy and unstructured data, like physician notes and observations. NLP can extract the value of this data without adherence to common structure. By default, the output of any system, interoperable or not, should be human-readable. Linguistic and cultural interoperability is very important and should be further studied.

Learnings from the LCIM

Our proposed four levels for EHR interoperability build upon the LCIM foundation, tailoring it to the specific needs and challenges of EHR systems. We suggest a path that focuses on advancing a health or medical record through shareability, integrability, interpretability/usability, and maintainability (column 5 of Table 1). It maps to four levels of EHR interoperability expectations, but it prioritizes the purposeful portability of the electronic health records themselves rather than the comprehensive connectivity of the systems.

Our suggestion of the four-level model was solely based on the study of the selected publications. We were encouraged after to learn that the Healthcare Information and Management Systems Society (HIMSS) supports similar four levels of interoperability.⁵⁶

Recommendations

If meaningful inference and use of the health record are the priority, then multi-directional exchange and unified data structures may not be necessary conditions for all use cases in healthcare.

It’s important to note that interoperability standards and approaches have evolved significantly over time. For instance, the CDA standard is gradually being superseded by more flexible and comprehensive standards like FHIR, which offers improved flexibility, modern web technologies, and better support for modular implementation. As healthcare needs and technologies continue to evolve, we can expect further refinements and potentially new approaches to emerge.

Regulations are necessary, like the efforts of the Office of the National Coordinator for Health Information Technology (ONC) to secure communications across different health systems and providers.²⁴ Government policies, reimbursement strategies, and workforce training can accelerate the integration of interoperable EHR systems.²⁴ We agree with Kuziemsky and Peyton³³ to recommend taking a longitudinal, socio-technical approach to understand the overall ecosystem where technology, processes, and people interact over time when studying interoperability.

Limitations

We understand that our review has limitations. Similarly to other scoping reviews, it may not have delved in deep analysis of all approaches, detailed the capabilities of each standard, or uncovered all technical implementations and limitations. Given the fast and vast advancements in the field, some other relevant publications may exist. However, we’ve taken steps to minimize these limitations through our sound search, selection, information extraction, quality assessment, and inclusion criteria. The result provides a necessary summary of key definitions, working concepts, common problems, major players, alignments, and gaps. Our synthesis, analysis, and reporting were concluded in the second half of 2024. Since then, we did a quick search which uncovered that the topics/titles of the newer publications address more specific systems, platforms, and technologies including NLP and AI. It sets up for a good subsequent study.

Conclusions

Extracting meaning out of exchanged data is not limited to shared standards and data structures. Focusing on case-specific solutions that prioritize the state of health records, rather full system to system integration, can expedite interoperability. Meaningful use can be achieved even in one-way data flow rather than two-way equitable exchange. Artificial intelligence and natural language processing are showing positive results in the process and they need further exploration and verification. As EHR systems evolve, interoperability approaches must adapt to handle diverse and emerging data types effectively. Stronger regulations, laws, and public policies may be necessary to guide ongoing integrations and ensure effective interoperability.

Supplemental Material

Supplemental Material - A scoping review of electronic health records interoperability levels, expectations, approaches, and problems

Supplemental Material for A scoping review of electronic health records interoperability levels, expectations, approaches, and problems by Raghid El-Yafouri, Leslie Klieb in Health Informatics Journal.

Footnotes

ORCID iDs

Raghid El-Yafouri

Leslie Klieb

Author Contributions

RE searched and curated records, defined the research questions, screened records, and drafted manuscript. LK screened records, structured the manuscript, and contributed to the discussion and conclusions.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Supplemental Material

Supplemental material for this article is available online.

Appendix

References

Dias

Martins

Queirós

, et al.

Interoperability in pervasive health: is it tackled as a priority?

Proceedings of the 11th international joint conference on biomedical engineering systems and technologies. SCITEPRESS - Science and Technology Publications, 2018, pp. 57–65.

Walker

Pan

Johnston

, et al. The value of health care information exchange and interoperability. Health Aff (Millwood). 2005; Suppl Web Exclusives: 10–18.

Frisse

Holmes

. Estimated financial savings associated with health information exchange and ambulatory care referral. J Biomed Inf 2007; 40(6 Suppl): S27–S32.

Reis

ZSN

Maia

Marcolino

, et al.

Is there evidence of cost benefits of electronic medical records, standards, or interoperability in hospital information systems?

Overview of Systematic Reviews. JMIR Medical Informatics 2017; 5(3): e7400.

Clarke

Ashrafian

, et al. The impact of electronic health record interoperability on safety and quality of care in high-income countries: systematic review. J Med Internet Res 2022; 24(9): e38144.

Torab-Miandoab

Samad-Soltani

Jodati

, et al. Interoperability of heterogeneous health information systems: a systematic literature review. BMC Med Inf Decis Making 2023; 23(1): 18.

Tolk

. Architecture constraints for Interoperability and composability in a smart grid. In IEEE PES general meeting. IEEE Xplore, 2010, pp. 1–5.

Tolk

Diallo

Turnitsa

. Applying the levels of conceptual interoperability model in support of integratability, interoperability, and composability for system-of-systems engineering. Journal of Systemics, Cybernetics, and Informatics 2007; 5(5): 66.

Robkin

Weininger

Preciado

, et al. Levels of conceptual interoperability model for healthcare framework for safe medical device interoperability. In 2015 IEEE Symposium on Product Compliance Engineering (ISPCE), 2015, pp. 1–8. https://ieeexplore.ieee.org/abstract/document/7138703

10.

Tolk

Bair

Diallo

. Supporting network enabled capability by extending the levels of conceptual interoperability model to an interoperability maturity model. Journal of Defense Modeling & Simulation 2013; 10(2): 145–160.

11.

Adel

El-Sappagh

Barakat

, et al. Distributed electronic health record based on semantic interoperability using fuzzy ontology: a survey. Int J Comput Appl 2018; 40(4): 223–241.

12.

Adel

El-Sappagh

Barakat

, et al. Ontology-based electronic health record semantic interoperability: a survey. U-healthcare monitoring systems. Academic Press, 2019, pp. 315–352. https://linkinghub.elsevier.com/retrieve/pii/B978012815370300013X

13.

D’Amore

Mandel

Kreda

, et al. Are meaningful use stage 2 certified EHRs ready for interoperability? Findings from the SMART C-CDA collaborative. J Am Med Inf Assoc 2014; 21(6): 1060–1068.

14.

de Mello

Rigo

da Costa

, et al. Semantic interoperability in health records standards: a systematic literature review. Health Technol 2022; 12(2): 255–272.

15.

Arksey

O’Malley

. Scoping studies: towards a methodological framework. Int J Soc Res Methodol 2005; 8(1): 19–32.

16.

Levac

Colquhoun

O’Brien

. Scoping studies: advancing the methodology. Implement Sci 2010; 5(1): 69.

17.

Tricco

Lillie

Zarin

, et al. PRISMA extension for scoping reviews (PRISMA-ScR): checklist and explanation. Ann Intern Med 2018; 169(7): 467–473.

18.

Page

McKenzie

Bossuyt

, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 2021; 372: n71.

19.

Munn

Peters

MDJ

Stern

, et al. Systematic review or scoping review? Guidance for authors when choosing between a systematic or scoping review approach. BMC Med Res Methodol 2018; 18(1): 143.

20.

Harzing

Alakangas

. Google scholar, Scopus and the web of science: a longitudinal and cross-disciplinary comparison. Scientometrics 2016; 106(2): 787–804.

21.

Bramer

Giustini

Kleijnen

, et al. Searching Embase and MEDLINE by using only major descriptors or title and abstract fields: a prospective exploratory study. Syst Rev 2018; 7(1): 200.

22.

Azarm-Daigle

Kuziemsky

Peyton

. A review of cross organizational healthcare data sharing. Procedia Comput Sci 2015; 63: 425–432.

23.

Adel

El-Sappagh

Barakat

, et al. A unified fuzzy ontology for distributed electronic health record semantic interoperability. In: Dey

Ashour

Fong

, et al. (eds). U-Healthcare Monitoring Systems. Academic Press, 2019, pp. 353–395. https://www.sciencedirect.com/science/article/pii/B9780128153703000141

24.

Zhang

Saltman

. Impact of electronic health record interoperability on telehealth service outcomes. JMIR Med Inform 2022; 10(1): e31837.

25.

Heart

Ben-Assuli

Shabtai

. A review of PHR, EMR and EHR integration: a more personalized healthcare and public health policy. Health Policy and Technology 2017; 6(1): 20–25.

26.

Sonkamble

Phansalkar

Potdar

, et al. Survey of interoperability in electronic health records management and proposed blockchain based framework: MyBlockEHR. IEEE Access 2021; 9: 158367–158401.

27.

Sreenivasan

Chacko

. Interoperability issues in EHR systems: research directions. In: Lee

Roy

Samui

, et al. (eds). Data Analytics in Biomedical Engineering and Healthcare. Academic Press, 2021, pp. 13–28. https://www.sciencedirect.com/science/article/pii/B9780128193143000021

28.

Ellouze

Bouaziz

Ghorbel

. Integrating semantic dimension into openEHR archetypes for the management of cerebral palsy electronic medical records. J Biomed Inf 2016; 63: 307–324.

29.

Ebietomere

Nse

Ekuobase

, et al. Crafting electronic medical record ontology for interoperability. The African Journal of Information Systems 2021; 13(3). https://digitalcommons.kennesaw.edu/ajis/vol13/iss3/2

30.

Sonsilphong

Arch‐int

, et al. A semantic interoperability approach to health-care data: resolving data-level conflicts. Expert Syst 2016; 33(6): 531–547.

31.

Roehrs

da Costa

Righi

, et al. Toward a model for personal health record interoperability. IEEE J Biomed Health Inform 2019; 23(2): 867–873.

32.

Dridi

Sassi

Chbeir

, et al. A flexible semantic integration framework for fully-integrated EHR based on FHIR standard. In Proceedings of the 12th International Conference on Agents and Artificial Intelligence. SCITEPRESS - Science and Technology Publications, 2020, pp. 684–691.

33.

Kuziemsky

Peyton

. A framework for understanding process interoperability and health information technology. Health Policy and Technology 2016; 5(2): 196–203.

34.

Bhartiya

Mehrotra

Girdhar

. Issues in achieving complete interoperability while sharing electronic health records. Procedia Comput Sci 2016; 78: 192–198.

35.

Martinez-Costa

Cornet

Karlsson

, et al. Semantic enrichment of clinical models towards semantic interoperability. The heart failure summary use case. J Am Med Inf Assoc 2015; 22(3): 565–576.

36.

Legaz-García

Martínez-Costa

Menárguez-Tortosa

, et al. A semantic web based framework for the interoperability and exploitation of clinical models and EHR data. Knowl Base Syst 2016; 105: 175–189.

37.

Maia

De Muylder

Reis

ZSN

. Archetype development process: a case study of support interoperability among electronic health record in the state of minas gerais, Brazil. J Med Syst 2019; 43(3): 57.

38.

Plastiras

O’Sullivan

Weller

. An ontology-driven information model for interoperability of personal and electronic health records. In Proceedings of eTELEMED, The Sixth International Conference on eHealth, Telemedicine, and Social Medicine. Barcelona: eTELEMED, The Sixth International Conference on eHealth, Telemedicine, and Social Medicine, 2014, pp. 130–133. https://www.thinkmind.org/index.php?view=article&articleid=etelemed_2014_6_40_40145

39.

Moreno-Conde

Moner

Cruz

, et al. Clinical information modeling processes for semantic interoperability of electronic health records: systematic review and inductive analysis. J Am Med Inf Assoc 2015; 22(4): 925–934.

40.

Sachdeva

Bhalla

. Using knowledge graph structures for semantic interoperability in electronic health records data exchanges. Information 2022; 13(2): 52.

41.

HL7 standards product brief - CDA® release 2 | HL7 international. https://www.hl7.org/implement/standards/product_brief.cfm?product_id=7

42.

About DICOM- overview [Internet]. DICOM. https://www.dicomstandard.org/about

43.

Integrating the healthcare Enterprise (IHE). IHE International. https://www.ihe.net/

44.

Index - FHIR v5.0.0. https://www.hl7.org/fhir/index.html

45.

ISO 13606 standard - EHR interoperability. https://www.en13606.org/information.html

46.

Beale SH Thomas . openEHR-Home. https://openehr.org/about/what_is_openehr

47.

What is SNOMED CT. SNOMED International. https://www.snomed.org/what-is-snomed-ct

48.

What LOINC is. LOINC. https://loinc.org/get-started/what-loinc-is/

49.

International classification of diseases (ICD). https://www.who.int/standards/classifications/classification-of-diseases

50.

ICH Official web site: ICH. https://www.ich.org/page/meddra

51.

LOINC / SNOMED CT – Regenstrief Institute Inc. and SNOMED International . https://loincsnomed.org/

52.

Adel

El-Sappagh

Barakat

, et al. An extended semantic interoperability model for distributed electronic health record based on fuzzy ontology semantics. Electronics 2021; 10(14): 1733.

53.

Weber

. Applying bidirectional transformations to the design of interoperable EMR systems. J Healthc Inform Res 2020; 4(2): 138–150.

54.

Slater

Gkoutos

Hoehndorf

. Towards semantic interoperability: finding and repairing hidden contradictions in biomedical ontologies. BMC Med Inf Decis Making 2020; 20(10): 311.

55.

Ayaz

Pasha

Alzahrani

, et al. The fast health interoperability resources (FHIR) standard: systematic literature review of implementations, applications, challenges and opportunities. JMIR Med Inform 2021; 9(7): e21929.

56.

Interoperability in healthcare | HIMSS. 2020. https://legacy.himss.org/resources/interoperability-healthcare

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.23 MB

A scoping review of electronic health records interoperability levels,expectations,approaches,and problems

Abstract

Keywords

Introduction

Methods

Search

Study selection

Information extraction

Quality assessment

Information synthesis

Findings

Levels of EHR interoperability

Basic levels of interoperability

Semantic interoperability

Higher levels of interoperability

Expectations from EHR interoperability

The role of EHR interoperability standards

Information models and transfer standards

Terminology and vocabulary standards

EHR ontologies

Other interoperability approaches

Problems with EHR interoperability

Problems with heterogeneity

Problems with standards

Problems with ontologies

Learnings

Discussion

Levels, expectations, and approaches

Problems and gaps

Learnings from the LCIM

Recommendations

Limitations

Conclusions

Supplemental Material

Supplemental Material - A scoping review of electronic health records interoperability levels, expectations, approaches, and problems

Footnotes

ORCID iDs

Author Contributions

Funding

Declaration of conflicting interests

Supplemental Material

Appendix

References

Supplementary Material