The healthcare sector is currently experiencing heightened expectations regarding interoperability. Federal agencies are ramping up enforcement against information blocking while promoting an Interoperability Framework and expanding the United States Core Data for Interoperability (USCDI). These initiatives signal a demand for greater accountability from providers and technology developers. Meanwhile, industry leaders are exploring emerging concepts like “conversational interoperability,” which involves clinicians using natural language to query electronic health records (EHRs) for immediate information retrieval. This approach reflects optimism about how new technologies, particularly artificial intelligence (AI) and large language models (LLMs), could simplify clinician interactions with complex systems.
Yet, historical context suggests that enthusiasm for technological advancements often outstrips practical realities. Various waves of interoperability initiatives, from early vocabulary standards to Fast Healthcare Interoperability Resources (FHIR), have struggled with a persistent barrier: the lack of clean, structured, and clinically valid data to support them.
Understanding Conversational Interoperability
Conversational interoperability is gaining traction and may see significant developments within the next 9 to 12 months as AI-driven interfaces demonstrate their potential. This concept is appealing as it aims to reduce the friction clinicians encounter when navigating EHRs. However, it is essential to recognize that AI can only retrieve information that exists within the records. If the underlying data is incomplete, unstructured, or inaccurate, the results of a natural-language query will also be flawed. In essence, poor data leads to poor outcomes.
Moreover, LLMs come with their own limitations. They can produce confident yet incorrect responses, a phenomenon known as “hallucination,” and require substantial computational resources. Without structured inputs, these tools have the potential to magnify existing gaps and errors rather than address them. Despite compelling vendor demonstrations, practical applications often expose the fragility of systems built on shaky data foundations.
The Challenge of Unstructured Data
The reality is that a significant portion of healthcare data remains unstructured. Vital information regarding symptoms, treatments, and patient context frequently exists in free-text notes or is housed in disparate systems, making it inaccessible for structured queries. When such critical information cannot be reliably extracted, clinicians are left with incomplete views of their patients, which undermines both the quality and safety of care.
Standards like FHIR provide mechanisms for packaging and transmitting data, but they do not guarantee that the data is clinically meaningful. In practice, FHIR often serves as a container for inconsistent or incomplete information rather than ensuring usability across different systems.
Structured and clinically valid data are crucial for several reasons. Without a reliable data foundation, every interoperability initiative—whether conversational, semantic, or technical—remains incomplete.
The Case for a Universal Medical Coder
One potential solution to this ongoing challenge is the development and widespread adoption of a universal medical coder. This system would be capable of translating clinical concepts into structured, standardized, and contextually accurate representations at the point of care. Such a tool would map free-text inputs and unstructured documentation into consistent, clinically valid codes across various vocabularies, including the International Classification of Diseases (ICD), Systematized Nomenclature of Medicine Clinical Terms (SNOMED CT), and Logical Observation Identifiers Names and Codes (LOINC).
The benefits of a universal medical coder extend beyond regulatory compliance and billing efficiency. Its true value lies in its ability to establish a solid clinical data foundation. By capturing concepts in real-time within a clinician’s workflow, it ensures that data remains accurate, complete, and interoperable across systems. This alignment would enable interoperability frameworks like FHIR to fulfill their promise, as the data contained would be as usable as the framework itself.
Healthcare leaders should approach technology advancements with a critical eye. While conversational interoperability is a compelling concept, it should be viewed as one element of a larger framework. The root issue remains the need for substantial investment in data integrity and fidelity. Only then can advanced applications like conversational interfaces, predictive AI, or population health analytics achieve sustainable impact.
This balanced approach is essential. While the industry thrives on innovation and enthusiasm, it must also maintain realistic expectations. Impressive demonstrations should not divert attention from the essential work required to build structured, clinically valid datasets. Policymakers, vendors, and providers must collectively recognize that true interoperability is not achieved through user interfaces or standards alone. It requires that every patient encounter yields usable, exchangeable, and meaningful data.
In conclusion, the healthcare industry’s renewed focus on interoperability is both necessary and overdue. Regulatory enforcement against information blocking, the expansion of USCDI, and ongoing innovation are all critical steps. However, these efforts will not realize their full potential unless the industry prioritizes structured, clinically valid data as the foundational element. Concepts like conversational interoperability highlight both the opportunities and risks present in this moment. While such trends may enhance usability, they cannot compensate for poor data quality. A universal medical coder, applied consistently across care settings, offers a practical solution to the long-standing challenge of data integrity. Addressing this core requirement is vital for the healthcare sector to move beyond cycles of over-promised breakthroughs and to achieve truly interoperable, patient-centered care.
