Interoperability, or the seamless electronic exchange of information, has long been a goal of the healthcare industry.
And while interoperability has continuously improved in the EHR era, it still has a long way to go.
So where does healthcare interoperability go next? Becker’s posed that question to health system IT leaders across the country. Here are their responses:
Debi Brobst. CIO of ProMedica (Toledo, Ohio): Healthcare interoperability is driven by data, which drives patient care. AI and machine learning are becoming embedded in technology, automating routine tasks and predicting outcomes. The requirement of FHIR interfaces assists us in data standardization, allowing us to exchange data. Government regulations, including TEFCA, allow us to exchange data on a larger scale.
The standardization and availability of data at a national level will help improve the electronic exchange of healthcare data.
Chuck Christian. Chief Technology Officer of Franciscan Health (Mishawaka, Ind.): With the investment that healthcare has made in EMR systems over the last 15-plus years, the majority of all patient encounters are captured within these electronic systems. With healthcare being provided in multiple different settings (i.e., ambulatory/physician offices, urgent care, virtual visits, behavioral health, hospitals, postacute care, etc.), that information does not flow from one setting of care to another without a level of friction. The continued use of federal regulations to remove that friction has made an impact; however, there is still a good deal of work that needs to be accomplished. Health information exchanges were a focus of the early regulations that came with HITECH and “meaningful use” in 2009 and the years that followed.
In 1996, as part of the initial HIPAA legislation, one of the identifiers that was included was related to a national patient identifier, which was meant to create a simple method to provide positive patient identification; however, there was a concern that this would be used negatively and present a potential threat to a patient’s identity and was removed from any federal funding allocations. While there are multiple methods to “determine” a patient’s identity using demographic, multiple data paths, etc., these methods are approximately 85-88% accurate. If you look at the total number of patient care encounters each year, this will mean that there is a potential to misidentify patients 12-15% of the time. Establishing a more accurate method of patient identification would create the ability to ensure a safer method of data exchange/interoperability
Many of the physicians I’ve interacted with while implementing EMR systems have shared that they don’t need to know what they already know, but they need to know what they don’t know (about the patient); but better yet, they would like to be presented with the information that they need to know that is specific to why they are seeing the patient during any given encounter. This is a tall requirement; however, with the advent of the enhancements being made in health information exchange, higher-functioning EMR applications and the inclusion of AI, we should be in a position to start making the EMRs smarter so they can retrieve information that is germane to the current patient encounter. In order to have the computer systems automatically retrieve and present that previously unknown information, we really need to ensure that we are gathering information for the correct patient; if not, this enhancement could very well introduce a new level of concern for patient safety.
Josh Glandorf. CIO of UC San Diego Health: Healthcare interoperability is evolving from basic data exchange to deeper, more meaningful integration across systems and care settings. The next step is achieving semantic and organizational interoperability, where data is not only shared but understood and used effectively for clinical decision-making and care coordination.
What’s next:
— Longitudinal, patient-centered records that follow individuals across providers and settings.
— Widespread adoption of FHIR and open APIs, enabling smoother data access across platforms.
— Integration of nontraditional data, including social determinants and wearable data.
— AI-powered analytics leveraging standardized, interoperable datasets.
What will help:
— Stronger policy enforcement (e.g. TEFCA) and aligned privacy rules.
— Standard adoption (FHIR, SNOMED, LOINC) to enable consistent data use.
— Trusted data-sharing frameworks and robust governance.
— Patient empowerment through access to and control over their own health data.
In short, interoperability is moving toward enabling true data utility — and success will depend on aligned standards, shared trust and active participation from both providers and patients.
Jeremy Meller. CIO of Children’s Healthcare of Atlanta: Common standards like the Trusted Exchange Framework and Common Agreement (TEFCA) and similar frameworks are a strong foundation for healthcare interoperability. However, we need more transactional and descriptive health data standards to better enable exchange of information at a more detailed level between disparate electronic health records. Enhanced standards will provide gains to patient and provider experience, supporting improved care coordination and patient engagement.
Priti Patel, MD. Chief Medical Information Officer of John Muir Health (Walnut Creek, Calif.): Despite tremendous progress over the past decade driven by federal legislation, technological advances and the adoption of standards like FHIR and USCDI, healthcare interoperability remains far from solved. The next chapter must focus on bridging persistent gaps, reducing cost burdens, ensuring patient privacy, and embracing AI to streamline and scale data exchange.
Healthcare interoperability is no longer about whether data can be exchanged — it is about whether it flows effectively, securely and meaningfully. As we look ahead, AI will be a key enabler, public health technical infrastructure needs to be modernized, and patients must be empowered to control their data. We must address both the technical and human factors, so clinicians are not buried in data, patients are protected, and the system works seamlessly everywhere, for everyone.
Keith Perry. Senior Vice President and CIO of Carilion Clinic (Roanoke, Va.): We need to focus more on resolving issues with disparate and proprietary systems. Although there has been improvement, navigating the hurdles of multiple systems and the lack of standardization remains challenging. Increasing the adoption of FHIR APIs across the industry is crucial to standardizing and simplifying data movement and making it easier to work with compiled data. Leveraging AI can accelerate this standardization, including unstructured data, while simultaneously streamlining clinical workflows to enhance data sharing.
Trust and data security barriers continue to pose significant challenges. Determining who is responsible for data protection and preventing breaches is a persistent issue affecting interoperability. Additionally, we must address trust issues between health systems and alleviate fears about the use of data beyond patient care. Regulatory complexities further complicate matters. While TEFCA appears to be a step in the right direction, its widespread adoption will most likely be slow.
Michael Schnabel. Vice President and CIO of UT Health San Antonio: The next phase of healthcare interoperability is rooted in intelligent, context-aware data exchange that enables real-time decision-making, precision workflows and scalable research. We’re advancing this vision by integrating Epic with cloud-native and AI-augmented platforms — leveraging technologies that automate audit processes, surface billing risks and enhance operational oversight. Interoperability now extends beyond simple connectivity; it’s about orchestrating data across systems using AI to interpret, prioritize and deliver insights where they matter most. By aligning clinical, administrative and research data through secure, AI-enabled infrastructure, we’re reducing friction, enhancing care quality and building a foundation for the next generation of data-driven healthcare delivery.
Albert Villarin, MD. Vice President and Chief Medical Informatics Officer of Nuvance Health (Danbury, Conn.): The future of interoperability is not just about moving data — it’s about making that data actionable, contextual and aligned with care delivery goals. As clinician informaticists, we sit at the intersection of data science, clinical reasoning and systems strategy. Our role will be essential in steering the next generation of interoperable, intelligent healthcare systems.
Jim Whitfill, MD. Senior Vice President and Chief Transformation Officer of HonorHealth (Scottsdale, Ariz.): For the last several years, the U.S. has been focused on solving interoperability challenges by implementing the 21st Century Cures Act along with TEFCA as the frameworks for providing exchange of data across healthcare organizations. It will be interesting to see how the current administration feels about these efforts since it was federal enforcement that was driving much of the adoption. However, one of the major challenges of interoperability is the lack of standardization around much of our EMR contents, especially for data that is qualitative or narrative instead of discrete. For example, health systems can share documents called CCDs today, but many clinicians would note those CCDs don’t contain the narrative information they need to understand the thinking of the clinicians caring for that patient at the sending organization.
With the development of large language models, one could imagine the ability to summarize a chart with specific prompts to give the context required by the receiving clinician. While there is certainly risk attached to chart summarization, this could also offer a real solution to the current lack of interoperability on the topics most needed by clinicians.
Denise Zabawski. CIO of Nationwide Children’s Hospital (Columbus, Ohio):
1. Automation through AI: Interoperability works well at a macro level but still involves manual steps for patients to connect providers across systems and to health platforms and applications. A national master patient identifier could help automate those linkages, eliminating manual steps. Providers may think they have a patient’s entire record, but don’t know for sure that the patient created all the appropriate linkages. This is a big problem for elderly patients who are often less tech-savvy and may not have someone to help them with this process. Important information about medications and prior visits can be missed. AI-powered automation could create the linkages for a comprehensive medical record, while prompting the patient for approval. These linkages could easily extend to fitness devices and applications, creating a personal copy of health history and health trends. MyChart and other patient portals have come a long way, but there are still opportunities to make them more user-friendly. It would be great if patients could get a text that alerts them to anything abnormal with a link to look at the details. Today, patients must log in to MyChart, pull up lab results and scroll through all the values, not really understanding if the results are serious or if there are next steps.
2. Connecting hospital systems: If everything you do and everyone you see is in the same hospital system, you have access to everything you need and your care process should flow smoothly. This process flow breaks as soon as you move outside the system, whether it’s an ER visit when you are on vacation or seeing a specialist at a different hospital. Patient orders and results don’t move across systems, which makes it difficult for providers to view a holistic picture of a patient’s current health status without extra clicks and time to find the data in other systems. For example, Nationwide Children’s operates nurseries in adult hospitals. Lab orders are entered in the adult hospital system but also must be manually entered into the pediatric hospital system. It is extremely difficult to create point-to-point order-and-result interfaces between hospital systems, and impossible if the hospitals are competitors. Building custom point-to-point interfaces is time-intensive, expensive and error-prone. Ideally, a provider should be able to enter an order one time in one system and have it automatically flow to another. Then the result can follow and flow to the adult medical record and the pediatric medical record together. This ensures the results are immediately and completely visible to everyone on the patient’s care team. Having the ability to do this between primary care providers, outpatient surgical centers and hospitals would reduce a tremendous amount of friction, manual labor and errors.
3. Health devices and applications: Certain health devices and applications can help create a comprehensive picture of health based on the device and manually entered data. There is a lot of opportunity to make this information easier to consume. Consumers may purchase wearable health trackers and be impressed with the capabilities and AI algorithms that alert them to their readiness and overall health, but users must log in to an app to see the information. As consumers, we need more standardization in the interfaces between health apps and devices (like HL7 or FHIR). That standardization can revolutionize how we take care of ourselves by linking glucose monitors, fitness devices and other vitals monitors into a user-friendly summary. In being able to see an overall summary with everything connected together, technology can make it much easier for people to do the right thing for their health, in collaboration with their care team.
