This includes not just data from the EMR and other primary hospital IT systems, but legacy IT systems, mobile computing on tablets and smartphones that are now becoming the standard for patient rounds, bi-direction data sharing to the cloud, remote patient monitoring, telemedicine and now artificial intelligence (AI). 

The promise of integrating these systems is use of application programming interfaces (API) and the new Fast Health Interoperability Resources (FHIR) HL7 specification for healthcare interoperability.

Interoperability needed with artificial intelligence

The U.S. Food and Drug Administration (FDA) has cleared well over 500 clinical AI algorithms to assess patients and aide patient care. Of these, about half were cleared by the FDA in just the past two years. However, there are hundreds, if not thousands, of non-clinical AI algorithms now operating within health IT systems to help manage data. This includes data mining medical records to identify trends or patients for population health, analytics, hospital department business management, inventory tracking and many other IT areas that are not patient facing. 

But all of these algorithms need interoperability within the established workflows on electronic medical records (EMRs) and other hospital IT systems, Trivedi said. Interoperability is needed not just to enable pulling in data throughout a health system, but also to prevent glitches in the system or patching by IT, which consumes staff time. He also said AI algorithms might be applied to help improve interoperability between systems.

"Definitely AI is here, front and center, but at the same time, at a policy level, there are lots of demands for some sort of regulation and guardrails. It is safe to say the horse is out of the barn, and the industry is optimistic about how AI may revolutionize interoperability and healthcare standards and delivery," Trivedi explained.

HIMSS supports setting industry-wide standards for interoperability, and more work needs to be done with AI governance, testing and compliance. Testing and comparison of these algorithms is an area that needs a lot more focus, Trivedi explained. 

Another AI trend Trivedi mentioned were a couple EMR vendors announcing at HIMSS the integration of ChatGPT into their systems. He expects to see a lot more of these types of integration with AI that can offset staffing issues and improve workflow.

The need for equitable interoperability 

The COVID-19 pandemic exposed a lot of inequities in healthcare, not only between the access and level of care of different socio-economic groups, but also from a health IT standpoint. This exposed a breakdown between local, county, regional, state and federal level data exchanges and the inability to get health data as quickly as people wanted because of large variations in how data is collected, stored and shared.
 
"This was really exposed during COVID with the public health reporting of COVID information," Trivedi explained. "It exposed the need to modernize our public health information systems. We also talked a lot about the $35 billion of investments the industry has made in ambulatory, inpatient and meaningful use to get to where we are now, but that did not apply to everybody. This did not include public health systems, long-term care, post-acute care, behavioral health, emergency medical services, and many more of these components of the healthcare ecosystem were left out." 

The pandemic made it clear more investment is needed to bring these disparate pieces of the healthcare mosaic together for easier sharing of data. 

In addition, there were already efforts underway to collect more data on social determinants prior to the pandemic, but the inequities that were very evident during COVID care have boosted this as a bigger priority post-pandemic. Trivedi said this data is needed to better address the inequity cracks in the healthcare system. He also said this has led HIMSS to support the concept of "equitable interoperability."  

Better data access should lead to better health outcomes, so HIMSS wants to help get more complex data points wired into healthcare systems. This includes social determinant data traditionally outside of health IT, such as income levels, the type of neighborhood someone lives in based on address, access to grocery stories and pharmacies, if a person owns a car to be able to pick up meds. HIMSS also wants to help get more complex data points linked into patient health records, such as mental health, sexual health, reproductive health. Sometimes these types of more sensitive health data can not be easily redacted is often just not included in a patient's record. So sometimes key elements about what is going on in a patient's life are unknown.    

"What we see here is that the people who can be best served by interoperability wind up being cut out of the picture, and this just increases the disparities," Trivedi explained. 

Many of the key interoperability data collection points were developed years ago by hospital informatics experts, but with the need to bring in new types of data it is time for new blood and stakeholders to help in gathering social data.

"When we talk about social determinants of health, we are talking about capturing data in places [that are] completely different. We are not talking about clinics or hospitals anymore, so there is a huge component that involves education both for providers, community health workers and the technology workers so they understand what technologies are available and how they can use them to drive better interoperability," Trivedi explained. 

Broadband internet connections is a big area where more data is needed to help understand patients and their access to the rest of the world and to enable improved access to healthcare. He said this area also will see more focus in the coming years to address interoperability with the patients themselves and wider use of wearable monitor technologies.