What Does Interoperability Really Mean?


William A. Hyman
Professor Emeritus, Biomedical Engineering
Texas A&M University, w-hyman@tamu.edu
Read other articles by this author

Interoperability in the context of EHRs is much discussed, but remains somewhat loosely defined. According to the ONC definition, adopted from IEEE, interoperability is “the ability of systems to exchange and use electronic health information from other systems without special effort on the part of the user”. The word “interoperability” appears 365 in ONC’s draft Interoperability Roadmap including an addition to the definition which says “Interoperability is made possible by the implementation of standards”. The latter perhaps brushes over that standards may be a necessary condition but are not a sufficient condition.

Despite all this there is of course some vagueness in what interoperability actually means. One issue has to do with what electronic health information. Is it the entire medical record, or selected portions of a medical record? And what constitutes “special effort”? Ideally this could mean no effort, ie the information just shows up in the receiver’s system, and they know it came, can readily find it, and it is fully integrated with the information that was already there. Or special effort might mean some level of effort is required (open this, click that, etc) but that this level isn’t special, presumably meaning not arduous, which immediately raises the question of arduous to who. It is also curious that only “from” is mentioned, ie the definition assumes the perspective is that of the recipient, not the person or system sending the information. Thus not addressed is how much effort, special or otherwise, the sender had to undertake in order to access, select, format and send the data, and perhaps confirm receipt. In addition, except perhaps in the case of lab, imaging or other isolated services, the original sender might want something back after the recipient treats the patient, ie a return of the new provider’s results and findings. Does this happen automatically or does the new provider have to undertake some effort–but not special–to make this occur? Or is the original provider so good at its job that it makes sure it gets information back. There is also then a frequency issue with respect to how often such information is exchanged, especially if there is ongoing treatment and/or multiple tests. Emergency room visits raise a different data flow question since here the ER event is not a referral and they would ideally be sending documentation of its encounter to someone, or several people, or many people.

The best example of interoperability is when there is in fact a single medical record which all providers interact with. However this isn’t really “inter” since inter implies between things and a single integrated EHR is one thing, not several different things that interact with eachother. Large healthcare systems have these single systems as I was recently reminded when a clerk at such a system asked me if there was any change in the non-system doctors I was seeing. They didn’t need to ask me about other system doctors because they knew that information. If I chose I could get all of my medical care from this system and thereby assure that all of my providers have that single system access. But for various reasons I have not done that including convenience and that one of my system doctors left to go elsewhere and I then had to choose to stay with the doctor or stay with the system. I chose the doctor. Multiple systems also results in multiple patient portals each with its own microenvironment of data.

It should be noted that here single system means the very same system since two providers with the identical EHRs will each have separate patient data even though they are using the same EHR product. In the absence of truly integrated (single) systems, when (if ever) all EHRs can send and receive patient data in a consistent and useable format, the “use case” questions will still exist. The core questions to support almost any use case (including the 56 “priority use cases” in the Roadmap) are what data, sent when, sent where, on whose initiative, reviewed by who, and with how much effort? And this still leaves the cybersecurity issue since it has been well demonstrated that the more information there is in one place, the more desirable it becomes to hack that place and steal that information.