How Should Interoperability Actually Work?
The generic idea of medical record interoperability is that diverse record systems from diverse providers and diverse vendors should be able to send/share/exchange/display patient information in order to enable better individual healthcare, and perhaps support a variety of public health interests. As ONC puts it, “EHR Interoperability enables better workflows and reduced ambiguity, and allows data transfer among EHR systems and health care stakeholders. Ultimately, an interoperable environment improves the delivery of health care by making the right data available at the right time to the right people.” However, there is a significant difference between an EHR being capable of processing information once it receives it and the logistics of triggering that flow of information. (Note that medical record interoperability is distinct from medical device interoperability which address medical devices communicating with a record or display system, or in some circumstances communicating with each other.)
The generic record interop idea has a relatively simple implementation when all of the providers are part of a single healthcare enterprise and that system has a single record system. When that is the case various specialties each generate their individual records. These are stored collectively and an enterprise provider can also look at the records generated by others. Even here however the records may be, to varying degrees, separate entries rather than an integrated whole. When this is the case the right data, at the right time may be available to the right people but only if they diligently look through a variety of disparate records. This can be further challenging in a retroactive review in which it will be easily shown that the data was there if the provider had “only bothered” to look for it. Here healthcare systems must be distinguished from EHR systems, given the now classic tale of two hospitals from the same enterprise, which happened to be across the street from each other, not having access to each other’s records even though patients might move back and forth between them. Information sharing deteriorates further if the patient ventures outside of the enterprise such that the central records are not readily accessible to their outside providers, and records generated by the outside provider may not find their way into the central record. In this regard, a perhaps unintended consequence of provider networks is that if they also mean a single record system then insurance cost savings interests and quality of care interests may actually coincide.
The more complex situation of multiple record systems owned and operated my multiple providers raises the question of how information exchange will be initiated and followed, not technically but logistically. What does interop mean in this situation? It should mean that each provider can access the other providers’ records, preferably without human interaction by the record owner. If this were not the case then record exchange could only occur during business hours. For many interactions this might be acceptable, but it would not serve in an after-hours emergency. Such open exchange of information also raises authorization and confidentiality questions. In addition, cybersecurity concerns generally increase with the number of people who can access a system. Moreover, in an emergency the provider might not know where relevant records reside. In particular, if an emergency patient was incommunicado there would be no way to ascertain what records existed, let alone obtain them, even if the record systems were compatible. After emergency treatment the record of that encounter would presumably be of interest to the patients’ more regular providers, but how does the ER record get back to those providers? Perhaps patients will have to carry medical identification which includes record availability information and access codes. Or there could be a central repository containing information on where various records exist, if not containing the records themselves.
There is a substantial difference between the compatibility of computer data bases and the interactive human governance of exchanging data between such databases. An analogy here might be that you and I might both have the same database software and if I could get the data you have in a machine-readable form I could successfully load into my computer. However, I might not know your electronic address, or phone number, or when you will be there. Moreover, I might not even know who you are.
Technical compatibility is a necessary but not sufficient condition for a functional interoperability process. This process (or processes) is an information control problem that I believe has not seen adequate discussion. The “use case” may be of help here in which particular scenarios are described in which a patient with records in one or more systems presents to a new provider and it is desirable for the new provider to access the existing records, and to report back on the current encounter. How does this really work?