A Closer Look at Interoperability in Healthcare

Interoperability is important. How do I know? Because everyone says it is. The question is: Is it the word interoperability that is important or some capability that the word is intended to represent? As a word, interoperability is important in the same way that apple pie, motherhood, and patriotism are important. These words engender warm, fuzzy feelings and no one would ever admit that they are not important. Since important is the opposite of not-important, and none of these words is allowed to be not-important, they must all be important.

Now that you know that interoperability is important, what is it exactly? I don't mean something general like "able to share information," I mean in detail, like those e-mail protocols I discussed a few weeks ago. What, precisely is going to be shared? Will information be requested, or sent unsolicited? In what format? How will the requests be acknowledged? What, exactly is the recipient of the information going to do with it? Who is responsible for doing the work necessary so that the recipient can use what has been sent? Are there different combinations of what might be sent and how it might be used that would, in essence constitute different classes or types of interoperability?

So, you see, just believing that interoperability is important doesn't get you very far. Most importantly, it doesn't help you determine whether, just because you are promised "interoperability" that you will get the type of interoperability you need. Insisting on interoperability that you will never exploit is a waste of everyone's time and money.

When the government requires certain types of interoperability regardless of whether anyone needs them or not, it is an unnecessary intrusion - another example of specifying the means by which something must be done, rather than concentrating on the desired end result, such as better patient care.

Still, the notion that things and information created for some purpose might be usable by others, under other circumstances, is a worthy goal and one that is relevant to a broad range of activities. A recent article, "The Elusive Universal Web Bytecode" by Alon Zaka, a researcher at Mozilla.org, addresses interoperability and provides a useful set of requirements that interoperability should fulfill. I will translate them from geek-speak to something more general.

For information or data to be "interoperable" it should:

• Support all the languages, formats, information, and data types.

• Demand only modest resources to generate, receive, and process.

• Require only readily and inexpensively (free) available tools (software).

• Have a compact format for transfer.

• Conform to some identifiable standard.

• Be platform-independent.

• Be secure.

In the context of EHR, interoperability refers to the ability to occasionally exchange information, on an ad hoc basis, about individual patients with some ER or clinic down the street or halfway across the country.

Only those things should be exchanged for which there is an immediate, clear-cut need. Sending gobs of data, things like quality measure reporting, does not fall into the interoperability category. Recipients should only request the minimum information necessary to satisfy their immediate need and should retain as little as possible to avoid placing privacy at risk. Of course, you want connections to your billing system and your labs, but those involve interfaces (a different kind of interoperability.)

Achieving interoperability can either be simple and easy or extremely difficult. Using the above criteria makes it simple; even a fax will suffice. The government seems to have, instead, chosen the extremely difficult route that cannot begin until detailed specifications have been standardized for every bit of information that one might conceivably want to exchange, while at the same time neglecting to define anything clearly.

If you are a non-adopter or a reluctant adopter of EHR and you look deep in your heart of hearts, I think you find that whatever may be the basis of your reluctance, it has nothing to do with the kind of interoperability that the feds claim is so vital. The typical physician could satisfy most of their needs for information about a new patient if they could just have access to some legible, well organized, and intelligible (no abbreviations) notes that describe the patient's current condition, working diagnosis and past problems, the treatment to date, and any other tidbits (such as allergies or genetic traits) of significance.

If interoperability is so important, the simple approach is best. There are no impediments to implementing it today (contrary to what many have said). Opting for a complex approach instead is tantamount to saying that interoperability is not that important after all, because complex approaches are notable for consuming time and resources and for producing few useful results.

Faxes and structured documents meet the need and there are theoretical reasons to consider them optimal. So, when it comes to interoperability the KISS approach is best - Keep it Simple (to avoid being) Stupid.