Unicorns, Pink Elephants, and Interoperability

Unicorns, Pink Elephants, and Interoperability
Three things you aren’t going to see in your lifetime

Let’s be fair, it’s quite possible with the technological advancements we experience on an almost daily basis that we’ll see a genetically engineered unicorn, or a pigment modified elephant. As for interoperability, fat chance. Heresy, you say? “There’s such a need.” “There’s so many companies working on it, even touting it now.” “It’s federally required for gosh sakes.” Those are all true statements, and I don’t suggest we won’t make incremental progress, but the holy grail of interoperability will be cracked and pitted and will ultimately fail to deliver on the promise of a seamless, accurate, and reliable health record.

For most of us, we’re so caught up in the dream of interoperability and it’s illusory promise that we haven’t dug into the nitty gritty details to see what the end result will actually look like. We’re focused on creating this spider web flow of data from disparate system to disparate system in hopes of keeping everything synchronized and up to date, with relevant and accurate clinical data. The truth is, on our current trajectory, it’s not going to work, and here’s why.

Let’s take a look at a typical elderly patient, this one happens to be technically astute, as in my 90 year old dad. He is considered to be a heavy healthcare system user. By the way, this example is not unique to thoseUnicorn_interop considered to be heavy users, but rather any patient that experiences encounters annually with multiple physicians. On his care team he has a primary care physician, a cardiologist, a nephrologist, a gastroenterologist, and for good measure let’s throw in an acupuncturist that also serves as a nutritionist. At home the patient has a shoe box full of medications, some prescribed, some over the counter. Each provider, sans the acupuncturist, has a patient portal. None of the portals represent a complete record for the patient. In fact, they are typically out of date or inaccurate. While they may present an accurate record of my dad’s experience with that particular provider, they do not incorporate the external information. To add to the complexity, his recent hospital resulted in yet another portal, with information from his stay. Thanks to copious notes by my sister and I, we were able to convey critical care elements to his other physicians verbally and by attending subsequent visits with my dad. Did I mention each portal has it’s own access information also? Needless to say, my dad doesn’t use or rely upon any of them. Despite the surveys that state patient portals aren’t working because they aren’t being adequately marketed, the reality is the current state is flawed, and that’s the reason for the lack of patient acceptance. This is the state of patient portals today.

The Direct Dilemma

But the Direct Protocol and the push for interoperability will change all that, right? Wrong.  First, let’s talk about Direct. It’s a messaging protocol for the transfer of clinical records across systems. Will it be effective? Incrementally, yes. Much in the way a 10 speed bicycle is better than a single speed. But it’s still a bicycle. Again, let’s step back into the tangled web of an elderly patient’s care. In a perfect, Direct and interoperable world, you’ll have data flowing from the primary care provider to the cardiologist, from the cardiologist to the nephrologist, from the nephrologist to the hospital, and back again, all being synchronized with the existing record at each node (practice). Keep in mind that most systems are incorporating Direct technology like email, so in many instances this communication is being manually generated and sent. This should be bi-directional, meaning either side should be able to send information to the other. Once the C-CDA (acronym for Consolidated Clinical Document Architecture) has arrived, it needs to be imported into the database. So what happens when an
error occurs, much like what happens with an HL7 lab interface? Let’s also keep in mind that the current C-CDA record is a limited data set. Many providers want to see more from each other. Great, the C-CDA allows for custom mapping of additional elements. Well how will that flow across systems? Adding to the complexity, all of these communication channels will be managed by different vendors, in many cases the EHR vendors who’ve been mandated to create this interoperability.

Where does all this take us? The first case for the failure of our current interoperability trajectory is complexity. Anyone who’s been on the technical side of this industry for any length of time knows the reality of maintaining interfaces. Don’t let the Direct element fool you – these are interfaces. While Direct provides the communication conduit, the underlying data within the C-CDA needs to be mapped and maintained. That leads to the second point of failure – cost. With complexity comes cost. This woefully inefficient architecture will be costly to maintain, and who is going to foot that bill? Let’s say a message sent from the primary care provider to the hospital regarding a recently admitted patient fails to import correctly into the hospital Information System. In most cases, the primary care’s EMR company will need to work with the hospital’s IT team to correct the problem. This scenario will play itself out thousands of times a day all over the country. I don’t believe the industry can sustain this resource demand nor the financial drain it will cause.

The solution is to part from the fragmented records that exist today entirely. The patient record needs to become patient centered. A patient centered record, by my use of the term, means a single patient record. Only the initial care rendering provider will create a patient record from scratch. Subsequent physician encounters will merely be updates to that original record. This approach eliminates the need to keep multiple records synchronized across disparate systems. It eliminates the limitations of the C-CDA format that provides only limited patient data and often excludes the narrative elements that many providers consider to be an important part of getting up to speed on a referred patient.interop

The technology to create this approach to medical records exists now, and a handful of companies are offering this model. Market forces dictate the easiest, most efficient answer ultimately prevails. Despite current market force obstructions, like well intentioned meaningful use requirements that have served to grow existing data silos taller, ultimately this model will prevail and the long hoped for promise of EMR and interoperability will be realized.