Web Log


Why Fast Healthcare Interoperability Resources (FHIR) abd HL7 Argonaut Will Fail (and What To Do Instead)

8/29/2016 9:12 AM

Last Friday I'm doing the usual research about what is going on in the world and come across "Argonauts Navigate Between Heavy-weight and Light-weight Standards," posted by Andy Oram (Actually it was Thursday, but the second part was published on Friday).  Part of the article states

Congress’s recent MACRA bill, follow-up HHS regulations, and pronouncements from government leaders make it clear that hospitals and their suppliers won’t be off the hook till they solve this problem of data exchange, which was licked decades ago by most other industries. It was by dire necessity, therefore, that an impressive array of well-known EHR vendors announced the maverick Argonaut project in December 2014.

Let's be clear: health information exchange (HIE)is not like that any other kind of industry.  This is a departure from what we normally espouse.  Normally we would say that 'your problem is not unique and we have solved the various pieces of your problem for someone else previously.'  This time, I'm going to side with the doctors.  If you are a bank, all you care about is what the balance is.  Literally the only question becomes "can the current resources cover the intended action?" If you build cars the questions are a little more complex: "do I have enough widgets and labor to complete the car?" in addition to the financial part "does the customer have enough resources to cover the intended purchase?"

In healthcare, we don't have the luxury of  having a commodity level service, like the financial industry does (yes, you all do it the same) or a commodity level manufacturing service (yes you all do it the same) where all questions have a concrete response.  Either the check will bounce or it won't.  Either you have enough camshafts or you don't.  In healthcare the product is both more esoteric, that is, How do you feel?" and  more subjective, or, "What are the symptoms?" that said, we need a complete history of what has been done to any particular patient making this kind of information exchange unique in the world.  Sure your bank can show you all the activity in a particular account (My idiot bank can show me nine months anyway, so I only get back to June or July of the previous year at tax time making it wholly ineffective) but we don't have a medical 'bank.'  I don't want those financial idiots having control of my medical information anyway.

A new protocol is simply going to further muddy the waters, cause confusion and be done worse than about anything that has come before.  Swinging back to the side of the tech guys, we have SOAP, REST, Web API, WCF and any of a hundred XML based web services.  We don't need a new protocol, we have more than we know how to use now.  Further, imagine the what the backlash would be if every other industry wanted to come up with it own communications protocol.  The Arrogance!  Even worse, the last time the medical industry got it's hands on a 'protocol' they came up with HL7, which is an unmitigated train wreck with documentation three inches thick and a standard so complex that nobody can follow it. Let's assume FHIR is completely successful for a minute.  Let's assume the powers that be have belled this particular cat and there is a clear, concise question that generates a clear, concise answer.  They haven't been able to do this so far, but let's assume they do.  To whom do you ask this clear, concise question? 

Aye, but there's the rub.  You have a brand new shiny language to ask questions in, and you don't know who to use it on.  Let's ignore for a moment the fact that having the data out in the world somewhere, where ever it sits means that we can't find trends and perform analytical analysis on the entirety of the data, let's just do the absolute minimum and simply get the medical records for one patient who has an accident or an incident away from this primary care physician.  You guessed it, that can't be done.  Is your doctor in Chicago Medical Center East or Chicago Medical Center West?   Is she Jane P. Doe, MD or Jane Q. Doe, MD?

How do you peel this particular orange then?  The only way to satisfy the requirements set forth above is to have a central repository, an history of record, stating this procedure was performed for this patient by this particular practitioner/practice at this date and time at this location and described by a set of terminology that we can all agree on.  We have the database, and we (collectively) have the terminology. Another benefit would be the analytics we could perform on this database to look at things like population health and maybe identify emerging trends.

I am appalled and aghast that nobody seems to look at the overall problem we are trying to solve and that anyone who crows about ANY solution in the correct place gets accolades heaped upon his or her shoulders regardless of the obvious (to us only apparently) flaws of said solution.

A couple of weeks ago we had a discussion about FHIR and came to the same conclusions.  Take a minute and share this article with your the members of your networks so we can get healthcare fixed.  Take a look at our other articles and blogs and share them too.

We are Sentia Health.

This post also appeared at http://sentiahealth.com/WebLog/Details/26 today, 8/29/2016, and on linked in.

Date Written Comment By Comment