Web Log


Latest Healthcare Trainwreck: IBM Watson Partners with FDA on Blockchain Driven Health Sharing

1/17/2017 12:21 PM

In the latest installment of "that's now how it works, that not how any of this works" by Anne Zeiger on EMR & HIPAA  The FDA picked a privately held company to partner with.  First, The FDA can't 'partner' with anyone, they are a government agency.  they can hire IBM to solve a problem, but they should not.  Second, IBM's Watson is a Natural Language Processor.  It doesn't have anything to do with health or databases.  Further, you can do the about 95% of what Watson does with any speech recognition software and a Google/Bing search.  Third, Blockchain is nothing new.  Blockchain is just a database.  That means that it is a central repository kind of like we have been advocating for over a decade.  

So what we have here is IBM using its brand recognition to even further muddy the waters of healthcare interoperability.  As stated above, Watson has nothing to do with healthcare, at least until we start needing natural language processing.  We might want it, but I doubt we will ever need it and it will be completely useless until we have that central repository database we here at Sentia Health have been advocating for years.  We have been advocating it for years, because we went ahead and built it years ago.  

Blockchain is the power behind bitcoin.  Any time you make a bitcoin transaction, the transaction details are added to the database.  We have had several articles about this over the years that you can look up and read.  The 'added' portion is important because it is like a ledger system.  All our databases use this 'added' notion because it gives us the ability to see all the transactions, in case something gets sideways, you can manually correct it.  We stated that the financial industry model (a clearinghouse) wouldn't work  because we need to see all the history for any given patient.  While the central repository is similar to the Blockchain concept, it isn't the same.  Blockchain was designed and built to manage financial transactions.  Person 1 pays Person 2 X amount of bitcoins.  In medicine, a patient could present with a headache in the left dorsal region.  They could present with his of her heart on the wrong side.  He or she could present any of several hundred thousand ways, with any of several hundred thousand concepts applying to several hundred thousand situations or body structures.

The abundantly clear single point of failure is that all the practitioners in all the world are going to have to hire software developers to scrape data out of an existing Electronic Health Records Management System (EHR), hook up to the Blockchain's web service and transfer said data in the correct format (ICD codes won't work because that only documents procedures performed, not documenting patient encounters).  There is a database for this called SNOMED_CT.   Here are two problems with using SNOMED_CT: first, the Blockchain people don't know about it, probably, and second, the existing EHRs don't use it anyway, so have no ability to translate what is in their EHR database into SNOMED_CT codes. 

Clearly, the FDA has no idea what they are doing.  Clearly IBM has no idea what they are doing, unless this is a publicity stunt.  If it is a publicity stunt, shame on them for creating noise that obscures real work done on real solutions.  Clearly, Blockchain isn't capable of handling this kind of data, currently.  ...and maybe ever.

Speaking of real solutions, those of you who follow along know that we here at Sentia Health never point out demonstrable idiocy without offering an alternative, better solution.  Here it is: our insurance company vends its own EHR based, of course, on SNOMED_CT.  Use of that EHR is free for medical practitioners.  we have designed and built a way to pay practitioners and practices in real time for procedures performed.  Our insurance costs $10/month plus the actual calculated cost of the care spread out over the insured population.  This is going to equate to about 20% less than ACA Exchange policies, and anywhere from 30% to 50% over private policies.  We will save even more as adoption increases and doctors start using our free system (with integrated reporting) instead of paying $30,000 average per year on buying an EHR and another $42,000 per year on efficacy and compliance reporting.  Even better, the practice won't have to have a medical coder or a billing department as payments happen in real time.

This new kind of insurance will be wildly popular because it doesn't cut payments to practitioners and is easier and cheaper for patients.  As they flock to our EHR, we build that central repository that the FDA is bumbling around looking for.  This isn't a pipe dream or vaporware.  We are planning releases in India, Central America and the Caribbean this year.  Yes, we will have to build some web services so the die hard, old school users can integrate their (literally) several hundred million dollar hospital EMR installation with our free system (that doesn't make much sense does it?) but that won't take more than a couple of weeks (again literally,  Our sister company Sentia Systems generates most of the software) to build.

So yes, the FDA is fumbling around showing us they don't really understand the problem, IBM is spinning, posing and posturing.  In the meantime, we have the real solution.  Maybe you should call the FDA (or your senator/congressman/president) and have him or her read this.  If you are going to partner with someone, partner with someone who can really solve has already solved the problem.

Date Written Comment By Comment