A National Universal Health Record (UHR) Database – Doable Any Time Soon?

Could a single, mammoth database solve all our health data needs? Margalit Gur-Arie, whose writing and ideas I greatly admire, has been arguing for one quite passionately on her personal blog in a couple of recent posts (part I, part II).

The crux of her posts is this:
– There should be a single, standardized national database to which physician practises, and ultimately EMR vendors, must submit mandatory data, “in real time”. The requirements will be along the lines of current Meaningful Use requirements.
– This database will be accessible to vendors and entrepreneurs alike, and can have multiple EHRs or apps built atop them.
– Since the patient data is available, and easily accessible (no one “owns” the data, they only own the proprietary bells and whistles they perform on the data), this is a near perfect patient utopia.

It’s a great idea and perfect for an ideal world. Except:
Massive databases cause massive headaches, as commenter Omowizard pointed out. There is a price to pay for data available at all times, all places, and by everyone. And if I may add, in Gur-Arie’s model, it’s not clear who’s left holding the bag. Presumably the government. Which opens a entirely different can of worms about data ownership.
Real time updates of data is no joke. At my current place of work, we perform quasi-real time (twice daily) updates of patient visits to client databases from a central repository. The sheer volume is enough to bring down the database servers for a good hour or two.
– We haven’t been able to agree on a standardized schema passed for a healthcare database. What are the odds of this idea ever catching on?
How are we going to mandate data population? After physicians and care organizations, will EMR vendors be the next recipients of government bribes/largesse/sops to induce them to populate the database?
– Gur-Arie herself points out that American enterprise being what it is, if there are no financial benefits to data ownership, they’re going to be a hard sell.

And while it’s easy for me to write a smart alecky blog post about the infeasibility of the mammoth database idea, I shudder when I think of what we have now: disjoint EHRs that don’t “speak” to one another, walled gardens and proprietary ownership of data that pretty much lock physician office in, PHR offerings from companies like Microsoft who will do God knows what with OUR health data.

I don’t think there are any easy answers. I’m leaning more towards an open source health “OS” platform rather than a single database. But at the very least, Gur-Arie offers some great food for thought.

About the author


Priya Ramachandran

Priya Ramachandran is a Maryland based freelance writer. In a former life, she wrote software code and managed Sarbanes Oxley related audits for IT departments. She now enjoys writing about healthcare, science and technology.


  • Howdy John,

    Thanks for reviewing my humble opinions. I would like to clarify some of the points you rightfully bring up in the above.

    1) Yes, big databases are big headaches, but with today’s technologies and the advancing big data technologies, these headaches ought to pale by comparison to running hundreds of thousands of on demand interfaces between hundreds of thousands of systems.
    2) Who’s holding the bag? Unequivocally the Federal Government. If I have a choice as to where my data is hosted, I will always pick the Feds over corporations. Neither one is particularly enticing, but corporations are less so than a tax payer funded entity.
    3) Agreeing on a schema for the database should be a step by step process instead of trying to boil the ocean as we now do. Start with the most essential codified data elements (ICD, CPT, Rx, allergies, immunizations and labs). We already have agreement on those, so start there and move forward incrementally.
    4)Mandating data population and use of it in real time should make use of the current Meaningful Use framework, and make this a certification requirement for vendors. No additional bribes needed.

    You know, it’s not as difficult as it sounds. When you do electronic prescribing you access real time prescription history from Surescripts, on demand. Same thing in an ER. It’s not trivial, but it’s not beyond our abilities, and it’s not something that will get bogged down in local politics and profit seeking misbehavior.

    If the government would throw in a bare-bones web based EMR of their own, displaying and collecting the basic elements only, and qualify it for incentives, particularly for doctors who lack the cash for private fancy (and complicated) stuff, I bet progress will be made by leaps and bounds.

    Sometimes I wonder though if we are truly trying to make progress, or just spread some cash around to those who don’t need it most.

  • Margalit,
    Just so you know, Priya Ramachandran wrote this post, not me. So, I can’t take credit, but I will take any responsibility if needed.

    I’ll let her answer some of the points you make. One missing element in the above discussion is the issue of privacy. You know how the privacy nazis will come out of the woodwork and do an #OccupyUniversalHealthRecord all day for something like this. Of course, I personally love the idea.

  • My apologies, Priya.

    Privacy?…. Oh well…. At least the government is not very likely to bombard us with Viagra and Depends running as a scheduled task in that database….

  • This is actually quite doable, and in a pretty awesome way – pardon my technical babble, but a nice highly scalable NoSQL database with an easy to use REST API, proper security controls and possibly an OAuth based authentication/authorization layer would be the perfect technology backend for this project.

    With such a platform in place, mobiles, tablets, web applications and desktop applications could use REST-based web services to communicate and inter-operate with this “platform”. Heck, this platform could even be used as an web-based service bus to allow distributed applications and systems to inter-operate THROUGH the platform (kind of like universal service discovery and UDDI was supposed to do).

    Also, if you’re worried about a standardized data model for healthcare, it’s really not too bad. We already have them in X12 for billing, HL7, CCD, etc. Make a new simple one based on all of these formats and start with simple requirements, and it’s doable.

    Thought #1 – start with a national patient identifier – the same system can be the core of this “single” EMR database.

    Thought #2 – I am almost certain this is what the GE/Microsoft partnership was about.

    Thought #3 – I am sure this can be done, and was trying to fit together some of the pieces myself, but you need a lot of support.

    Thought #4 – Have lots of lawyers.

    Who’s in? 😉

  • A few more thoughts toward Margalit (by the way, I like the direction of your thinking – knock down walls, not put them up!):

    – Facebook has a database that includes more than 800 million users. If they can do it… it’s doable – and as we know, they are already facing the privacy issue head on.
    – would you trust a ridiculously funded non-profit or a NGO with your data?
    – if we build the connectivity, does the government even need to build the front-end “bare-bones” EMR? I would prefer they let people build and compete on those front-ends, and they provide useful back-end services instead.
    – mandates are nice and all, but if the solution is good enough and actually solves a problem that needs to be solved, maybe people and corporations would use it on their own accord if it was.. you know.. free, supported, fair, and actually worked, and then it was marketed well.

  • Jon,

    You are my hero!!!

    The people that insist that this is not doable, all have financial interests to keep this from happening, and even more financial interests to keep things as convoluted as possible so they can charge fortunes for “solutions”.

    I don’t think it is absolutely necessary for the government to provide a simple and certainly ugly bare-bones EMR version. I am just looking for a way to put things in motion. I have no doubt that private enterprise can rock the house if the data was made available.
    I somehow think that government is in the best position to kick start this effort, seeing how everybody is playing along with Meaningful Use, but hey, if IBM wants to step in and give Watson something to sink his teeth in, I would have no objections… 🙂

  • I’ve been awol for personal reasons, but I love how this discussion has taken off. Thanks everyone for your comments.

    Margalit, Lol @ viagra ads. You have a good enough reason for why government involvement would be good. I’m guessing you want something along the lines of public television – something that is ultimately tax payer funded… Good luck with getting political conservatives to agree on that.

    Jon, I would love for an NPI to take off in the US. And I did have hl7 in mind when I wrote this post. I think the Facebook analogy breaks down a bit because the ultimate goal of fb is profit. Something of the scale Margalit is envisioning, without the profit motive for enterprise buy in, is going to be very difficult to pull off. I don’t think it’s impossible… Just very tough.

    And like I mention in the post itself, it was easier for me to be negative Nellie on this… but the more I think about it, the more I  feel like there’s a third option we haven’t even given a thought to – community owned, community driven project a la Wikipedia or Mozilla. That might be something worth considering… If there’s a well protected, centralized database – free of government and corporate involvement – that I can feed my health data into, I might even be tempted to type up my own medical records, and maybe those of my loved ones. A Volunteer/community driven database… That’s what I’d really love to see.h

  • Priya,
    I find your idea of a community driven repository really really interesting. I’d have to think a little more how that would come together, but it could be quite interesting if done right. Plus, how weird is it that people will go and obsess over making a wikipedia page for some strange person, event, topic, etc that they have no real relation too, but they won’t spend time taking care of their patient record?

  • John – I was hung up on “corporate altruism” when I wrote the post, but then there’s plenty of Internet altruism going around… In terms of models, I personally like Wikipedia and WordPress.

    Wikipedia is a cake to use. With bbcodes and such, they made the barriers to entry so easy that anyone that felt like could post something they felt qualified to. I don’t think it’s weird at all that people post on wikis… Clearly they have that emotional connection which makes them think they are related.

    I like WordPress for its business model. There’s the free software that anyone can download, and the WordPress.com component that they’ve monetized. That might be something worth looking into for a free health database – e.g. printable data for the free version, and XML based data for paying customers or whatever. Initially it might be just a few technogeeks entering/maintaining their data (which they already do… remember the self-logging trend)… But soon it can become a case like where users tell their PCPs to look up their records in XYZ database, and maybe that will in turn get business interested in getting this off the ground.

  • lol—let’s try this again. Government + Business = Fascism. Melding health care and government will not solve any problems, only create new ones. UMRs, EHRs, EMRs are not sustainable without creating the mark of the beast. Period.

    Beware the governmental industrial complex.

  • @ALL
    I just hope everyone get’s an email before I’m deleted into the “memory hole” of the internet.

    Big Brother thanks you for your diligence.
    –Ministry of Truth, or MiniTrue.

  • Jonny,
    Your other comment went too far down a road that lead away from constructive discussion around healthcare IT. So, I did delete it. This one will remain.

    I don’t agree with your comment, but that won’t be the first or last time that will happen.

  • 1) We won’t have anything until we have “Standards” in place! This has been needed for several years now!

    2) Clinical Data will ‘have’ to be able flow like liquid – codified data.

    3) Patient ID – Like your SS#

    4) Absolutely no need for a Master Database. Only need is to be able to gather/aggregate patients information from local, regional, state etc….. repositories from across the US.

  • @1: which standards, for there are many.

    @2: codification will subvert any real attempt at making #1 a reality because there will be no central control

    @3: mark of the beast, RFID. No chip, no service. That will be the endgame for this proposal.

    @4: where are you going to aggregate all of the data? Into a master database? LOL. Linking up to local systems from disparate systems will be a logistical nightmare pock-marked with security holes that would make a Swiss cheese producer jealous.

    The only answer is Big Brother, Big Business, the Cloud, and the eradication of personal privacy. Not much when you think about it though; who cares because we having nothing to hide anyway…..Right.

Click here to post a comment