Modeling Health Data Architecture After DNS

I was absolutely intrigued by the idea of structuring the healthcare data architecture after DNS. As a techguy, I’m quite familiar with the structure of DNS and it has a lot of advantages (Check out the Wikipedia for DNS if you’re not familiar with it).

There are a lot of really great advantages to a system like DNS. How beautiful would it be for your data to be sent to your home base versus our current system which requires the patient to go out and try and collect the data from all of their health care providers. Plus, the data they get from each provider is never in the same format (unless you consider paper a format).

One challenge with the idea of structuring the healthcare data architecture like DNS is getting everyone a DNS entry. How do you handle the use case where a patient doesn’t have a “home” on the internet for their healthcare data? Will the first provider that you see, sign you up for a home on the internet? What if you forget your previous healthcare data home and the next provider provides you a new home. I guess the solution is to have really amazing merging and transfer tools between the various healthcare data homes.

I imagine that some people involved in Direct Project might suggest that a direct address could serve as the “home” for a patient’s health data. While Direct has mostly been focused on doctors sharing patient data with other doctors and healthcare providers, patients can have a direct address as well. Could that direct address by your home on the internet?

This will certainly take some more thought and consideration, but I’m fascinated by the distributed DNS system. I think we healthcare data interoperability can learn something from how DNS works.

About the author

John Lynn

John Lynn

John Lynn is the Founder of the, a network of leading Healthcare IT resources. The flagship blog, Healthcare IT Today, contains over 13,000 articles with over half of the articles written by John. These EMR and Healthcare IT related articles have been viewed over 20 million times.

John manages Healthcare IT Central, the leading career Health IT job board. He also organizes the first of its kind conference and community focused on healthcare marketing, Healthcare and IT Marketing Conference, and a healthcare IT conference,, focused on practical healthcare IT innovation. John is an advisor to multiple healthcare IT companies. John is highly involved in social media, and in addition to his blogs can be found on Twitter: @techguy.


  • OK you have officially said the scariest thing I’ve heard this week/month and maybe this year. I guess depending on what you mean by modeling. Somethings that dns does I would applaud, most I would be horrified at:

    1) It’s public – do you want your info to be out there at all?
    2) It’s discoverable – do you want people to find your records?
    3) It’s not authoritative – how are you doing to disambiguate the rightful owners?
    4) Registrations cost money – should you have to pay to have your own health data?
    5) Registrations expire – if you don’t have the money to keep your account active, it just goes away and someone else can take that address?

    On the other hand lessons from dns that would be nice:

    1) It’s a concept that has stood the test of time (since 1983)
    2) it uses characters not numbers
    3) it can be abstracted – logical vs. physical
    4) it scales
    5) Both computers and people can use it

    So – I guess it really matters what you mean by modeling. Why wouldn’t just an email address work – as is suggested by BlueButton+. And don’t get me wrong it’s not that I’m not a fan of DNS, I had to understand DNS well enough to explain the migration of to to Paul Mockapetris in 1988. I just worry that it’s not the right model for a Health Data Architecture.

  • Greg,
    Good. I love saying scary things. The best ideas come out of things that are scary, but that as you dive deeper you can find a happy balance with the scary.

    I think for the first couple, the key is that the location is public or at least discoverable. Not the information that’s stored at that location. It’s like saying that someone knows your address knows where you live, but doesn’t have a key to the house unless you give it to them.

    The money one is a tough question an worth exploring more. In fact, it’s a corollary to what I describe above as far as “what do you do when someone doesn’t have an address?” Do you charge for the service? Is it a feature that comes with a primary care provider? I think there’s a lot of models, but those would have to be fleshed out.

    The beauty of this is we take the principles we like and apply them and not the ones we don’t like. Or at least we find workarounds for those other issues. I think it broadens the thinking around distributed data though.

  • This goes back to the idea that people want there medical records easy to get. Some do, most don’t…or at least don’t care.
    Those that don’t care are generally fortunate enough to be healthy/have a healthy family.
    We are getting to the point of having to force people to buy health coverage…if this is the case, why do we think people need or want a simpler way to get their medical records?
    Until there is either great demand OR we are forced by the gov’t to do this…I don’t see it happening.

  • John Brewer,

    I think you have hit on a fundamental truth. People actually just want to be healthy. Looking at health records kind of reinforces the fact that we are mortal and perhaps we are losing the battle, which, well isn’t what everyone wants to think about.

    More commonly, people are are in denial. Yes, I went to the doctor – but I’m fine. You are right, there is a motivational and cultural barrier to people managing their own health at all.

    But (to John Lynn’s point) – if/when people want to get their records – It does seem like some level of centrally navigable abstraction from people to their records would be nice. I just think email is better, simpler and already adopted in BlueButton+.

  • Interesting concept — using DNS to access potentially all patient data regardless of format. I’ll just add one more problem with this at this point — IP bandwidth.

    The current IPV4 system is limited to 2**32 IP addresses, and we’ve long ago exceeded that. Currently, one resolves this problem by dynamically allocating from a pool of available “hard” addresses “on the fly”. We were supposed to by now be well into using IPV6 — 2**128 IP addresses, whereby nearly every organism or object could have its own unique “hard” address. We’re simply not there yet.

    When that finally happens, lots of interesting things will result, like everyone’s computer or smartphone could act as its own web server. We’ll have to wait awhile to see how this plays out.

  • It sounds like an interesting solution until you spend a little time with the idea and realize that it not only doesn’t solve the real problems of sharing healthcare data, a/k/a “interoperability”, it also introduces some new ones

Click here to post a comment