Cue the John Williams soundtrack…

I am not shy about sharing my opinions. Pausing here to wait for laughter to subside. OK.

Often that is a positive, and means that I can help push a group to surface important issues that otherwise get swept under the rug. But sometimes it gets in the way, too.

I also work for Microsoft. I am on my second stint at the company, having returned in 2006 to help found the Health Solutions group after a 10 year journey through the startup world. I am really proud of what we’re doing here in HSG, and I am extraordinarily lucky to be in a position where I have a ton of freedom to choose where I focus.

A few months ago, when I read Wes and David’s “Simple Interop” posts and then spoke with Arien during the genesis of the NHIN Direct project, I got super, super excited. I spend pretty much all my time trying to get the right information into the hands of patients and providers, and it is really hard in the existing environment. NHIN Direct, with its ambition to short-circuit some of the policy and technology barriers that have slowed this down to date, represents the best chance we’ve had to really create a ubiquitous way to move beyond this quagmire.

Because of this belief, I have dedicated a huge percentage of my personal time, and that of my team, to push the project forward. The last week or so, I have been very vocal about my concerns about taking an IHE/XD* approach to NHIN Direct. I am comfortable with a number of alternative approaches, but have been championing one (SMTP) that I believe is the best path to quick and sustained success.

Somehow along the way, the conspiracy theories seem to have started to emerge. As I understand it, the central theme is that the reason “Microsoft” is pushing so hard for SMTP is that we are looking to extend our world hegemony with Outlook, Exchange and Hotmail.

My first inclination here is to giggle a bit, and then say “great idea!” But I really care about the success of NHIN Direct — so I wanted to offer a serious and public response as well.

First — the incremental units of our email products that we might sell as a result of NHIN Direct are completely irrelevant to that business. None of those groups have even heard of this project. The vast majority of providers are already Mac or Windows Office users already.

Second — we currently sell an HIE product (Amalga UIS) that includes support for the IHE/NHIN protocols. Selling more of these is incredibly material to the HSG business (and my personal commitments). So if I were simply shilling for Microsoft products, I would be better served espousing a different position.

Everything we are doing in HSG comes back to getting the right information in front of the right people so that they can make better decisions. We sell products to help with that: Amalga on the enterprise side and HealthVault for consumers. Will NHIN Direct help those products? Yes — but only if it gets used by providers and patients. Without that, it’ll just be another silo.

That’s my concern — I do not believe that the business dynamics of an exclusively-IHE/XD* implementation will enable that ubiquitous usage.

I won’t go back over all the reasons for that, especially as we are engaged in some very positive conversations about blended concepts that might get us to a consensus position. I just wanted to be clear — the Death Star is not at work here, and I’d be more than happy to speak with anybody one-on-one more about it if they have questions.

Comments (3)

  1. David Tao says:


    Thanks for the clarification. So now all the Luke Skywalkers can feel safe! Seriously, I am encouraged by the recent conversations moving towards consensus, and I hope that the conspiracy theories (all of them) stop getting air time.

    But now I have a very pragmatic question to you as chief architect for HealthVault. One of the benefits, that is a goal of NHIN Direct, is that patients can receive clinical info at their Health Internet address. That might be a personal e-mail account but could also be a PHR. Prior to ND, one of the pitfalls was that most PHRs had a vendor-specific API/SDK that ecosystem vendors would have to "code to" to push information into that flavor of PHR. Or they'd just have to give the patient an "electronic copy" and the patient could upload that into the PHR, as I've done with HV. But I think that ND's success will imply that more PHR vendors (at least those supporting ND such as Microsoft, Google, and the EHR vendors who have tethered PHRs) would offer the patient's PHR record as an address for an ND message, which in turn might contain a CCD, CCR, or other attachment. So, voila, standards-based interoperability with PHRs, rather than every one being different, hooray! Am I right?

    I assume that there will still be a place for tight integration of discrete data (such as home medical devices) into PHRs, and that's something specialized and different. But for the basic simple interoperability, such as pushing clinical summaries to patient PHRs, ND should save us all (as vendors, providers, or patients) a lot of time.

    One thing I'm not clear about — if a provider sends a document to a patient's PHR such as HV, will it just appear there automatically, or will the patient have to do something to "accept and file" (or reject) the message? It still seems like the patient needs some control. Or do they just authorize once-for-all that sender to update their record, without having to accept individual transmissions?



  2. Sean Nolan says:

    David, thanks for the note — Completely agree that a ubiquitous NHIN-D network will put patients lightyears ahead in being able to get their information, at much lower cost to the providers that have to actually do the sending! Convincing providers and vendors to write to our API, Google's, Indivo's, Dossia's…… has been a key inhibitor to adoption (and something we as PHR vendors have all be looking to fix, with NHIN-D being the most promising avenue).

    The question of when the data "drops" to the patient is a good one, and not something I've locked on yet, although I can give you the current state of my thoughts. The challenge of course is a dynamic between patient control, which they want, and patient convenience in not having to "ok" things too frequently, which they also want. The two similar features we support work as follows:

    1. Inbound faxes are immediately posted as "files" and anybody who has access to see your "files" will see them without further review.

    2. Inbound CCR or CCD documents are immediately posted as complete "packages" (e.g., "your visit summary from NYP on 6/1/2010") and anybody who has access to see those CCx data types will see them without further review. The user also receives an email encouraging them to "reconcile" the package — which is a semi-manual process of extracting the specific items from within the package (medications, labs, etc.) and adding them as discrete data elements.

    These models seem to work pretty well because in *general* the only folks who have an ability to send data (either because you gave them your fax number or explicitly authorized the provider to work against the HV API) are people you trust. Of course, if the fax number gets out, you have a problem.

    I think the most likely initial option will be for us to follow the model of "disposable addresses" that many of the big webmail providers now use to reduce spam. A patient will be able to create as many NHIN-D addresses as they like for their record, and if they choose give a different address to each provider. If an address becomes compromised, they can disable it without impacting everybody else who is sending to you.

    This is the really fun part of the problem! I can't wait to dig more deeply into it as we go forward and see what the right balance is for folks.

    Thanks again…


  3. David Tao says:

    Thanks for your explanation. I'm very encouraged that you confirmed that the desired outcome is the same one I had hoped for, re sending to PHRs in a common (ND) way. I agree with you on the "current state" of your thoughts re the patient drops. Even though I have a HV account, and have done (and generally liked) the "reconcile" step you mentioned, I wasn't aware that someone could "FAX" a document to HV. So there's a "phone/FAX number" that can be associated with my HV account to receive FAXes? How would I get one (I couldn't find a place to put it, and besides I can't just make up a phone number)? I'd like to try it.

    As for the "disposable addresses" that might be needed, although I (as a consumer) would probably not want to do that unless I absolutely had to. Sure I could have my addresses follow some convention like,, but managing all those aliases could become a pain, especially for a person who has lots of providers.

    You're right, figuring out these details of the patient experience, trying it, and iterating until we get it right, is the fun part. Thanks.


Skip to main content