Matthew Holt

Kaiser tiptoes into HealthVault & tells THCB about it, with UPDATE

Kaiser Permanente signed on this morning for a pretty extensive pilot with Microsoft,
allowing its 159,000 employees to copy their online health records into HealthVault. This is a big coup for Microsoft and a fairly ambitious move for KP which to this point hasn’t said much publicly about the data transferability it was going to provide for its members. This is a clear signal.Kp

Assuming that the pilot is a success, presumably all Kaiser members using My Health Manager (over 2 million now and heading to 3 million at years end) will soon be able to move their data to HealthVault. We are potentially seeing the first real example of mass scale data interoperability onto a platform not connected to a health care organization. And obviously, Google is playing in this same space too.

Once the data is collected in HealthVault, there are lots of possibilities for what can be done with that data, and what services can be offered.

Back in the days when Justen Deal was causing havoc with HealthConnect, I had a somewhat unorthodox interview with Permanente’s Andy Wiesenthal — in which (without KP’s PR folks knowing) I called him in a taxi on a cell phone late on a Friday night. Perhaps it’s a mark of how far THCB has come (you decide if it’s good or bad) that in regular business hours on Friday, KP’s publicity machine lined me up for a pre-release interview with Peter Neupert, Corporate VP of Microsoft Health Solutions Group and Anne-Lisa Silvestre, VP of Online Services at KP.

I spoke with them about what they intended to do, why they were
doing it, and where they thought that it was taking American health
care.

Here’s the interview.

UPDATE: I apparently missed something quite important in a comment
made by Anna-Lisa Silverstre.  Here’s how Steve Lohr reported it in the
NY Times:

Kaiser chose the Microsoft technology, even though Ms. Silvestre is
a member of a group of health professionals advising Google. Ms.
Silvestre said Kaiser also looked closely at Google, but was
particularly impressed by Microsoft’s technology for protecting the
privacy and security of personal data. And she noted that Microsoft and
Kaiser are using the same Web-based format, called a Continuity of Care
Document, or CCD. Ms.Silvestre said Google is initially focusing
its efforts on a different format for Web health records, called
Continuity of Care Record, or CCR.

But these different Web
standards, analysts say, are not a long-term obstacle to consumer
health records from rival companies being able to share and view
information. Google, for example, is committed to supporting both
formats, and so is Microsoft.

“Google will certainly be a player
in this, and we look forward to working with them,” Ms. Silvestre said.
“We’re all for consumer choice.”

However,
some people are reading rather more into this choice in formats.
Apparently the CCD format makes the data viewable rather than
delivering discrete data elements that would make it interoperable. I’m
hoping that some who know more than me will chime in here, but my
impression is that it’s the equivalent of moving a copy of an html file
than handing over a full database. Please comment on this below.

11 replies »

  1. In response to the anonymous rant:
    Actually according to the HealthVault sign-in page, they accept openID from at least two providers. I’m going to assume they aren’t lying on the sign-in page.

  2. As for Authentication of the Individual, does this mean that the we all need to update to the new MS Passport routine for Secure access to their HealthVault PHI data?
    I’ve had MS Passport since `99. When I populated my PHR data, I was required to re-configure my Passport codes to their new more robust alphanumeric password code set.
    Does this mean all users will need to have a MS Passport in order to share their PHR on the MS platform?

  3. One clarification – Be Well seemed to be largely scoffed at by a number of people in the blogosphere when it was announced but some of the ideas and use cases were very interesting even if it is a ways before truly implementing them.

  4. Attended the Microsoft Conference and some points:
    – Number of Sr. Senior Engineers there so you know that a number of vendors and partners are treating this quite seriously. Vendors don’t commit scare resources like this unless they really think it is something worthwhile.
    – Much of what Troysterr said is true from a technical perspective. Have to really read more posted on the Microsoft site but one of the biggest potential biggest issues I see down the road is that Microsoft is building the data elements in a very ad hoc fashion (now up to 32 distinct data elements that have considerable range in variation) and using very granular data elements that are very small.
    While this approach has its advantages (e.g., less likely to need frequent versioning of the Healthvault code, fits into Microsoft’s privacy and security approaches), I bet they are going starting having some real headaches when they several different transformations available for a particular data element.
    – Microsoft is being very heavily driven by the input from the partners I talked with. Not just cheap PR rhetoric. This includes just about the entire data element process but perhaps more importantly what “works” and won’t work in a particular setting/environment.
    Going to be very interesting to see what role that the early adopters of Amalga like New York Presbyterian and St. Joseph’s System play and how this effects HealthVault.
    – Kaiser made big waves but there were a number of large companies that supposedly made substantial commmitments to HealthVault including CVS/Caremark, Humana, U.S. Military Health System, Ascenion Health, and others. Detail is in the details though as always.
    For example, the CIO of Humana said that HealthVault will be integrated with “My Humana” website after some pilots are conducted this year with select Humana employees and clients (large employers likely). What exactly does this mean though and what data is going to be pulled from EOBs and other data sources from My Humana into HealthVault?
    – I also wonder for example if somebody like Walgreens (a big partner with Google) will be willing to partner with Microsoft too. Didn’t see anybody from Walgreens there (doesn’t mean alot because there are at least a good 400-500 people) but hopefully this doesn’t degenerate into a “Healthvault” vs. “Google Health” partner arms race. It in the end consumers would be the ones who lose.
    – Finally have a bit more clarity on how Microsoft is going to monetize this (e.g., adding robustness to increase search numbers through Live Search and take a sliver of the overall search market) and some potentially unique use cases of advertising down the road geared toward existing online advertising players in the health care space (e.g. pharma/biotech) and relatively new ones (e.g., providers).
    – Be Well Fund winners had some very interesting ideals. This was largely scoffed at but I talked to an old friend who is a research scientist at Children’s Hospital in Seattle now (one of the Be Well Fund winners for their idea regarding the transition of pediatric diabetics to adult practices) and seems very interesting.
    Microsoft was still a bit vague on how the independent developers and entrepreneurs are going to contribute but frankly this is where I see the most interesting potential use cases for HealthVault down the road.
    Grad Conn also mentioned that Microsoft is going to be setting up a site (Copernicus something) that is going to be throwing out an idea a week that Microsoft would either think would be “cool” for a developer to work on. Have to see what this is really about before commenting on.
    – Up to 36 applications now (sees like everyone is from a large vendor or organization like the AHA) with another 50 apps in the pipeline in the near future. Still, 2008 is really another development year for HealthVault and Microsoft readily acknowledges. Even remains to be seen how far they get in 2009 but they are positioning themselves for the Home Health Care market by partnering with a huge number of vendors (over 50 devices already with some of the key vendors with consumer devices including Polar) in this space and enjoying interoperability with their devices thorough the Connectivity Center.
    Anything is thinks this is a “bad thing” just is a really a cynic or has a beef against Microsoft. There are several issues/concerns I still have with this approach but they are putting their money where there mouth is in terms of the resources and energy they have devoted to this.

  5. I don’t have a horse in the race, but I feel compelled to clarify/correct some of the comments by “Anonymous”:
    1. Yes, Microsoft is the author of .NET. However, it’s in use by millions of developers the world over, so its proprietary nature is hardly a hurdle.
    2. False. Microsoft .NET is not necessary to develop for HealthVault. It uses industry standards, and Microsoft is even developing toolkits for Java and Ruby. They do provide richer toolkits for .NET, but that doesn’t preclude anyone from using Perl, Python, COBOL, or whatever they wish. http://blogs.msdn.com/familyhealthguy/archive/2008/04/14/opening-up-the-vault.aspx
    Also, Visual Studio Standard edition is free if you choose to develop on the .NET platform but don’t want to pay for the Professional edition.
    3. This is no more true for .NET than any other platform. There is a rich ecosystem of free support in the form of blogs and user groups for .NET, just like there is for Open Source.
    4. It is true that C# isn’t Java. Also, Chocolate isn’t butterscotch, and Ford isn’t GMC.
    5. Only Kaiser has publicly expressed their view that Microsoft’s security model is impressive. I have to believe, though, that at least some of the dozens of other companies that have signed on with Microsoft have investigated the security model and find it at least adequate.
    6. This is only bad if (1) you are correct that they won’t share their code, (2) it is true that their code is of value to other parties that already have the toolkits provided directly by Microsoft, (3) the industry is materially injured by the absense of that code, and (4) you accept as fact that “open source” is good and anything else is bad.
    7. Because they want to and they believe it will help people. They may be right, or they may be wrong. No harm, no foul.
    8. See #6. Also, the objective here is to help patients, not the Open Source community.
    9. Apparently Kaiser sees additional benefit in interfacing with HealthVault. Again, if there is no value, no one will use it. No harm, no foul.
    10. True, but you don’t indicate why that’s “bad”.
    11. Also true. Since the vast majority of computer users run Windows (or Macs w/ Windows emulators), that hardly seems like a disadvantage. I could also name a few applications written for Linux/OS X that don’t run on Windows. But that’s all irrelevant to this discussion, since HealthVault and Google Health are both standards based, and accessible via the Web in a cross-platform environment.
    I’ve been in healthcare IT for 15 years, and I see this as a very positive sign that true interoperability may be a reality in my lifetime. I look forward to other large healthcare companies working with HealthVault and Google Health (and Revolution Health, if they are successful). It’s all good for patients. The religious war between Microsoft and Open Source can continue outside on the patio.

  6. I’m grateful for all comments. My sense is that what Jamie Ferguson says directly contradicts what little bit I understood about CCD. And I stress I was going off what I was told by someone else.
    Many of Anonymous’ criticisms of Microsoft are true, but I don’t think they are relevant. The fact is that despite the success of the open source movement the vast majority of software is not open source, and of course those vendors including Microsoft are not going to release the code. So their attendant business and security issues are what they are.
    There are two issues intertwined here. Can Kaiser Members move all their data in a format where it can be then moved into some third party’s data schema via Healthvault? The point is not whether it’s open source the more relevant question is, “can all my medical records from Kaiser, or at least the data elements in the CCD/CCR, be moved over to HealthVault in such a way that they can be included in the EMR database of another company such as a competitor to Kaiser like Sutter?” Can the third party just see the records, or can they take data from them and include it in their applications? And would the answer be different if KP choose CCR over CCD?
    I’d like to be more hard hitting and tougher about my questioning but I simply dont understand the technology enough to ask the right questions, which is why I’m asking for my readers help.
    The overriding business question is, are health plans (e.g.KP) really moving towards letting consumers take their data with them? That’s what Anna-Lisa Silvestre said.

  7. Over the next few days and weeks there will be a media love fest over the recently announced partnership between Kaiser Permanente and Microsoft HealthVault. The two large corporations, one not-for-profit and the other very much for profit have agreed to exchange Kaiser Permanente health data. The details of the actual mechanics are not widely konwn other than Kaiser Permanente and Microsoft touting that the exchange will take place via XML schema of CCD.
    Why is this bad for Health 2.0, Web 2.0 and the FOSS movement? Oh let me list the ways:
    1. Microsoft coding technology is proprietary to Microsoft.
    2. Pay to play; developers must purchase Microsoft Visual Studio.NET to code HealthVault applications
    3. Pay to fix; developers must purchase support calls when something goes wrong on the Microsoft side.
    4. C# ain’t JAVA
    5. Only Kaiser Permanente seems to think Microsoft privacy and security model are impressive.
    6. It is doubtful that Kaiser Permanente will open source the code required for interoperability between Microsoft Health vault and Epic Systems (the Kaiser Permanente EMR vendor).
    7. Why is a health care company writing a software add-on to a vendor product to interoperate with another vendor product?
    8. The code can never be open source so the only beneficiaries are Microsoft and Kaiser Permanente.
    9. Kaiser Permanente already has this feature available on thier web site. So why are member premiums being used to pay for redundant functionality?
    10. Microsoft does not use OpenID.
    11. Microsoft applications don’t run on Linux or for the most part Apple.
    There are additional problems with this announcement that just boggle the mind. For instance as of 9:20 PM PST The Kaiser Permanente website is offline. Not exactly confidence building.
    None the less it is disturbing that a major health care company that has a not-for-profit status and serves nearly nine million people has made such a public endorsement of a public company. A company that many of its members who live in California probably would rather see Kaiser Permanente, partner with a more open framework for PHR.
    Side note: Mathew we’ve come to expect a bit more hard hitting pieces from you. This seems like a fluff piece for Microsoft and Kaiser Permanente. Where is the indepth analysis and tough questions?

  8. As the Kaiser Permanente director of health IT strategy and policy, I am more than a little familiar with these formats. Probably few would care about this detail, but the reporter actually got that last bit exactly backwards. CCD represents the data in HL7 v3 CDA XML for improved interoperability.
    CCD is the recognized federal standard for interoperability of medical summaries, required in 2008 FEHB contracts and in the 2009 CMS call letter, and it has greater specificity than the competing standard chosen by Google.
    Both of these standards may deliver discrete data elements and both may be viewed as web documents. CCD has more specific requirements for each data segment template, whereas with Google’s choice you may have to accept nonstandardized data that different providers stuck into the segments.
    Jamie Ferguson

  9. Matt,
    I’m no expert on the technical details, but I’m sure that the difference between CCD and CCR is not that one provides interoperability while the other does not. The point of both of these standards is to make data transmittable across systems and thus produce interoperability.
    CCD (I think!) transmits data in xml format, and thus is more ready-made for web viewing. It is also broader than CCR in that it incorporates HL7, which (I think again!) provides a much richer set of clinical categorizations.
    In other words, the sense that I get is that CCD is more advanced than CCR. It’s also newer, and I think HealthVault and Kaiser are pioneers in using it.
    Now please, someone who really knows something about this, chime in.

  10. At the Microsoft Healthvault Solutions Conference and interested to see what they have to say over the next two days. Agenda is full of a pretty varied group from all segments of the healthcare community. Even the big national health plans have set some senior people which is a surprise to me.
    About the only notable absence I can see from the list of attendees is Epic. Even some token representation from Cerner (which is kind of ironic given the tongue lashing that their CEO gave PHRs just about a week ago).