Why we need an independent health data utility

This is another in the numerous “death of Google Health” stories that have been appearing since Friday when the Google blog announced the pulling of the plug. I must admit to being more than a little pissed off with Larry Page or whomever it was within Google that made the decision. After all, Google Health was only introduced a tad more than 3 years ago (premiered at HIMSS in Feb 2008; launched officially later that year). And just nine months ago they hired a new product manager and debuted some interesting new features connecting to the new wave of personal sensors.  I know that Wall Street has been telling Google to focus on fewer products and that Page as new CEO has decided to do that but for a company as rich as Google the effort involved in keeping Google Health alive would be trivial. And props here to our friends at Microsoft who are integrating HealthVault into their wider health care business.

The sunsetting of Google Health has meant an outpouring of articles from the factual (Deb Linton at Health 2.0 News), to the historical (John Moore at Chilmark) to the winners/losers assessment (Fred Trotter) to the mega-quotes including mine (Marshall Kirkpatrick at ReadWriteWeb). There’s also been a steady stream of both sad and (sadly) happy people commenting on the Society for Participatory Medicine listserv, and Mr HISTalk was his cynical self–basically saying that tech know-nothings should stay out of our complicated health care business. He’s wrong and now Google is wrong, and here’s why.

With the very notable exception of HealthVault and (hopefully) some new innovation from Dossia, we are now dependent on a number of small companies to maintain the emerging data utility layer. The data utility layer in health is the place that is going to collectively store all the data that is being generated. Apparently Google didn’t have the real patience for two rapid developments.

First, with a combination of the Direct Project and the stipulation in the meaningful use regulations that EHR users share data with other providers and with patients, individuals are going to find that more and more data about themselves is available and easily accessible. Whether or not it’s a Farmville-type hit, the ability to capture all that information in one place is very important. Currently it’s also very time consuming to put together so very few people do it. But I do know of instances where people have laboriously entered lab values into Google Health just to store them. Sooner rather than later that data will be available much more easily in machine readable format, and as those barriers to use fall so the desire to look at that data will increase.


Other than Kaiser members and veterans almost no one is in a health system that will capture even all their “medical” data for them. And even at Kaiser and the VA patients are basically beholden to the organization’s version of what they do or don’t need (with the Blue Button being an honorable exception) in terms of both data and the ability to analyze and play with it.

But most Americans aren’t in KP or the VA. It’s much more realistic that a typical American will be using multiple providers. My own family of three has three different main care providers (one Ob-Gyn, one PCP, one Pediatrician) in three different organizations-none of which share electronic data about us. And we don’t even use specialists much! However, in order to get the meaningful use dollars those organizations are going to have to free the data and allow patients to take a copy (the original idea behind healthdatarights.org).

Don’t think that–even if they’re successful–HIEs and meaningful use requirements are going to get all this data in one place. Direct won’t do that either, although it should make it much easier for providers to share data with each other. But we know provider organizations won’t take the lead. And health insurers?  They can’t even present the data they ought to have to members now. I’m in an HMO for Anthem Blue Cross and their “360” view doesn’t think I’ve visited a doctor in two years because the claims are run through an IPA and Anthem doesn’t see them. But there is going to be much more data to come and many more people are going to be interested in storing it somewhere.

That may not mean weekly interactions with that data–that’s not the way most health care interactions happen. Most health care interactions are transactional (appointments, refill requests, lab values, et al) and the big health care players more or less stopped the Googles of the world from providing those services–although via its MedFusion purchase Intuit and some others are now linking EHRs to those features. But lack of regular interaction with that data doesn’t mean that aggregation of that data for patients is not very important. I don’t look at my previous tax returns, but I sure as hell will if I get audited. And I want my health care data available when I need it too.

Second, there’s about to be a big uptick in sensors automatically tracking sleep, calories burned, weight, pills taken, etc, etc with no need for manual data entry from the individual. I think this is going to revolutionize how people behave. We’re in a long term cultural shift but, just like Prius drivers paying more attention to the miles per gallon meter than traffic lights, we’re all going to be looking more and more at indicators and incorporating them in our feedback loops. (Here’s Thomas Goetz’s great article about it). Already the folks at RunKeeper are incorporating data from multiple devices and Healthvault does the same thing. (BTW Runkeeper tongue in cheek claims to be ready for prime time but it’s not quite on the health data side!).

Here again we haven’t quite figured out how to incorporate different data streams with each other–who knows what the impact of running on sleep is? And we don’t have any good way of tracking it against health yet–but isn’t it a better than even bet that we will learn something when all that data is being collected and integrated at just a slightly greater level than it is happening now? After all, Runkeeper has 6 million downloads and Zeo already has more data than all sleep labs combined.

I can’t tell you the true  potential for improvement in human welfare that can come from this, but I can tell you that we’re at the beginning not the end of a process. And given their huge resources I have no idea why Google wouldn’t just keep their Health data storage ticking over at a minimal level. Perhaps Larry Page could have taken fewer rides on his jet to compensate?

So what to do about the loss of a big player so soon after the game has started?

For now we’re going to have to depend on Microsoft more than is healthy, and we’re going to have to hope that the combination of smaller players and point solutions with open APIs will suffice. And maybe there is a viable play for a technology company finding a way to make this a viable independent business. But I think it’s too important for that. Eventually we’re going to have to depend on one of three solutions.

First, players in the current health care system could step up to the plate and provide a common utility for storage and data transfer. No, I’m not holding my breath either, but if they’re paid to do so and forced into it by regulation, it’s not impossible. After all that’s why they do everything else.

Second, the government could take this over. Yes, I don’t think this idea will survive Deb Peel either, but given that health data is by definition life-long and we don’t have much faith that tech companies or provider organizations will stay around in their current format, it’s a plausible possibility. After all they already have health data on millions of people–not to mention Social Security and who knows what else– and via Blue Button are making it available. Why couldn’t they create a storage layer too that we know will be there long term.

Third, couldn’t some major non-profit Foundation step up and provide this as a real public service? I know it seems far away, but in some ways this was the original idea behind Dossia? Rather than Google closing down Google Health could they give it (code & data) to an unimpeachable non-profit with enough money to keep it going for while? Other Foundations and the government could step in to fund it. I believe there’s already a strong group of potential developers and more out there who could run it.

Obviously right now none of these are on the table. But if we know that more and more data is coming, and we believe that there are positive values to individuals and society in being able to store and use it (not to mention research on it), can we not make it a priority?

19 replies »

  1. That’s one way to do it. France’s Carte Vital contains a chip that stores the info, but, yes, you’d need data standards and utilities to upload and download info to and from the chip. It’s certainly an alternative to one or multiple master storage utilities, or putting all the info in one place in the cloud.

    There’s an alternative technology called Humetrix which was shown at last Oct. 2.0 meeting that uses a USB interface to aggregate and access existing records already stored on peoples’ sites or in the cloud which can, with your permission, give access to your info to providers.

  2. “Health care providers, who hold the data of greatest value to patients…”

    JD, what data is that? Insurers hold, or can assemble, all diagnosis lists, all procedures and tests and visits and admissions, all medication lists (they actually give those to providers), and diagnostic facilities, like national reference labs hold results.
    Hospitals may have a bunch of results as well, but ambulatory physicians have very little that patients and others don’t already have.

    I would say that most of the hard data is not locked down by providers. Why isn’t there a huge outcry that insurers and labs release the data to consumers? Did all major payers connect do Google Health? Surely they are not threatened by customer mobility or visibility into redundant testing?

    And why are we thinking that Google is a good fit with a PHR? Did we imagine that many millions of people would spend lots of time on their PHR page, so they can respond to ads in a way that makes it worthwhile for Google? Or was it Google’s index &search abilities that would have made PHRs profitable for Google?
    Microsoft is a much better choice for something like a PHR and I can think of several others, including the government, before Google.

  3. OK.. I’ll bite…
    Is this card electronic? How do you intend to “update” the card with the information from the latest visit? Doesnt there need to be some “download” mechanism from the EMR (Enterprise system) into your card?

    The issue is not with the storage mechanism – it is the Enterprise update mechanism and other consumer apps that can read/write it.

  4. why are we anguishing over the failure of “enterprise” solutions to this problem? why can’t we simply carry this information with us on the back of our insurance cards and update it when we see our providers? why do we insist on making this so hard. . .

  5. JD–you’re right of course, and before I start crying I just offer the hope that — given that the taxpayer is subsidizing 60% of the health systems out there not to mention dumping another $30bn into the EMR trough — that we legislate and enforce data mobility with the same vigilance that we use for, oh I don’t know, something that matters just as much — say marijuana possession (750-K arrests a year)

  6. And now the gov’t wants to tax it, read your mail and turn it off when it suits them. Sounds like a box to get out of.

    The internet is self-sustaining, unlike Mcare and Mcaid. It is a viable medium.

    The gov’t created TV standards, also, but it doesn’t write the script (yet).

    Don’t you know the word “retarded” is politically incorrect?

    If my little comment is the most retarded you’ve ever read, then you should read your own.

  7. Thanks to the usual idiotic bombast and ideological venting inspired by this post – along with many thoughtful comments from people who actually understand the problem – I suppose we all need to keep repeating the same idea over and over, Matt. There is no real marketplace for health care information mobility.

    Like most every kind of health care market, personal health information exchange has not only failed, but is choking to death on its own failure. There should be an ATM network for the transport and retrieval of health care data, so that patients can move about the health care system without a medication interaction error killing them, and Google Health and HealthVault were the market’s last viable options. Now, we’re down to one.

    As too many of us in health IT have been saying for years now, any PHR-type application – or any platform like Google Health that supports multiple PHR-type applications – is only as useful as the data it can access, display and transport. But clearly, even the brand heft of Google was not enough to get a critical mass of health care providers to open up their systems and processes to enable this data “liquidity.”

    And why should they? Data outflow means potential patient outflow, i.e., the loss of a customer to a competitor. Even worse from a provider’s point of view, accepting data inflow up the same pipe from PHR-type apps/platforms for a new patient means loss of revenue on all those tests, work-ups, marginal admissions, and other services that access to PHR would preclude.

    Health care providers, who hold the data of greatest value to patients, are – whether they admit it or not – exactly NOT interested in customer mobility, nor in avoiding redundant services. They are interested in customer retention and revenue maximization. Why would any provider – or health plan for that matter – invest in collective IT systems like PHR platforms that make it easier to move your business, along with your data, to a competitor?

    The demise of Google Health is one more example of the perennial power of inertia in health care, and primacy of strategy and money over patient empowerment and quality of care.

  8. This is the most retarded comment I have ever read. If it weren’t for standardized and structured data in existence there would be no Internet for your comments to be heard. Now that is freedom. The government is also responsible for creating the Internet (google Arpanet) so once again your freedoms using electronics was driven by government. You seem very limited in your capacity to think outside your office. I hope this anger doesn’t bleed over into your bedside manner.

  9. You are exactly right. I do not want any Mcare or Mcaid money spent on stupid useless meaningless care. I do not believe you can show any meaningful use in any of this BS. People are not free when the government pays for stuff. Doctors are not free when they are told what they must record and how they must record it. There is no untainted “evidence” when the gov’t funds the research.

    Wake up.

  10. Thanks for the thought provoking article…

    I think that there needs to be a shift in our broader healthcare culture – both within healthcare and among consumers – before there is wide acceptance of the need /use of a health data utility.

    For consumers, this shift could be driven by radical risk/cost shifting onto consumers. Or a killer health / lifestyle / wellness app, perhaps some of the sensors that you identified. Or something unforseen.

    I think that it will take even more radical changes among healthcare providers. EHRs (largely) don’t interoperate because the healthcare organizations that buy and implement them don’t prioritize this. And the idea of incorporating patient generated information into care processes is mind blowing for many health professionals. The use of data in healthcare – for quality improvement, operational measurement, for research, etc. – is also in its infancy…and thwarted by myriad cultural, emotional and legal issues (obscured by technical challenges).

    I have little faith in meaningful use or HITECH…or even ACOs driving this change. It will take radical change in our healthcare system to drive the need for an independent health utility,

    If healthcare organizations or consumers wanted…needed access to this data…it could happen. Problem is that they presently don’t.


  11. Ok so I guess you don’t want government money in any capacity either. No medicare, medicaid, meaningful use money, etc.? Or want to be able to leverage the massive amount of health data available to you for retired service members/veterans? That doesn’t sound like a very healthy response.

  12. Matthew… I believe your question is incomplete… It should be
    “Why we need an independent health data utility SOMETIME IN THE FUTURE”

    Personal health information (PHR apps etc) aimed at the consumer are all important and will be required… but are ahead of their time. Our nation needs to go through some more changes before this will become important and desired by the masses…

    I am not saying there arent folk who use it today.. just that adoption is and will be very low for a few years (2-3 at least). This is why Google shelved it.

  13. Hasn’t there been a fairly consistent message of consumer indifference to PHR’s that has limited commercial demand for this product? Isn’t that where Dr. Koop started?

    Perhaps us consumers are too stupid to know why it’s essential. Or too busy. Or perhaps it’s simply too much trouble, owing, as Jim Speros suggests, to the complexity of integrating diverse personal health data sources. We’ll all be on Medicare before converging data standards fixes this problem.

    France’s Carte Vital is looking better and better. . .

  14. “We” do not need this.

    “We” do not want it.

    “We” want the government to keep its paws off our healthcare information.

    If “you” want it, knock yourself out. Create a product and a market and live long and prosper. Don’t make anyone play who does not wish to either freely pay or play.

  15. Matthew – thanks for a terrific, thought-provoking post. As an early proponent of PHRs as platforms, I was delighted when Microsoft and Google launched HealthVault and Google Health, respectfully. We started Project HealthDesign in 2006 without them – actually urging our grant applicants to imagine that they existed. So watching Google Health’s death spiral for a while has left me a bit dispirited.
    I’ve thought a lot about this data utility layer over the years and I have to say I’m not enthused about any of the three approaches you’ve outlined. On the one hand, I agree that the layer is really important (my post last week on how all the apps come together speaks to that), but I’m not sold on the notion that there should be one. The computer industry has generally figured out how to manage with two or three competing platforms (Windows v. MacOS, Android v. iOS) that push each other and developers seemingly tolerate having to write versions for each platform. My sense is that the data utility layer will need to evolve quite significantly in the coming years as we add more and more data types, more ways to analyze across data, more ways to make the data social and so on. While the stability that a single platform would bring might be nice, I think it’s important for platforms to emerge from the market – for companies like RunKeeper to build out from their starting point and grow tentacles into the health care system, for companies to use the Indivo/SMART Platforms approach to bridge the health care and the consumer data divide, and for HealthVault to continue its push. What’s most important is that real competitors push each other to get better and better. That’s why Google Health is a loss, but perhaps others will step up.

  16. Matthew, you are looking for a solution in the wrong places. The healthcare IT market needs to be seriously disrupted so the needs of patients and providers are met. With very few exceptions, established vendors, government and foundations have never developed or launched truly disruptive solutions to anything!

    Established EMR vendors are focused on adding more features to their existing products (the very products providers have resisted for years and that have never met the needs of patients), not on developing new products that meet the currently unmet needs of these two groups.

    Government has the power to compel providers to do its bidding, but government has rarely, if ever, created disruptive solutions that meet market needs. It typically relies on established providers as it is doing with HITECH and Meaningful Use which, in my opinion, will cost a staggering amount of money but will not get the job done.

    And foundations? They have neither the creativity — disrupting markets and meeting needs isn’t what they do — nor the clout of government to force anyone to do anything.

    Personally, I’ll put my money — and have — behind entrepreneurial companies who defy “conventional wisdom” and develop new approaches that have the potential to change healthcare IT.

  17. is it possible that the primary barrier we’re facing is that health data standards are just too doggone complex? That by the time we sort out the CCH format and the moving-targer HL7s the whole thing becomes too complex for even a Google to make happen?

    What about VA’s Blue approach — a simple ASCII download which gives you your data and did so with sufficient structure to be uploaded into an EHR?

    Yes, complexity has benefits for some. But those benefits come at the price of complexity, high maintenance and too-high cost.

    What if Thoreau was right? Again?

  18. This is absolutely true, and puzzling that there is not a run on this space by companies large and small. The author’s reasoning behind the need for such a service is very apt. One challenge is the lack thus far for a standard data format; once a standards set emerges from increased ubiquity of EMR (MU catalyst), apps will be developed that very easily use one’s data to produce solutions that have value both to patients and payors, providers, and other stakeholders.