Beyond HIT Operability: Open Platforms Are Key

I want to begin by sharing well-known information for the sake of comparison. Both the Apple and Google Android platforms welcome the introduction of new and (sometimes) highly valuable functionality through plug-n-play applications built by completely different companies.

You know that already.

Healthcare IT companies welcome you to pay them great sums of money for enhancements to their closed systems. This is on top of substantial maintenance fees that may or may not lead to hoped-for updates in a timely fashion. (With all due respect to the just-announced CommonWell Health Alliance, Meaningful Use does mandate interoperability. The participants are, in effect, marketing what they have to do anyway to try to differentiate themselves from Epic.)

The respective results of these two divergent approaches are probably also familiar to you.

Consumer technology has taken over the planet and altered almost every aspect of our lives. These companies and industries have flourished by knowing what customers will want before those same customers feel even a faint whiff of desire. We are both witnesses to and beneficiaries of dazzling speed-to-solution successes.

Back on planet health IT, the American College of Physicians reports that the percentage of doctors who are “very dissatisfied” with their EHRs has risen by 15 percent since 2010; in a poll, 39 percent said they would not recommend their EHR to colleagues and 38 percent said they would not buy the same system again.

I will argue that the difference between health IT and every other progressive, mature industry is the application of open source, open standards and, most importantly, open platforms. These platforms supporting interoperability and substitutability have enabled Apple and Google—and NOAA weather data, the Facebook Developer Platform, Amazon Web Services, Salesforce, Twitter, eBay, etc.—to drive innovation and competition instead of stifling it. They have created markets where everyone wins—the client, the application developer and the platform company.

The keys to open platforms are application programming interfaces (APIs) through which a platform-building company (i.e., Apple, Google) welcomes the contributions of clients and other companies. The more elegant the API, the more it can support true interoperability.

In the platform, innovators see market opportunity—an invitation to build on the open platform without time-consuming business negotiations. For the open platform provider, APIs also provide key benefits:

  • Increased value of the system and platform for clients (the right kind of lock-in)
  • Expanded solutions for the client while the company focuses on system and platform
  • More rapid and affordable innovation and customization

Most importantly, as described by Kenneth D. Mandl, M.D., M.P.H., and Isaac S. Kohane, M.D., Ph.D, in a 2009 New England Journal of Medicine article entitled “No Small Change for the Health Information Economy”, open platforms create new markets driven by innovation of substitutable choices:

The Apple iPhone, for example, uses a software platform with a published interface that allows software developers outside Apple to create applications; there are now nearly 10,000 [currently well over 300,000] applications that consumers can download and use with the common phone interface. The platform separates the system from the functionality provided by the applications.

And the applications are substitutable: a consumer can download a calendar reminder system, reject it, and then download another one. The consumer is committed to the platform, but the applications compete on value and cost.

Specifically regarding health IT, Mandel and Kohane say:

The system should be sufficiently modular and interoperable so that a primary care provider could readily use a billing system from one vendor, a prescription-writing program from another, and a laboratory information system from yet another. Individual systems do not need to perform all functions … Just as consumers may swap out applications on their iPhones, physicians should be able to readily replace one referral-management system with another … the platform should be built to open standards, accommodating both open-source and closed-source software …

Imagine, as Mandel and Kohane did, that health care organizations did not have to look to a single proprietary vendor for all solutions. What if they could go to a competitive marketplace of ever-evolving applications in a host of categories that could all plug and play, integrate and enable collaboration? The NEJM article envisions prescribing, physician order-entry, medication reconciliation and drug-safety alerts available in the Medication Management app store. Under Public Health Reporting, physicians can purchase and download plug-n-play notifiable disease reporting, bio-surveillance, and pharmacosurveillance tools. For Decision Support, laboratory-test interpretation, genomics and guideline management are available.

The creative potential for a healthcare app store is huge.

The open-platforms vision is starting to become a reality in health care. Open source code and standards support collaborative communities at Open Health Tools and the federal government’s Open Source Electronic Health Record Agent (OSEHRA). Commercialized versions of the VA’s much-lauded VistA system incorporate API-enabling open development tools like the OpenVista Interface Domain (OVID), Medical Domain Web Services (MDWS or “meadows”) and FM Projection, a tool for the MUMPS database that enables structured SQL views of system data for various uses. The open source MIRTH Connect interface engine can work with any of these open platforms to enable easier and more affordable interfacing than proprietary options.

Open platforms in health IT are inevitable. Exactly when OPEN becomes health IT’s de facto reality is impossible to determine. But we can be certain that it will happen because healthcare businesses focused on quality improvement and cost-effective care will demand it.

Edmund Billings, MD, is chief medical officer of Medsphere Systems Corporation, the developer of the OpenVista electronic health record.

40 replies »

  1. Thanks for sharing! Very useful
    Your article is from 2013, but it still has merit.
    I would like to add that the healthcare sector is using EMR and EHR software development more and more widely today. EHR is a real-life, patient-centric record with immediate and secure access for permitted users. In 2018 the global EHR software market was worth $20.4 billion and by 2023 it is expected to be $33.294 billion. you will learn about its potential in detail. Take a look!

  2. Payer IT Outsourcing Market’s valuation expected to reach $10793.7 million.

    Get customized details @ ( http://bit.ly/10FCSeo )

    The global payer IT outsourcing market is mainly driven by outsourcing standard transactional services like electronic data interchange (EDI) gateway services, production maintenance services, and trading partner management services. Thus, the payers can achieve better data quality, reduction in pending claims, and speed up the turnaround time on claims processing. The market was valued at $7833.3 million in 2013, and expected to reach $10793.7 million by 2018, at a CAGR of 6.5% from 2013 to 2018.

  3. 20% unemployment where I live. If I went concierge, I wouldn’t have any patients, and my patients wouldn’t have any doctor. Not an option when you don’t live in a city. Do only rich people deserve to get medical care?

  4. Seems like this thread has moved way off course. Back to basics: using Apple as a model of open source software seems a stretch. Publishing an API to enable apps to run on your iPhone is hardly open source. While Vista garners a lot of visibility as a model for interoperability, when you look under the covers, you discover that the VA is running 130+ instances of Vista to make it all work, and data collected at the Hines VA in Chicago (particularly images) is not necessarily available or able to be updated by the clinician working at the VA in Houston. I would argue that what we have is a fundamental architectural problem: enabling applications to “interoperate” is not necessarily solved by everyone writing “open source” software. Rather we should look to the services architecture model that basically drives the Internet as the best option for the future.

  5. You are all witnessing the depreciation of medical care via this unregulated experiment using medical devices considered to be magical by the CEO who posted recently.

    The practice of meicineis an art and science, but one thing it is not is magic.

    Shame on all you HIT magicians. Show us the evidence. Quit the hype and deception.

  6. Huge redistribution of wealth is occurring, being taken from the patients and medical system, and enriching the pockets of vendors and consultants, without providing any meaningful benefit. Government wants HIT paid for by HITECH, but skimps on covering life saving drugs and medical care.

    Whatever you propose as a solution, consider it an experiment and study outcomes prior to wasting $ billions on devices and treatment that serve as impediments to safe care.

  7. Shiva,
    Terrific comment, except the “compel” thingy 🙂

    So let’s talk shop. As you note, API integration in other sectors is between enterprise software packages, which to me implies some sort of discipline and accountability for quality assurance. This is good and this is a natural evolution from the current state of EDI, and to be fair to health care, it is not much different than current status in B2B commerce elsewhere. Actually, our “clearinghouses” are very much like the VANs in the rest of the transactional world, and in the last five years or so, they too transitioned to RESTful exchange via XML.
    So I am guardedly optimistic.

    The problem here is that in order to expose robust APIs most of health care’s software would need to undergo massive refactoring, because just like all other B2B software, it was not built from the ground up to expose any useful interfaces. Honestly, I have no idea if all this tinkering is even cost effective. To me, it looks like the big HIT houses may find it more profitable to buy out new technologies and integrate them on their own into the main product on a case by case basis. Obviously, other than not working very well (as some avid shoppers of bits and pieces discovered to the chagrin of their shareholders), this creates very brittle code and only hastens the end of life of the product.

    So we’ll see how it goes, but I am expecting major “rewrites” of the original software packages, to a more modern architecture (or at least some sort of architecture), before we see any type of serious opening of APIs. And I seriously doubt that little guys will be able to plug and play ala Google Driver. Health care data is much too “precious” and sensitive for that. Hence the hopes that all those CCD things can trickle it out through millions of patients willing to do the bidding of data-poor entrepreneurs. Maybe…. I don’t think so though.

  8. Margalit,
    You are correct that the need is for an “OPEN API” (regardless of Free or Open Source or other models).

    Open API has been used in many other verticals without the “small problem” you state because there are real enterprise applications that can leverage the API besides your “99c apps”. Enterprise applications will have their own release cycles managed independently from the layers above or below. This is nothing new… just not there in Health IT.

    And yes, these enterprise applications will need to talk to the main datastore via the open API and yes, the owner of the datastore will likely continue to be the incumbent EMR vendor.

    IMHO, we need a Meaningful Use Stage 4 requirement 🙂 to “compel” EMRs to expose a North-bound R/W API just like they are now required to have an East-bound (depending on how you look at it) CCD interface to applications across the continuum of care.

  9. You advocate for something that will never happen in our country with our legal system.

    I was involved in firing a doc in our own group – for quality reasons. Naturally, we got sued, case dragged on (with continuing legal bills) until finally a settlement was reached. And that was for someone who was employed by us in a State where people can allegedly be fired at will by an employer.

    Now, suppose the person being “run out of town” is a competitor. Lawsuits will start flying, the Justice Dept could even weigh in on “restraint of trade”, yadda, yadda, yadda. We will have a lasting peace in the Middle East before anything like that happens in our country.

    And do you think that all the various “stakeholders”, “consumer activists”, etc. want docs to regulate each other without their “input”. Again, don’t hold your breath.

    Other than the fact that your suggestion is naive and will never happen, I like it!

  10. @Lying for a Living – I agree.
    Maybe open APIs would be a better term. It is completely irrelevant to the average user if the source code is available for download, because the average user is not going to start writing code. For the non-average user, i.e. very large hospitals, modifying the source code on your own presents a nightmare of maintainability, considering that a bunch of other people and the original open source governing body are also continuing to update and develop.
    Open source is great for utilities to build products from, but has very little real value for commercial products.

    As to open APIs, there is a small problem here as well, because the envisioned apps will have to interact with the main datastore and each other in order to be useful, which is almost never the case for the multitude of iPhone 99 cents apps (at most they read a contact list or something else that’s trivial). So someone has to be the master of the database, and that master is ultimately responsible for its integrity.

    Now imagine a dozen or so interchangeable apps for eRx or CPOE, all with their own release schedules, all depending on each other’s behavior, connecting, updating and modifying that database, pulled in and out by the user, or multiple concurrent users, whenever they wish; some using one app others using another.; some using one version, others using another.
    If you haven’t fainted yet, now imagine that this entire thing is open source and users can make changes to their apps on their own, recompile and deploy as easy as pie.
    If you don’t see the problem here, I would take a wild guess that you never really worked in IT before.

  11. Where does CMS’s money go? To The Scooter Store.

    I’d love a government takeover.

  12. With great trepidation…

    For years I have been calling on THCB for physicians to reassert control of medicine. This is a bigger thing than it might sound like, for by “control” I mean “governance” — much, much more than “leave me alone, I’ll control myself, thankyouverymuch.” I mean that medicine must operate very like the old Guilds, and this means several things. As a `fer instance: all of you physicians know another doc you’d NEVER refer to because he’s dangerous. So why on earth does he still practice medicine? A Guild would have found some other way for him to put his training at the service of the public without endangering it and without depriving his family of a livelihood, but this means the Guild Masters are actually supervising him. If he does not care to be supervised, the Guild reports this fact to the Prince, who forbids him to practice medicine due to his non-membership in the Guild. It means the State Board of Medicine disappears and a private (but chartered) Guild replaces it. Every state should have at least two such Guilds — I figure Missouri (where I live) should have five. I realize all this has been abused in history. But abusus non tollit usus, as the Schoolmen say. Make it work, docs. Because what we’re doing in practice is making one very huge nationwide Guild that probably won’t work very well, and runs contrary to our national character. This I think is very bad for medicine, and frankly for the American Experiment. But it means you docs have to actually go to your local professional society meetings, shove your hands into the muck up to your armpits, and then move on to higher levels and do it again. You’re getting a Federal takeover because physicians have opened the door to it by accepting insurance payments, and have left a huge vacuum. But if your objection is to being supervised at all, you will get no sympathy here.

    Next — I lay the responsibility for lack of interoperability at the feet of hospital CIOs and boards primarily, but physicians share blame. When capital projects are considered at hospitals — buy a sexy new DaVinci Surgical Robot? Or buy new HIT from a small company that can’t support us as well as the big boy can, but hangs its hat on standards, interoperability, custom hardware, and the like? — who argues for what, and what wins? You all know the answer.

    There are other strategies CIOs could use to force the issue with the big boy vendors, but they ARE short-term expensive to implement. Same argument applies then. And what is the average tenure of a hospital CEO? Three years? Four? And why do they leave? This is a 20 year project that should have got started in earnest about fifteen years ago. Constant turnover of leadership measured on short-term results ain’t going to get you there.

    So yes, by all means use public standards and GNU Copylefted software when you can. Great. But the main problem is people’s attitudes — including physician attitudes.

  13. “Health IT is just one part of the complete take-over of our healthcare system by for-profit corporations”

    Hmmm… CMS (Medicare, Medicaid, CHIP), VA, DoD account for an increasing share of the health care $$$ — by now — what? — up to half or more? Gotta love you cats and your reflexive hyperbole. Corporate takeover? I thought the Goopers are all out there railing against a “Government Takeover”?

    Which is it?

  14. As usual, southern doc, you hit it right on the head. Thanks.

  15. Chill out, docs.

    Health IT is just one part of the complete take-over of our healthcare system by for-profit corporations that has occurred over the past 15 years. The system is about serving the needs of Wall Street, not patients or physicians. These corporations (insurers, pharma, hospital chains, IT, etc.) are just following the natural laws of unregulated capitalism. It’s too late to do anything about it.

    If you want to get angry at someone, how about the Quislings among us, the medical societies and the others among our profession who speeded up this take-over?

  16. I don’t think we want to get too hung up on the term “open source.” It still, unfairly, conjures up images of man-children furiously coding in their mothers’ basements. Mention the term to many hospital administrators and they return a blank stare followed by a barely controlled look of terror.

    It’s not that open source can’t make a very significant contribution to the development of effective health IT. It can. But Apple is so far from an open-source company, and they still provide an open platform for disparate app developers. The platform model is what we need to discuss, it seems. How do we get to that point in the development of effective health IT? How do we break the vendor lock-in model but good?

  17. “Is the vendor lock-in / secret-software business model helping either physicians or patients?”

    All caps NO.

  18. “What does physician compensation have to do with bad software products?”

    There are plenty of fine HIT products. Spare us your simplistic broad brush. The domain, though, is so routinely data-intensive that it is utterly unrealistic to expect physicians to document meticulously and in full in the time allotted by the payment system.

    Do I REALLY have to explain that to YOU?


    What do you think is just compensation for your training, skills, and experience levels?

  19. I’m for open source data and platforms which is inevitable

    Smart entrepeneurs have accepted that reality

    Dr. Rick Lippin

  20. Quack,

    Why pay for data entry, when they can “influence” you to do it for free?
    (By the way, being sent to the gulag was “re-education”, not “punishment”.)

    Please do not impugn the motives of our fearless leaders!

  21. “influence” not “coerce,” sort of like…not “murder” but “suicide with assistance”

    What does physician compensation have to do with bad software products?

    I know the EMRs help you guys collect all kinds of data for public health. That’s great. Go forth. I just want them out of the way of my medical practice. They are slow, they take away time spent with patients. They are cumbersome to use. I’m not opposed to you collecting all kinds of data to crunch. I’m just not the one to be collecting it and entering it. I’m here to see patients and fix what’s wrong. What you need are data entry people, not doctors.

  22. “influence”, not “coerce”.

    You are free to go totally paper concierge. Some are doing so.

    Look, I routinely bite the hand that feeds me. I think physicians should be paid more like lawyers (better than lawyers, actually), in particular with high hourly rates that expert cognitive skills ought get. dx is way undervalued relative to px. That is certainly not of MY making.

    btw, I’m a statistician, not a Health IT geek.


    Nothing would make me happier.

  23. What, no smart a** remarks about being compelled to use EHRs?

    Does that mean you agree that fines are an attempt to coerce use?

  24. Ahhhh… yes, Victimhood. Overpopulates TCHB comments. I have to buy my Kleenex at Costco just to keep up.

  25. BobbyG,

    And of course you are not the angry one. You are the beneficiary of others being forced to use poorly written software, not the victim.

  26. BobbyG,

    And when the penalties for not having an EHR kick in, you don’t think that is an attempt to compel use?.

  27. You’re the one crying. I merely mock you. Properly so. I’m not the angry one.

    Go concierge. Show your Stones.

  28. Testy, testy. EMR ppl get so angry when we call them out on their scam.

  29. Quit. Do something else. Or step up and go Concierge. Quit whining. You are not “compelled” to take 3rd party reimbursement. Be a Bob Lambert instead of a Drama King.

  30. If you think that purchasing those things, implementing into your workflow, and then decreasing the volume of patients seen is “nothing” then I have a bridge to see you in Brooklyn…

    Yes, compel. As in not accepting bills by any means other than electronic. By requiring “quality” measures, the data on which can only be gathered electronic.

    When they wrote HIPAA, the software companies bribed the Clintons into including mandates for their fecal products. Instead of improving and lowering the price on said products, they bought a government mandate for their purchase by their target market.

  31. We physicians were not duped. We wouldn’t buy, so the software companies got Congress to compel us to use their fecal products.

  32. Excellent and important post.
    Only we physicians would be near-sighted enough to be duped by the EHR industry taking advantage by sliding open source right by us.

  33. Thank you for this brilliant article! It states a large and costly problem within the healthcare system today. Just as the medical field treats patient problems as they arise, rather than treat the holistic problem, so do healthcare institutions buy their IT products. When a problem arises or a data piece is needed, the existing vendor is not always flexible enough to meet the need. So, another vendor is sought and often these two products don’t speak to each other. I like the idea of open source, but are these platforms HIPPA approved?