Uncategorized

Clinical Groupware: Platforms, Not Software

Kibbe

Clinical Groupware is rapidly gaining acceptance as a term describing a new class of affordable, ergonomic, and Web-based care management tools. Since David first articulated Clinical Groupware's conceptual framework on this blog early last year — see here and here — we've been discussing Clinical Groupware with a growing number of people and organizations who want to know what it is, where it's going, and what problems it may solve, particularly for small and medium size medical practices, their patients and their institutional/corporate sponsors and networks.

Clinical Groupware heralds a shift away from medical applications that are primarily based in local hardware and software. It creates a more fluid functionality in those applications, and empowers communications as well, by leveraging Internet connectivity, Web-based data resources, and new services (i.e., capabilities) performed upon these data by agents or applications.

In other words, Clinical Groupware is about platforms that can integrate modular applications, which in turn are supported by subsystems of data services. Although it is still in its infancy, Clinical Groupware is an end-to-end digital revolution in health IT.

It is still too early for a single best example of Clinical Groupware to have emerged. The creation of platforms, modules, and data services in health care has begun only recently, fueled by and borrowing from developments in popular computing that include search, social networking, geo-location, identity management, photo and music-sharing protocols, and remote storage.

Clinical Groupware is sometimes understood in terms of "remote hosting" or an "application service model" (ASP) of software. It is true that this might be a starting point for some users. But as a phenomenon, it is far more powerful than simply running a software program over the Internet instead of on your computer or local area network.

Tim O'Reilly uses "Internet as operating system" as a short-hand way of describing the robust complexity of features and functions available to users of today's browser-based and mobile computing platforms. This approach contrasts markedly with the older client-server computing model. In client-server arrangements, a computer-resident operating system coordinates access to applications and machine resources on a single or, at most, a few computers on a network. In the "Internet as OS" model, the Internet itself coordinates that access across large numbers of computers and users. 

The browser or the smart phone may be the means of gaining access to this new and rich "compu-cology," to coin a term. But what really matters most of the time is what is happening between your interfacing device and the many applications on the net that it can reach.

Consider the difference between the mere delivery of an application, such as an ePrescribing software program, over the Internet, versus the richness and complexity of two very popular, although very different computing platforms, Google apps and the iPhone with its app store.

Google's core competency is, of course, its search technology, which almost instantaneously takes the search string from your browser or mobile phone and serves it up to Google's proprietary software at one or more of its massive server farms.  But Google also offers free (or very inexpensive) applications such as calendaring, email, photo organizing and sharing, word processing and presentations, mapping, etc. most of which are capable of sharing, indexing, and processing several different types of information in the background in a connected manner. Thus, at the push of a button while in Picasa Web Album, Google's online photo storing/organizing application, one can publish individual photos, or whole albums, to groups of people in one's Gmail account, while also allowing those people to upload new photos to some albums, but not others. It is also easy to place photos on a map location, view both photos and maps in Google Earth, and then share these with others. In each case there are complex data look-ups and indexing occurring, mediated by Internet protocols for identity management and access permissions, in the background.

The iPhone is a more proprietary platform – a "walled garden" in the jargon of the day – that integrates multiple data processing activities, some of which are hardware resident and others that occur online. Its wireless capability supports access to the Web, which can integrate with the built-in GPS location services that are in communication with satellites circling the earth. This arrangement can tap into a world-wide technical infrastructure that can help you find the nearest Chinese food restaurant or get to a nearby hospital trauma center. It can allow you to search for a doctor, map the location of the doctor's office, and get performance ratings on that physician's or organization's quality and service. Many different applications "run on" the iPhone device, but they depend on what O'Reilly calls "network available services" for value creation that far exceeds the features of the phone itself.

In other words, these new Web-based platforms allow distinct functions to interact with and leverage one another, creating a robustness of capability and productivity that was unthinkable in earlier, more limited hosted arrangements. Thinking of these platforms as merely running remotely-hosted applications is to miss their possibilities. Clinical Groupware, a very powerful and practical medical application of this model, is the revolution ahead that will foster intense competition among vendors vying for platform real estate.

Still skeptical? In fact, the leadership at ONC/HHS have already realized that the future of Health IT lies in a whole that is greater than the sum of its parts.

In early April, ONC awarded $60 million to four institutions – Mayo Clinic, Harvard University, University of Texas Health Science Center at Houston and University of Illinois at Urbana-Champaign – through the Strategic Health IT Advanced Research Projects (SHARP) program. Each institution's research projects will identify short- and long-term solutions to address key challenges associated with health IT and meaningful use. John Halamka recently blogged about the Harvard research, which will "investigate, evaluate, and prototype approaches to achieving an “iPhone-like” health information technology platform model, as was first described by Mandl and Kohane in a March 2009 Perspectives article in The New England Journal of Medicine." Further, Halamka writes:

The platform architecture, described as a “SMArt” (Substitutable Medical Applications, reusable technologies) architecture, will provide core services and support extensively networked data from across the health system, as well as facilitate substitutable applications – enabling the equivalent of the iTunes App Store for health.

This new approach to a health information infrastructure was the focus of a June 2009 working group meeting at the Harvard Medical School Center for Biomedical Informatics and an October HIT meeting which brought together more than 100 key stakeholders across academia, government and industry in an exploration of innovative ways to transform the national health IT system.

One of the challenges facing the Clinical Groupware, modular application approach, is that of data exchange between apps and data integration among several different apps.  It is clear that the Harvard SHARP research grant will have these problems as high priorities for solutions during 2010 and 2011.

David C. Kibbe, MD, MBA and Brian Klepper, PhD write together on health care technology, market dynamics and innovation.

Categories: Uncategorized

32 replies »

  1. Hi very cool web-site! Person. Outstanding. Astounding. I’ll bookmark your web site and grab the feeds furthermore? My business is delighted to uncover many helpful data throughout your publish, we start to use figure out a lot more approaches to this specific consider, thank you spreading.

  2. “Web-based data resources” and software applications are definitely the wave of the future. But as “Joseph Stevens, MD” there still will not be an electronic substitute for the well trained and prepared medical professional. We need to keep our eye on the ball and ensure that the health, safety and welfare of a patient comes first even as we pursue efficiencies and new technologies.

  3. Nice comprehensive list! I agree that every software application intended for human consumption is built based on the Model-View-Controller (MVC) paradigm. if religiously observed this gives a separation of interests where one could substitute each part of the MVC for another. The basic theme here is that if we want data to do research on, learn from and measure quality with, somehow, somebody, at some point, is going to have to get that data in there.
    You can also look for series provided by Simulation Medical to get more insights on to the topic. Intramuscular and subcutaneous injections are the most common techniques for therapeutic injection.  This series will provide everything an organization needs to demonstrate these essential skills with their products.
    Simulation Medical Demonstration Kits have been designed to help companies showcase new technologies. 

  4. Are we looking for a solution that’s only specific to Health IT? Or is it just the same 0s and 1s that just need a set of HC specific restrictions?
    Your article mentions the SHARP initiative. Although a transactional ‘framework’ specific to the new ONC rulings/policies doesn’t specifically exist, an ONC explanation of why research institutes were the right choice over existing data transactional companies / projects would shed some light as to the direction of what such a framework is to become.
    Development concepts like RAD (Rapid Application Development), mysql, mongodb and Margalit Gur-Arie’s mention of MVC are just a trickle of examples on how the rest of the world invented frameworks and data-driven models.
    A look outside HIT might do us all good. You want data integrity, why not look at financial IT? You want scale, take a look at what’s happening with opensource mongodb – able to manage millions of transactions a minute.
    Whichever model(s) SHARP helps to develop, let’s hope it borrows from existing models that have already pounded the pavement.

  5. James;
    If I am understanding structured narrative correctly, it is sort of a template in which one dictates to fill in the blanks? We used to use this in our gross pathology room for measurements of the specimen and other standard items. This has some possibilities, but what free text capabilites are there?

  6. Bev: “But I am talking about truly clunky ways of doing things which take twice as long as the “old” way, and essentially convert the doc into a secretary. For example, could dictation be incorporated as a way to enter narrative, or something similar.”
    I covered a presentation on this topic at a recent local HIMSS Chapter event. Technology based upon EHR structured text tools was being discussed. The presenter was from http://www.MD-IT.com. A link to the presenter bio and an overview of the presentation is here: http://www.austinhimss.org/

  7. This vendor agnostic platform approach already exists and is in the early stages of implementation. Healthcare Interactive has developed a platform that allows all of the stake holders to communicate aggregate and provide third party applications to the consumers and providers of healthcare. All this is done securely and privately. Information is easily accessible by all parties based on system level permissions.
    Communications, Organization, Efficiency, and Management Capabilities all in one platform.

  8. “The old adage regarding real estate is that the top three value considerations are location, location and location. As healthcare transitions into more value-based payment designs, the same adage will hold true for health information technology.” – i really like this comment. I support this point of view.

  9. Agree on the general direction of the original post. Office based PC’s provide an interim solution to start collecting the data but they are way too limited to provide the benefits that are possible.
    I have just finished reading Safe Patients, Smart Hospitals by Peter Pronovost, M.D. In 10 years a child’s needless death from a center line catheter infection is leading to culture and data changes that are reducing costs and saving lives. The changes are reaching into more and more areas of medicine and more and more areas of geography including nationwide and international implementations.
    Among other issues, health care providers resisted changes because they were concerned that data would be used against them in court. We are moving to a point where providers that do not capture and use data will find that being used against them in court. I can hear a plaintiff’s attorney saying to the jury, “Can you believe any doctor could be so calloused as to ignore the life saying benefits of good data and changes in ways medicine is being practiced? This kind of arrogance cries out for not only compensatory damages but punitive damages as well to send a message to doctors everywhere!”
    Data is becoming more important and we will learn how to capture it easier and use it better. When we didn’t have it, we didn’t miss it. Now that we can begin to see what will be, we want it now! That’s part of the on-going story of technology.
    Thank big!

  10. “my concern is that for that to happen, data has to be in the computer first”
    This of course is the key regarding digitization of clinical data. Once the data are routinely and more or less universally digitized at “point of sale”, the endless concerns over how they are manipulated and presented can be easily dealt with. More or less no-brainerdom in regard to those.
    The hard part is getting providers in particular to get on with the program – a task apparently that will take minimally a generational change, so 15 years into the future.
    There is nothing inherently wrong with so-called legacy systems that cannot be fixed with some fresh paint applied to them, then gradually replacing them over time.

  11. Bev,
    I love cranky customers. You learn nothing from the happy ones, as few as they are…
    I do agree with you and all the docs out there, that the current hospital systems need a lot of improvement, and the ambulatory ones too, but maybe a bit less because it is easier to change ambulatory systems, so I think they are a bit ahead of the hospital mammoths.
    The heart of the issue though, is that if we want data to do research on, learn from and measure quality with, somehow, somebody, at some point, is going to have to get that data in there. There is no other way.
    So somebody is bound to feel like a secretary, and right now it seems that it is the physician.
    All that said, I am sure we can improve on user interfaces and remove all the billing related documentation, assuming of course that docs don’t want to get paid or that the payers and government don’t require documentation for payment any more, and add all sorts of other technological wizardry, but at the end of the day, the entire thing will still require data entry.
    This is why my biggest hopes are tied to innovation in this unlikely and humble area – data entry.

  12. Margalit;
    I sense you all think I am just one of those nay-sayer doctors who doesn’t want change. And I know there are plenty of those out there, many of them about my age. (I wonder if you get the same comments from the younger docs.) I just think you have to be careful, as I did when receiving constant criticism of our lab systems from clinicians, not to let the identity of the messenger obviate the usefulness of the message. Maybe if we’re all saying the same thing, there really is a problem!
    For instance, we found that the lab system was great for incorporating narrative type information like microbiology reports and pathology reports – which are far different than the simple numbers in chemistry/hematology, etc. But when interfaced to the hospital information system,most the time we got either garbage or important things truncated or left out. We had to teach the IT vendors why it was important that their systems be able to handle this! And then we got the “phrase of death” – “We’ve NEVER had this problem with another client.” If I had a nickel for every time I heard that in 21 years…..
    As for dictation, when I retired 5 years ago they were just experimenting with dictation systems for gross pathology and autopsies, and those still had to be carefully edited for many errors. (Foreign accents notwithstanding….) MD as HELL, for all his ranting, I believe is correct when he complains the ER computer has turned him into a secretary -I bet he doesn’t have dictation capability. What is available out there and what is actually installed may be vastly different.
    I am just saying, don’t stop listening to your customers just because they are cranky. There are real reasons for this crankiness, and they need fixing. The pace of progress is far too slow, especially with all the new $$ coming in and the high expectations in a short time frame. Patients’ lives are depending on this!

  13. Bev, most puddles in the desert are mirages… just saying 🙂
    What do you mean by THEIR way?
    For example, a long time ago I had to learn calligraphy and when I wanted to make text bold I used to put more pressure on the pen at a certain angle. Now, I have to highlight the text and click on the B icon. Not to mention, that I only have predefined fonts now, so my personal flourishes are impossible. Eventually, I got used to it.
    I think you know where I’m going here….
    That’s not to defend the usability of some systems, which is horrifically behind the times. It’s just that if we are going to use computers, it will have to be the computer way, and computers are ill suited to personal flourishes.
    Just saw your second comment, and yes the prices are obscene, particularly for hospital systems. For private practice there are much cheaper alternatives.
    I don’t know when you last looked at one of these things, but practically all of them incorporate dictation, some have scribbling like on a sticky note, and all have plain old typing.
    Just to be clear, I am not defending the status quo. We have a long way to go, but I think we are moving in the right direction, even if it’s a bit slower than we would all like,

  14. I don’t know about one or two ways of doing things, and this may of course vary by each hospital or office. But on achieving consensus, you speak the truth; we had the same problem when asking them how to configure lab reports on our system. It takes a truly skillful facilitator to keep them on subject and achieve some agreement.
    But I am talking about truly clunky ways of doing things which take twice as long as the “old” way, and essentially convert the doc into a secretary. For example, could dictation be incorporated as a way to enter narrative, or something similar.
    Part of this is a generational thing (younger docs are more comfortable typing and using the system, etc.); part is stupid stuff like, are there enough terminals on the nursing station or laptops in the halls,and a large part is vendors not trying to serve the customer.
    I believe the new approaches delineated above will help; these huge, 40-60 million dollar legacy systems which don’t implement well are really a problem……and yes,I know part of it is the customer, etc., but the fact is, if they don’t work well, they are not providing value for the huge expense. This is what hospitals and docs’ offices are grumbling about right now. They only see more big $$$$ down the drain. The vendors really, really, REALLY need to provide a different product.

  15. Bev, a computer system that’s going to add much value has to have one or two ways of doing something, and then you’re going to have to do things its way. One hopes that its way is a better way, and that’s where docs (and nurses, and pharmacists, et. al.) come in. The trouble (from my own POV) is that getting five docs in a room and looking for one or two better ways results in about ten mutually contradictory ideas. `tis a quandry.
    t

  16. Dr. Kibbe’s comment about the incumbents trying to minimize the importance or usefulness of innovation in health IT products made me laugh out loud. My only knowledge of health IT products is as a heavy user and observer of most other users (who always had helpful, if pointed, criticisms of our lab information systems). I can say categorically that most docs hate these legacy systems so badly that any innovation which makes their interaction with the computer easier will be jumped on like a man finding a puddle in the desert. We are very, very tired of systems which make us do things THEIR way instead of helping us do things a better way.

  17. sony servis

    nice discovery, and very informative post.. thanks to share.. it will help the novice bloggers a lot… keep posting this type of stuffs.

  18. Thanks for responding, David.
    I think Dr. Walker’s comment made me realize that we really don’t disagree much. We are just looking at the problem from different perspectives. I tend to view the big, enterprise software required to run the office, or the hospital, as the starting point and the layers of interoperability as secondary to that software.
    I may be wrong, but I think you are approaching the same problem with interoperability as the first concern. Granted, interoperability is now the issue on the table and, yes, it is possible to achieve communications through small modules that may very well be plug & play. When I look at it in this way, I do agree that some sort of disruptive innovation, be it in software design, or in business model, could open the floodgates and allow data to move freely.
    However, my concern is that for that to happen, data has to be in the computer first, which points me again to the basic problem that doctors are averse to getting data into the computer, and if we are ever going to exchange data, the enterprise application needs to drastically improve. Here I am half expecting a very disruptive innovation to allow physicians to capture data painlessly, accurately and in a clinical meaningful way.
    As to disruption, I am sure the disruptor carefully planned their entrance into the market and the act of disruption was anticipated, at least to a certain degree. As Dr. Walker noted, most disruptions do not replace the incumbents, as much as causes them to adapt to a new way of doing business. Those that don’t adapt are the ones who die. So I still have high hopes from the EHR incumbents and the fact that there seems to be a trend to open platforms and expose APIs. This sort of thing would be unheard of only a couple of years ago.
    Maybe the disruption is happening as we speak……

  19. Margalit and David, you are both right – some apps lend themselves to cheap iPhone-store-like functionality, while that model fails for the larger apps, such as the basic EMR database manager. But Margalit makes a very powerful point – one that could make clinical groupware seamless with current legacy constructs in HIT. We saw this when the nascent Internet evolved from bundled suites of functionality (AOL and Netscape) to search engines and direct server-to-server connectivity without the middle man. To me, this provides a path forward for our current juncture in HIT – pursuing widespread emr adoption (creating the market) via the AOLs of EHR (Allscripts, NextGen, etc) while the policy, industry, and research sectors continue the push towards a healthcare Internet and modular framework – as standards begin to emerge. The legacy vendors should not fear the transition, but rather adapt and find other ways to provide value to their consumers (lots of folks still use AOL email services; I myself continue to pay gor yahoo mail services).

  20. Another point, Margalit: disruptive innovation is not “unexpected” or “unplanned.” In fact, just the opposite. Consider the likes of Southwest Airlines, the PC, Google search, MinuteClinics, the digital camera, Linux, mini-mills for steel, and, of course, the transistor radio. These disruptive innovations were VERY carefully planned and, at least to their proponents and champions, very much expected to succeed. According to Clayton Christensen, the Harvard Business School prof who coined the term, and has written about it for almost a generation, careful design is a critical element of disruptive innovations in every case.
    I think what you’re saying, and I’ll acknowledge, too, is that the incumbents in any market facing potential disruption by innovative products or services tend to deny the reality of the challenge, often long past the point where their own adaptation to the change is possible. And this, too: if they recognize the challenge, they will often disparage its importance or significance. They’ll try to convince the market, including the financial market, that the new low cost, easy to use, less complex product or service is “like a toy” or is “lousy quality” compared to their own products and services.
    But don’t confuse disruptive innovation with “chance” or poor design or magical appearance. The closer one looks at successful disruptive innovations, the larger appears the excellence of business and technical design/planning that has gone into them.
    It is only an artifact of our own short-sightedness that they appear to occur unexpectedly or suddenly.
    Kind regards, DCK

  21. Randall: To your point, yesterday I received an inquiry from a hospitalist (a physician working full time in the hospital, and caring for many patients who don’t have doctors, or whose physicians don’t do inpatient work), that asked: is there a software product out there that works kind of like Twitter? You know, that will let the hospitalists on duty convey short messages to one another, and communicate with colleagues outside the hospital, too?”
    What goes without saying is that none of the current inpatient HIT vendors offer a product that does this kind of social networking, care coordination and management. But the doctors are starting to understand the potential value to them and their patients, and to ask for it.
    Margalit: Enjoyed your (long) comment. I have to disagree with you somewhat, however. Yes, some transactional apps might be complex, and difficult to see being implemented in the $1.99 iPhone model. But others might work that way just fine. Suppose I could send a modular app a CCR xml file and run a diagnostic algorithm or a preventive maintenance guidelines agains the data. Might this be a “plug-in” that would be worth a few pennies for each file thus queried and returned? Or what if my practice were able to add on a modular app that utilized the data from other apps, and in this case helped the practice manage inventories of supplies, including vaccines and items that have limited shelf life? Why not use a modular platform to map my patient population, or a population of patients with a particular set of conditions? Or a plug-in that would support managing free drug samples and at the same time schedule drug representative visit?
    I think it is very difficult to consider what value added modular apps might looks like, as we are so wedded to the current practice workflows and habits. I well remember when most of my friends looked at me askance and said “are you crazy? no one will want to buy a book from a computer!!”
    Kind regards, dCK

  22. “In fact the Government would be creating an open source EHR and allow companies to use the platform to build proprietary products on top of it (like Apache).”
    To paraphrase from above, I would like to see this

  23. Every software application intended for human consumption is built based on the Model-View-Controller (MVC) paradigm. The Model is the data layer; the Controller is the business layer, or the logic and processing; the View is the presentation layer, or what the user sees and interacts with. This paradigm, if religiously observed, provides a separation of interests where one could substitute each part of the MVC for another. For example, one could replace the database layer (the Model) without affecting either the View or the Controller. More commonly one can substitute the presentation layer (the View) from say a browser to a smart client or a mobile phone. One could have multiple Views for the same Model/Controller combination. In layman terms the View is the Client and the Controller/Model combination is the Server. Therefore every user accessible software is really a client/server application.
    Web browsers are Clients. They reside on the user’s machine, just like other downloadable clients, and allow the users to connect through the Internet to the Server. Other clients serve the same purpose. For example, I can use the web browser client to access Twitter, or I can use TweetDeck, which is a Client software that I have to download and install on my local machine. The Twitter Controller and Model are unchanged; the View however is different.
    Basically, browsers are universal Clients that can be used to expose various applications. It makes it somewhat easier for the user, since they already have a browser on their device, but many popular web based applications need a proprietary client. Skype is one example and everything that you buy from the AppStore and needs connectivity to work, is another example.
    So maybe we are not discarding the client/server model, but instead we are saying that the communications between client and server must be conducted over the Internet. Well, not quite. What we are saying is that there should be an option to communicate over the Internet. In fact a “legacy” client/server application could easily add another View that accesses the Contoller over the Internet. Whether the new View is a browser or a proprietary, downloadable, client is irrelevant. As long as the geographically dispersed group can access the underlying server over the Internet, this must be Groupware, I would think. It may even be simpler than that. It may not even require new Views, but instead the exposure of Controller functionality as Internet reachable services, that could be discovered and used by other software applications – Web Services has been around for quite some time, but never really delivered on its promise.
    As to substitutability of components, while technically possible today through Web Services, for some reason it is not happening in any industry. The Google apps example is not an example of a platform where components can be substituted. I cannot use Yahoo mail with Google Buzz. Google Apps is an example of a proprietary vendor adding more and more services to its base product which happens to be browser based. There are examples of platform owners that allow independent developers to build modules, like salesforce.com and Websphere Portal Server (WPS), but those modules are forever tied to the platform owner and cannot be used independently. Anything else would constitute suicide as far as the platform vendor is concerned.
    Now, I can see the Government funding the development of an open source platform and its underlying data layer (Model) and basic business services (Controller), while allowing independent developers to create various modules on top, modules that would include a View and new business services (partial Controllers). Such platform will require strict governance if data integrity and service reliability are to be preserved. In fact the Government would be creating an open source EHR and allowing developers to contribute functionality, or depending on the license model, allow companies to use the platform to build proprietary products on top of it (like Apache). I don’t think the latter is what people have in mind. If the former is the proposed model, you may get a bunch of selfless developers to contribute, but it is not financially feasible for a software company to develop, say, an electronic prescribing module to only run on the Government platform, unless of course that platform is the only allowed game in town. Besides, there is a vast difference between developing clinical and transactional modules, and developing a $0.99 iPhone application. And as I said many times before, platform governance is an ongoing significant expense. Unless the Government is prepared to fund it indefinitely, someone else will, and it will evolve into a private corporation just like the many others who built their fortunes by providing governance and support to particular versions of open source platforms. How, then, is this different than any other EHR vendor, like Eclipsys for example, that is opening up their platform to third party developers, or PatientOS (open source), or even Medsphere (VistA)?
    I do agree that we are due for a bit of disruptive innovation in the HIT space, but I sort of think that it will blast into the market unexpectedly, unplanned and unpredicted from an unlikely source. It wouldn’t really be disruptive innovation otherwise.

  24. The old adage regarding real estate is that the top three value considerations are location, location and location. As healthcare transitions into more value-based payment designs, the same adage will hold true for health information technology. Information capabilities only residing in a siloed system within a medical facility will not only place that facility off any main roads, but will likely marginalize it to the boondocks.