Thursday, July 27, 2006

Bill, have you lost your mind?

In perhaps the most bizarre story about Healthcare IT I've seen in some time, this story appeared in the New York Times:

July 26, 2006
Microsoft to Offer Health Care Software

Microsoft plans to offer software tailored for the health care industry, a change from its usual strategy of encouraging others to create industry-specific products using its operating system and programming tools.

One reason I consider this bizarre is in the apparent "superhuman leap of information technology hubris" I see in the details, by a technology company with core expertise in business-infrastructure software believing it is a "small step for mankind" to move to the bedside. This does not augur well for hospitals' limited resources and capital for clinical IT misadventures:

The company’s first step, announced today is to purchase clinical health care software developed by doctors and researchers at a nonprofit hospital in Washington. Microsoft is also hiring two of the three doctors who created the software system and 40 members of the development team at Washington Hospital Center. ... “This represents a change in our strategy,” said Peter Neupert, Microsoft’s vice president for health strategy. “This is the start for Microsoft. We’re just getting started.”

This is certainly a peculiar way to start. It appears what Microsoft has purchased is a fancy interface engine/display system developed in-house at one medical center, most likely without thought of portability, scaleability, and other critical enterprise issues, in a plan to compete with vendors of similar, field-proven products from health IT vendors with decades of experience. (As an aside, the fact that Microsoft has hired the whole development team raised the question in my mind as to if that team fled, considering my disappointment in the informatics strategic thinking, abilities and insights of former colleagues who are now at the Washington Hospital Center). In any case, how could Microsoft have gone down this route as a "start" in the clinical IT market? Read on:

The Microsoft model in the past has been to supply operating systems, database software and programming tools that outside companies use to make applications for specific industries. The idea is that Microsoft provides the underlying technology platform, but then industry partners build the applications for industries like health care and banking.

Mr. Neupert, 50, is leading Microsoft’s new strategy in health care. In 1998, after 11 years at Microsoft, he left to become chief executive of, an online retailer of pharmacy and health products. From 2003 to 2005, Mr. Neupert ... was co-chairman of a health technology subcommittee that published a report called “Revolutionizing Health Care Through Information Technology.”

Mr. Neupert returned to Microsoft last September, after persuading Steven A. Ballmer, the chief executive, and Craig Mundie, a senior strategy executive, that Microsoft should be doing more in the health care sector.

“I’ve had an opportunity to see how messed up the health care system was,” Mr. Neupert explained. “And to really have an impact, you need a footprint like Microsoft’s.”

(A "footprint like Microsoft's?" Does the "Bull in a China Shop" metaphor seem appropriate here? )

This story in many respects reflects familiar patterns. Notable is that strategy leadership in this initiative is apparently by a person whose healthcare background seems limited to having been CEO of I've also heard the promise of "Revolutionizing healthcare" in a Microsoft function before. From my essay "On Medical Informatics, MIS, and Leadership of Clinical Computing" written in 1998-2000, here is an observation I made in the mid 1990's:

... A letter I had written, " Broken Chord " (third story down) published in the journal Healthcare Informatics , further amplifies the skills and experience issue. The letter addressed issues of skills, insights, and roles of MIS personnel in healthcare settings. I described a Microsoft Healthcare Users Group conference attended predominantly by healthcare MIS staff and vendors, where I observed a panel discussion moderated by the (former) CEO of HBOC and composed of several other industry CEO's. The panel was discussing how they would revolutionize healthcare through their leadership in information technology.

During the Q&A period I asked the audience how many really felt they would revolutionize medicine through their leadership in IT. Several hundred--almost all in the audience--raised their hands. I then asked how many had ever taken care of patients or examined any textbook of medicine, such as Harrison's Textbook of Internal Medicine or the Merck Manual. A minority of hands went up. This suggested, in my view, a striking deficiency of knowledge, experience and insight on the part of the de facto clinical information technology leadership, complicated by cavalier attitudes regarding the essential role of clinical expertise.

The pretension of non-medical computer support personnel in offering profound thoughts on complicated medical matters that they understand, at best, at the level of popularizations, is troublesome. Often, CIO writings in journals of healthcare computing on subjects such as Electronic Medical Records remind me of this. A metaphor to describe the results might be that of a submariner presenting "expert" views on flying a B-52 jet plane. While both are military transporation, in technology specifics do matter.

More from the New York Times story:

The software system Microsoft is buying is called Azyxxi (pronounced ah-zik-see). It is designed to retrieve and quickly display patient information from many sources, including scanned documents, E.K.G.’s, X-rays, M.R.I. scans, angiograms and ultrasound images. It was first used in Washington Hospital Center’s emergency department in 1996, and has since been adopted at six other hospitals, including the Georgetown University Hospital, in the MedStar Health group, a nonprofit network in the Baltimore-Washington region.

The Azyxxi software, Mr. Neupert said, will be “our foundation,” adding, “You’ll find us expanding to a suite of health care solutions.”

This describes a fancy interface engine, a technology used to integrate legacy and disparate information systems that we were using in the late 1990's at Christiana Care, for example, in concert with a commerical Clinical Data Repository (Cerner OCF) to do likewise. I'm not sure what the major innovation here is, considering this system is proprietary to Medstar Health.

Additionally, adopting a proprietary interface engine as "the foundation" of clinical IT offerings iteself is a strategy whose logic escapes me. Generally, an EMR or CPOE is a foundational application in health IT. "Quick and dirty" might better describe this particular strategy outlined in the New York Times.

I'm also uncertain why Microsoft feels their name is going to inspire confidence in clinicians, when those same clinicians have had to deal with MS Windows with all its bugs and faults over the years in the home, and are aware of its serious and constant security issues. I would view the Microsoft imprimatur as a weakness in this industry sector.

More from the article:

The Microsoft plan, analysts say, could be risky. The software Microsoft is purchasing — the price was not disclosed — is a homegrown system that has not been used outside a few hospitals. It has no installed base of customers, and there are already several established suppliers of clinical information technology systems, including Cerner, Epic, G.E., Eclypsis and others.


Most of the big health care software suppliers, analysts point out, are also big customers for Microsoft operating systems, databases and programming tools. “This puts Microsoft in the uncomfortable position of potentially competing against its major customers,” said Dr. Thomas Handler, a health technology analyst at Gartner.

Tom, a former colleague and postdoc in the Yale Center for Medical Informatics, is putting this mildly. I'd say Microsoft is going to really anger some of its major customers, again a puzzling strategy.

At the Washington Hospital Center, the system [interface engine] has done its job. In 1995, before the system was introduced, the emergency ward handled 37,000 patients a year, waits stretched up to nine hours, and there seemed to be an urgent need for more doctors and rooms, Dr. Feied recalled. Today, the emergency department handles nearly 80,000 patients a year and 70 percent of them are diagnosed, treated or admitted in three hours or less. The staff has increased only 5 percent, and few rooms were added.

(As another aside, I do recall the CIO of Medstar calling me on the phone at Christiana Care circa 1997, in a panic, needing to bring his staff and several executives to Christiana Care on an "urgent basis" to review our clinical information systems, especially the clinical data repository. Apparently he had gone on a similar site visit to Texas, but that visit had gone poorly. I rapidly put together a site visit program and gave a good demo to my colleagues from Medstar. As thanks for my efforts that CIO flatly told me "there was not work for me "in his shop several years later I was looking for new opportunities. Ah, gratitude.)

“This is extremely interesting as a signal of Microsoft’s intentions in health care, but we’ll have to see what comes next and how this plays out,” said Dr. Blackford Middleton, an assistant professor at the Harvard Medical School and a health technology expert at Partners Healthcare, a nonprofit medical group that includes Massachusetts General Hospital in Boston.

Blackford, a former senior executive at a company that made an excellent EMR product, Logician, knows well the complexities of this industry. Microsoft, on the other hand, appears to believe that clinical computing and business computing are similar, and that expertise in one makes a company or executive an expert in the other. Bad assumption.

In addition, in the article the Microsoft executive stated his belief that "to really have an impact [in healthcare] you need a footprint like Microsoft's." I disagree. It's not size, it's expertise that counts. A lack of experience and probably even familiarity with social informatics is one major problem here. Sociotechnical issues, not technology, are probably the most important cause of health IT failure such as this recent case " Hospital Vexed by new computer system ", ironically at clinical IT pioneer El Camino hospital. Where, exactly, is Microsoft's expertise in this area (not to mention that these issues often seem ignored by the "experienced' vendors at their peril as the Vexed story illustrates)?

In summary, I think a significant amount of stockholder money and hospital capital, already in short supply, is about to be consumed in this "grand experiment" of leaping from software infrastructure supplier to the curing of all of medicine's ills at the bedside.

Coining a colorful expression for this phenomenon, the Syndrome of Revolutionizing Healthcare Through Ignorance and Hubris™ comes to mind.

Note: also see the posting "Anatomy of Healthcare IT Failure" on an actual, current story of health IT difficulty that occurs even with the involvement of the experienced vendors.

-- SS


Anonymous said...

It's interesting that you make assumptions without having all of the facts.

First, you admit that the system is, in your own words, "most likely without thought of portability, scalability. . ." or that other cricical enterprise issues have not been taken into consideration. Do your research.

Second, you assume that the stockholder money and hospital capital will be be in short supply, again without any knowledge as to how it will be introduced.

What if, when introduced, the system more than paid for itself in increased revenues to the hospital, as well as the number of lives saved, and improved quality of life for the masses?

Perhaps before taking the "glass half empty" viewpoint, you should get more information.

InformaticsMD said...

iIt seems the anonymous poster above is the one speculating. As a seasoned professional in healthcare informatics, I write from experience.

> What if, when introduced, the system more than paid for itself in increased revenues to the hospital, as well as the number of lives saved, and improved quality of life for the masses?

And what if,as pointed out by people such as Ross Koppel and by sites such as "bad Informatics Kills", it hurts patients once implemented in some other setting? Too bad for them?

InformaticsMD said...

The first poster also writes: "... you assume that the stockholder money and hospital capital will be be in short supply, again without any knowledge as to how it will be introduced."

Yes, I guess hospitals are just rolling in the dough and have plenty for IT experimentation.

Matthew Holt said...

Interesting post--MS is running out of rope on its core products, so feels it has to go vertical (a la Oracle)--and health care is just too big to ignore.

The biggest problem in health care IT is interfacing (perhaps I should say the biggest theoretically solvable problem). And presumably if you can drop in an interface engine (especially a really cool one--note that Sun has partnered with a bunch) into a hospital then you don't have to take out the core Clinical departmental systems that are the HIT vendors bread and butter...and of which they all try to remove their competitors products.

So my guess is that MS is going to try to leverage this new product into a consulting service play, in which they come in and fix the interface problem (or at least say they will!).

But basically they're a big monopoly utility which is out of really big ideas, and so are looking around the edges.

Will Oliver said...

Great post. It is a little strange that they are starting to go verticle, but they'll go where the money is. With the recent financial projections in HIT, government/lobby push, and the lack of quality products how can Microsoft miss? Even if they start with a MckHBOC strategy of buying and integrating, they can get the experience and build their infrastructure. I doubt they would be a player for serveral years because they have such a hard time with execution and follow through. Even with this, their name will get them in the door.

InformaticsMD said...

> The biggest problem in health care IT is interfacing (perhaps I should say the biggest theoretically solvable problem). And presumably if you can drop in an interface engine (especially a really cool one--note that Sun has partnered with a bunch) into a hospital then you don't have to take out the core Clinical departmental systems that are the HIT vendors bread and butter...and of which they all try to remove their competitors products.

It's not theoretical. We did this at Christiana Care in Delaware, an 1100 bed system, in the late 1990's and it worked very well. I don't know what's in use there now, but this was not rocket science.

I think MS would have been better served buying an EMR vendor - but that's expensive and requires core expertise in health IT.

The elephant in the living room that often is ignored is the finding in my field that clinical computing and MIS (business computing) are two different subspecialties of computing, requiring different approaches, methodologies, insights and experience to do properly. See some of the links in my post for more on that issue.