Friday, January 16, 2009

Throwing The National Research Council Report On Health IT Under The Bus, Part 1: MedStar Health

I wrote about the just-released National Research Council report on HIT at the post "Current Approaches to Health IT Insufficient ... and Other Master of the Obvious News."

Critiques on the National Research Council report are to be expected. It will likely have a major impact on the HIT industry and those with special interests in that industry. There will be critique, I expect, to a significant degree, unfortunately up to and including in today's political climate ad hominem attacks on its authors, I'm afraid. (I am definitely not implying the latter is the case below, however.)

The following critique, seen at the HISTalk industry-sponsored gossip site here, has me scratching my head a bit.

I'm scratching my head at the following commentary on HIT and the NRC report by an official at MedStar Health, Peter Basch, MD, Medical Director of Ambulatory Clinical Systems.

Points that leave me wanting are as follows:


... in spite of this clear support for funding and continued development of HIT, some media headlines have painted this report as harshly critical of the potential of HIT in general and EHRs in particular. This media misinterpretation resulted primarily from two faults inherent to the report:

(1) the NRC’s mislabeling of their recommendations as a change from what health IT leaders are advocating for; and

(2) the NRC’s inappropriate assignment of blame to EHRs as being the cause of dysfunction rather than their understanding that EHR functionality and implementation deficits are a result of a dysfunctional reimbursement system, which is based on volume of episodic care and verbosity of documentation.


My comments:

Did the NRC mislabel their recommendations for improving HIT as a "change" from what health IT leaders advocate? For example, I haven't heard many industry based health IT leaders advocating for -- or generally supporting with money - studies of interdisciplinary research in biomedical informatics, social science, or healthcare engineering, for example.

In fact, at the
Government Health IT Conference & Exhibition 2008 in Washington, the plenary session leaders and many attendees, including the "experts" from large HIT management consulting organizations, seemed to have little idea what I was talking about when I mentioned the study of social factors. They largely blamed doctors for HIT's ills, until I gave some examples where that was decidedly not the case, and the plenary members then agreed there might be other "unknown" issues at play. Interestingly, the attendees from the VHA knew precisely what I was talking about.

I believe the NRC recommendations are indeed "change we can believe in" compared to the usual and customary HIT industry chatter. I also believe the media reported accurately on the current state and potential of HIT under its current design and management paradigms (note the latter point; I am an HIT advocate, but only if done right.)


Did the NRC "blame" EHR's as the cause of dysfunction, or as a symptom of some deeper issues? The latter seems the case.

Are implementation and functionality deficits a sole result of the reimbursement system? Doubtful, see below regarding HIT in the UK and Australia.
While I agree we have a dysfunctional reimbursement system here in the U.S. that makes quality HIT harder to achieve (my own organization had to sue AllScripts over a malfunctioning E&M module, for example, see civil complaint PDF here), I believe there are far more fundamental issues at play in HIT failure.

Another point made in the HISTalk posting:

... While it is true many early adopter systems believed (at the time) that merely switching off paper medical records to EHRs would lead to improved and safer care , nobody has believed that, or has advocated that position in years

My comments:

"Nobody has advocated EHRs would lead to improved and safer care in years?"
... What about
the statement "To improve the quality of our health care while lowering its costs, we will make the immediate investments necessary to ensure that within five years, all of America's medical records are computerized" by the President-elect on Jan. 8, 2009 as just one example?

That's an ambitious timeline indeed for a technology our National Research Council tells us is not yet ready to provide such benefits without significant rethinking and rework. It certainly send a clear message to the public that HIT in 2009 is a magic bullet, a panacea towards better healthcare quality. It suggests the new administration has absorbed that view through ill informed HIT industry and pundit wishful thinking, and perhaps purposeful disinformation.

I can easily find myriad quotes expressing the same point of view, that EMR ipso facto equals better healthcare quality, on a google search "EMR improves care" or similar search engine query. I think those in informatics should be educating on the dangers in that view, not denying it exists.

A third point in the HISTalk posting:

... The NRC faults current EHR build as not supporting the cognitive support necessary to optimize care. This deficiency is obvious and abundantly clear to veteran EHR and HIT users – many of whom work on their own or with vendors on new and better functioning clinical decision support. However, let’s be fair as to the root cause of this deficiency. It is neither lack of vendor vision nor limitation of IT technology; it is lack of a market.

My comments:

"The root cause" of misdesign of HIT, its glaring deficiencies towards its users' needs after 40+ years of trying, is simply - market forces?

Are any of the sociotechnical ("people and their interactions with technology") issues at my website here or at other sites by those studying social-IT issues and misalignment of the business IT paradigm towards clinical computing relevant? Or, should such issues simply be thrown under the bus since under government stimulus a better market will magically appear?

Is the entire discipline of social informatics (
the study of information and communication technologies in social, cultural and institutional contexts) irrelevant?

What about the problems with the UK's massively government funded national HIT program (see here), and the problems in Australia (see here)? Were the reimbursement system and the market a cause of the problems there? Doubtful. Here's what the UK national program's former leader Richard Granger had to say:

"Sometimes we put in stuff that I'm just ashamed of ... Some of the stuff that Cerner has put in recently is appalling ... Cerner and prime contractor Fujitsu had not listened to end users ... Failed marriages and co-dependency with subcontractors ... A string of problems ranging from missing appointment records, to inability to report on wait times ... Almost a dozen cancelled go-live dates ... Stupid or evil people ... Stockholm syndrome - identifying with suppliers' interests rather than your own ... A little coterie of people out there who are "alleged experts" who were dismissed for reasons of non-performance."

How familiar to Medical Informatics experts that all is.

In summary, are the market and the reimbursement system the major causes of HIT's disappointing track record as claimed in the critique? Hardly.

Perhaps there are other far more fundamental "root causes" for HIT dysfunction that this critique of the NRC report is leaving out, including the issues above as well as additional factors mentioned in my posting here.

We as clinicians, especially those who are biomedical informatics professionals, cannot improve the situation by making ill informed, wishy-washy excuses for the HIT vendors. While the latter, as in pharma, might treat us, as well as CIO's and IT hospital personnel to nice perks, promote us in speaking engagements as key opinion leaders (KOL's), give grants, etc., they have produced mediocre to very poor products for the most part for decades. This wastes resources that healthcare can ill afford, ultimately harms or at best does not help patients, and needless to say makes reports such as the NRC report and Joint Commission Sentinel Event Alert necessary.

I am aghast at the stories I hear even today of HIT products such as EHR's and CPOE's with cryptic user interaction design
that inundate, confound and insult clinicians, force clinicians to "drink information from a firehose", or have other issues that waste time and create new opportunities for error.

To be quite frank, the IT industry spends exponentially more time and expense on design and debate over trivial and arcane features in computer operating systems (to ensure a "better user experience", the fluffy marketing-ish industry buzzword du jour for what used to be called "efficient interaction") than on HIT.

As in medicine itself, we cannot cure this situation by failing to recognize or by ignoring the symptoms, signs and features which are right before our eyes.

Ironic note - While
I do not know the author of the critique, a few of the case studies of HIT dysfunction and difficulty at the above referenced sociotechnical issues website here involved a high level person now at MedStar. I can assure readers that the issues then had nothing to do with "reimbursement" and "lack of a market" and very much to do with "lack of vendor vision" as well as "lack of hospital management and IS department vision" in the prior organization.

-- SS

7 comments:

Anonymous said...

The NRC has politely told the industry that their products are inferior and insufficient. Based on recent reports from the VA and safety alert from JC, the earlier works of Koppel and of Han, the conclusions of NRC are indisputable. The glove fits. The HIT industry should not rely on the users to fix its products' defects. Basch is just a rah rah guy for his company...sure he will be defensive.

InformaticsMD said...

Basch is just a rah rah guy for his company...sure he will be defensive

I'm basically tired of physicians advocating for technologies not ready for prime time, no matter what the cause.

-- SS

InformaticsMD said...

Dr. Basch replied to this post. I respond to several of his comments which appear in italics:

I am a practicing physician and early adopter of HIT - and have been described (pejoratively in the early years of the HIT movement)as a HIT "realist" - for refusing to "drink the koolaid" - and blindly support anything said or done on behalf of health information technology. Most of my writings and presentations have either been focused on, or included warnings and critiques re the overselling of HIT

I am also a very early adopter of HIT, have studied the clinical, technical and social issues at the postdoctoral level (as is required in any medical subspecialty to achieve the highest skill levels and professional credentials) and agree with Dr. Basch that HIT has often been oversold.

And much of my time within my health system is currently spent on helping to improve limitations in our existing and upcoming implementations. In fact, this post was the first time I have been critical of anyone [the NRC report - ed.] for underselling the potential of HIT.

I don’t think a realistic empirical study of HIT issues and the resultant report represents underselling HIT’s potential. Saying so undercuts the report.

The report represents an objective appraisal of how HIT performs at this point in time. I think its conclusion was clear:

"The nation has made a commitment to achieve the nationwide use of electronic
medical records by 2014. Many meaningful and useful steps can be taken today toward
this goal. However, this goal reflects expectations for improvement in the quality and
cost-effectiveness of health care that will require more than just wider implementation
of today’s health care information technology.
At the start of its work, the committee had set out to identify a variety of long-
term opportunities for greater involvement of the computer science research community
in addressing health care problems. And indeed, the committee did identify a number of
such opportunities, which are described above. But it was also struck by the number of
other opportunities for meaningful progress that do not depend on research—that is,
areas of improvement in which today’s information technologies are reasonably
adequate for initiating and sustaining meaningful progress and yet are not being fully
leveraged for health care. In addition, the committee was surprised to see how little
attention had been paid—across the board—to support for the cognitive functions that
clinicians manage, organize, coordinate, and use the vast amounts of information
needed for effective health care. It is in this domain that the committee believes
enormous leaps and bounds are possible, and also where a substantial number of grand
research challenges reside."


Either we are scientists who objectively approach HIT, both its potential deficiencies and potential benefits, or we are opportunists no better than those who promoted VIOXX due to “gray areas” in its pre-launch and postmarktet profile.

The subtitle of this post is "throwing the NRC under the bus." Since when does agreeing with the majority of a report, and publicly lauding it for its sober analysis and insightful recommendations count as "throwing something under the bus.

In my view, your essay and its points instill fear, uncertainty and doubt about the NRC report as "underselling" the potential of HIT and as containing major errors (e.g. regarding causation). A report with errors is a report that can more easily be dismissed ("thrown under the bus.") That is my opinion.

It is very true that I objected to three pieces of the report. First - that the current conventional wisdom is "implement HIT and everything will be fixed."… Basically, I was attempting to articulate the point that what the NRC called a change in direction was a change that most of us fortunately made several years ago.

You stated that “While it is true many early adopter systems believed (at the time) that merely switching off paper medical records to EHRs would lead to improved and safer care, nobody has believed that, or has advocated that position in years.”

I pointed out how that statement was profoundly untrue by citing the President elect himself, who is likely (not being an HIT expert himself) receiving such information from his advisors who in turn are receiving it from industry leaders, and via a simple google search which shows such statements are common.

If by "nobody" you mean an elite group of people involved as CMIO's, then I can say you've done a very poor job of making that view public. My own website that addresses that issue, for example, remains nearly unique since 1998 when it was started after my time as a CMIO. Thank god, then, for the NRC report ... the Web ...and for blogs.

Secondly, while it is very true that even the best EHRs are not ideal, many of them are good enough to show incremental benefit - even without payment reform.

As I wrote at my extensive HIT sociotechnical issues website:

"With regard to electronic health records (EHRs), a research article in the Archives of Internal Medicine entitled “Electronic Health Record Use and the Quality of Ambulatory Care in the United States” (Arch Intern Med. 2007;167:1400-1405, link to abstract here) reached what to many was a counterintuitive and paradoxical conclusion. The authors examined electronic health records (EHR) use throughout the U.S. and the association of EHR use with 17 basic quality indicators. They concluded that “as implemented, EHR’s were not associated with better quality ambulatory care,” and were bold enough to publish their findings, sure to be unpopular in the health IT industry."

Is the incremental benefit of inadequate HIT worth the expense at this time, when millions are without basic health insurance? See my posts on HC Renewal about medical centers spending $70 million, or $100 million, or more on clinical IT. e.g. here.

However, let’s be fair as to the root cause of this deficiency [in EHR suitability to task -ed.]. It is neither lack of vendor vision nor limitation of IT technology; it is lack of a market.

I pointed out the chaos in the UK and problems Down Under paralleled those here, if not being worse; yet in the UK the HIT industry had not just a market, but a well paid closed market. Granger himself wrote:

"Sometimes we put in stuff that I'm just ashamed of ... Some of the stuff that Cerner has put in recently is appalling ... Cerner and prime contractor Fujitsu had not listened to end users ... Failed marriages and co-dependency with subcontractors ... A string of problems ranging from missing appointment records, to inability to report on wait times ... Almost a dozen cancelled go-live dates ... Stupid or evil people ... Stockholm syndrome - identifying with suppliers' interests rather than your own ... A little coterie of people out there who are "alleged experts" who were dismissed for reasons of non-performance."

I do not believe market issues are a critical root cause of HIT’s deficiencies. The problems are far more profound. I fear many who lack formal medical informatics, IT, information science and social informatics backgrounds may not appreciate the issues fully.

As a licensed radio amateur, for example, I know a lot about radio; I even passed a series of exams far more thorough than the exam one takes, for example, to become a "CPHIMS" (certified professional in health information management systems, see my post here.)

I am not, however, qualified to comment on root causes for telecommunications blackouts to police such as we've suffered in this city numerous times.

Third - as someone who has always thought that documentation requirements for payment serve to create clutter and obfuscate wisdom - I think it is unfair to blame EHRs, or to label wasting time on documentation as something that EHRs created.

I wrote “Did the NRC "blame" EHR's as the cause of dysfunction, or as a symptom of some deeper issues? The latter seems the case.”

The blog author mentions that he expected... personal attacks against the NRC

In today's political climate, can you anyone credibly argue with that assertion, I ask?

... and appears to count my post as a personal attack against the NRC. Nothing could be further than the truth.

I have amended the paragraph from which you drew that unfortunate conclusion about your own posting to ensure others do not.

I would hope that readers will remember that the NRC report and my response - were both written in good faith.

Of course. Our interests both coincide with the ultimate consumer of the healthcare system – patients.

However, good medicine goes way beyond good intentions. Truly speaking truth to power requires listening to those in academia with far more experience than one's self, e.g., the authors and editors of the NRC report (Stead, Barnett come to mind).

-- SS

InformaticsMD said...

Addendum: for clarity, although the link was in my HCRenewal post, Dr. Basch's HISTalk essay is here:

http://histalk2.com/2009/01/14/readers-write-11509/ .

My own website, started in 1998 and entitled "Sociotechnologic Issues in Clinical Computing: Common Examples of Healthcare IT Difficulties", is here:

http://www.ischool.drexel.edu/faculty/ssilverstein/medinfo.htm .

Anonymous said...

Let's face the reality. These devices are not ready for prime time. The manufacturers should have tested their devices for safety and efficacy prior said devices' deployment. They should not be using the hospitals as a proving ground in an unconsented experiment getting feedback from the health care professional users (or the NRC, for that matter) as to how to improve the device, all the while, endangering patients who are the subjects without their consent. One third of the VA hospitals reportedly had "glitches" in their devices' software. The fact is that yhese are not "glitches"...these are defects consistent with the device not being ready to be used on patients.

http://www.cbsnews.com/stories/2009/01/14/national/main4720942.shtml?source=RSSattr=U.S._4720942

A RECALL IS INDICATED.

Anonymous said...

The problems with HIT are also found in other types of IT implementations. One of the problems is also the large consulting companies or "systems integrators." These firms are lead by smooth Partners and Directors. Having worked with, them I can say that they are primarily professional liars. They create "relationships" with the vendor and client that are all about reducing the need to perform.

Intelligent HIT implementations should be hosted and require minimal installation. Also, complexity should be minimized. Finally, the hospitals themselves have to have technologists, not ex-MBAs and hand shakers in important decision making positions. Playing golf well, or having nice hair and a Nordstrom's charge card don't help implementations go in on time or help good technology solutions get chosen.

InformaticsMD said...

Shaun Snapp wrote:

Playing golf well, or having nice hair and a Nordstrom's charge card don't help implementations go in on time or help good technology solutions get chosen.

In fact, such characteristics among technology leadership often harm such efforts. We don't tolerate the incompetent physician or surgeon very well due to the damage they cause, after all.

The current world economic crisis is fair evidence, in fact, that many of these "privileged" people in fact established this privilege through these social skills, but are in fact intellectually and ethcailly challenged in many spheres.