Sunday, August 29, 2010

Similar Conclusions on Health IT Via Observation and Via Research: Is HIT 'Mission Impossible'?

The Wall Street Journal reported on a study in Health Affairs entitled "A Progress Report On Electronic Health Records In U.S. Hospitals" by Harvard researcher Ashish Jha and colleagues.

An Aug. 27 WSJ Health Blog post was entitled "Only 2% of Hospitals Could Have Met ‘Meaningful Use’ in 2009."

While the topic of this Healthcare Renewal post is not about Jha's new article per se, I will provide the article's summary:

Despite all the talk about digitizing the health-care world, only 11.9% of U.S. hospitals had adopted at least basic electronic medical records by last year, and only about 2% had done enough to qualify for future government financial incentives, a study finds. The study, published online in Health Affairs, covers responses from 3,101 hospitals surveyed by the American Hospital Association


... It’s actually not surprising that hospitals were slow to adopt new systems in 2009, given the horrible economic conditions, difficulty of raising money for capital investments and uncertainty over what the final government requirements would be. “I’d be shocked if we didn’t see an uptick in 2010 and an even bigger one in 2011,” Jha tells the Health Blog. “But are we going from 2% to 40%? No. We might go from 2% to 5% [in 2010] to 15% or 20% in 2011.”


No surprises there. Jha has previously published research articles I've mentioned on Healthcare renewal such as at "Science or Politics? The New England Journal and "The 'Meaningful Use' Regulation for Electronic Health Records" and "Hospitals Under the Knife: Sacrificing Hospital Jobs for the Extravagance of Healthcare IT?":

... Ashish Jha’s research at the Harvard School of Public Health that compared 3,000 hospitals at various stages in the adoption of computerized health records and found little difference in the cost and quality of care. A New York Times story "Little Benefit Seen, So Far, in Electronic Patient Records" on those findings is here.

What is of interest to me are statements made by Jha in response to questions asked by the WSJ reporter Katherine Hobson. Hobson: "We hear from a lot of physicians and others who are doubtful about the benefits of electronic records, and we asked Jha about the evidence supporting quality and safety improvements."

Jha's response was quite interesting. There are a number of issues he raises:

[Jha] "There are really two sides of the coin. I can point to about 400 studies done really well that show electronic health records can have a substantial impact on improving quality and safety. [But] we had a paper published earlier this year [Electronic Health Records’ Limited Successes Suggest More Targeted Uses - ed.] showing that when hospitals adopted EHRs, there was no impact on quality. So how to reconcile that?"

[Hobson] Jha says we have evidence that “when high-performance hospitals develop their own systems, implement them and tweak them, they can work wonders.” But the benefits gained from purchasing an off-the-shelf product and failing to implement or tweak it to meet an institution’s particular needs are likely to be “far less pronounced,” he says. “If you adopt a new technology, and do it badly, you can end up making productivity worse” or causing harm, he says.

[There must be a lot of organizations meeting that description for such low impacts on quality to be the norm - ed.]

“This is not a plug and play.

I find this interesting for several reasons.

First point of interest:


On doing health IT badly, and on health IT not being "plug and play":

I came to the same industry-challenging conclusions a decade ago observationally that Jha came to through empirical research and review of the literature. On "doing health IT badly", in the late 1990's I'd written the following (now at my academic health IT site here):

... While clinical IT is now potentially capable of achieving many of the benefits long claimed for it such as improved medical quality and efficiency, reduced costs, better medical research and drugs, earlier disease detection, and so forth, there is a major caveat and essential precondition: the benefits will be realized only if clinical IT is done well. For if clinical IT is not done well, as often occurs in today’s environment of medical quick fixes and seemingly unquestioning exuberance about IT, the technology can be injurious to medical practice and biomedical R&D, and highly wasteful of scarce healthcare capital and resources.


Those two short words “done well” mask an underlying, profound, and, as yet, largely unrecognized (or ignored) complexity. This website is about the meaning of "done well" in the context of clinical computing, a computing subspecialty with issues and required expertise quite distinct from traditional MIS (management information systems, or business-related) computing.


I'd also made the observation, to the consternation of health IT marketing departments I'm sure, that health IT was not plug-and-play in an essay entitled "Cultures of mismanagement: toxic to healthcare quality" about my experiences as a CMIO:


... To make matters worse [in an ongoing EPR project], the executive team then gave a key EPR staff member, a Senior Resident who'd done an excellent job writing and programming custom templates for the EPR system, a difficult time on promised payment for his services.


They believed such a customization function was trivial and wasteful, and essentially reneged on their agreements with the Resident. When challenged by the informaticist and others that this person's services were essential, the views were met with indifference, if not disdain, for facts and logic. In fact, the executive team clung persistently to a mind-numbing leap of logic: they seemed to believe that just as home computers were "plug and play", so was [or should be - ed.] clinical IT. Their attitudes seemed to reflect a belief that the EPR team and resident were basically deceiving them.


... Unfortunately, healthcare IT is never plug-and-play, and in IT a person is either part of the solution or part of the problem.


Second point of interest:

Jha did mention two systematic reviews as describing hundreds of articles in the literature supporting potential benefits of health IT, one in May 2006 by Chaudhry et. al in the Annals of Internal Medicine (link), and one by Goldzweig et. al in Jan. 2009 in Health Affairs (link).

These reviews were rather weak on substance, though, in terms of robustness that could support spending hundreds of billions of dollars on health IT as national policy. The former article concluded:

Limitations: Available quantitative research was limited and was done by a small number of institutions. Systems were heterogeneous and sometimes incompletely described. Available financial and contextual data were limited.

Conclusions: Four benchmark institutions have demonstrated the efficacy of health information technologies in improving quality and efficiency. Whether and how other institutions can achieve similar benefits, and at what costs, are unclear.


The articles' conclusions matched my own writings from observation and critical thought about what I was seeing personally and hearing from Medical Informaticists in the field.

Third point of interest:

In the WSJ blog post comment section, several commenters including Dr. Jha opined on the issue of lack of regulation of health IT as a factor in its limited diffusion and benefit to date. In a comment of 8:06 am August 29, 2010 Jha wrote:

... On the issue of FDA approval [of HIT devices - ed.] — this is a tricky issue on which reasonable people can disagree. FDA regulates devices that are used on patients (not on doctors and nurses) traditionally so you could make a very reasonable argument that this is out of their scope.


This argument goes back to the customary notion of the clinician as "learned intermediary" thus providing a shield from liability for the vendors of commercial HIT systems.

Here I disagree with Jha. I would argue, again from observation, that this notion is obsolete.

For example, in an item from the HIStalk blog site news of 8/30/10 (link):

... [a reader] sent a UPMC internal document describing a quality improvement project at Shadyside [Shadyside Hospital, Pittsburgh - ed.]. Before Cerner was implemented, the ED stocked 95% of the meds they used in their automated dispensing cabinet and the other 5% were requested by tubing the paper order to pharmacy, resulting in a pretty good turnaround time of 28 minutes. With CPOE, the ED had no good way to alert pharmacy about those 5% of meds, so turnaround time shot up to over two hours. An ED nurse came up with a solution: tubing a yellow “Stat ED Med” card to pharmacy along with an index card listing the patient and med needed, dropping TAT back to six minutes.

From the medical informatics perspective, the lesson here is that the physician is effectively no longer a “learned intermediary” between computer and patient when a complete system of EMR, CPOE, decision support etc. linked to Pyxis medication dispensing and other machines is constructed.


This model from the earlier days of simple advice-giving and-records keeping health IT no longer holds (arrows represent interaction):


[Computer] <---> [Clinician] <---> [Patient]

Rather, the computer has become a cybernetic governor of care, as in a programmatic intermediary between clinician and patient regulating care operations:


[Clinician] <---> [Computer] ---> [Patient] (note one way arrow).

The fact that a “workaround” a.k.a. an “override around the cybernetic regulator of care” had to be created in order for appropriate care to be delivered in a critical care environment supports the latter conceptual and legal model.


You should not have to work around something that is not in the way.


In other words, the computer is now the intermediary. There only way around it are workarounds. Articles by Koppel on CPOE and barcoding workarounds are other examples of how the health IT system is now the intermediary between clinician and patient, not the clinician as learned intermediary between IT and patient.


One might also ask - what is the cognitive impact of continually devoting intellectual power to create work arounds, which add to the already complex multi-tasking of clinicians? It likely is not a positive contributor to the best care.


These and many other issue contribute to the low diffusion of health IT in 2010. It is interesting that much of this was known to people in Medical Informatics (at least those without industry conflicts of interest) through in situ observation of health IT projects, going back many years, in fact decades with regards to the pioneers (as mentioned in my post "Healthcare IT: How Much More Out of Control Can An Industry Be?").


Finally:


The question should be asked: will health IT ever achieve the benefits claimed for it? Will it ever "revolutionize", or even make substantial contributions to medicine in a cost effective, non negatively-disruptive fashion?


Will we ever stop "doing it badly" (per Jha) and start consistently "doing it well" (per me?)

I am beginning to believe that health IT will never be successful without a major revolution in the IT field itself.


In 1969, EMR and Medical Informatics pioneer Donald A. B. Lindberg, M.D., now Director of the U.S. National Library of Medicine at NIH, made the observation that

"... computer engineering experts per se have virtually no idea of the real problems of medical or even hospital practice, and furthermore have consistently underestimated the complexity of the problems…in no cases can [building appropriate clinical information systems] be done, simply because they have not been defined with the physician as the continuing major contributor and user of the information" (Lindberg DAB: Computer Failures and Successes, Southern Medical Bulletin 1969;57:18-21).


I believe the problem was -- and is -- not "computer engineering experts" but "business computing personnel."


First, there's the issue of the customs and traditions of business IT. As I've written before, these rigid traditions are antithetical to the poorly bounded, rapidly changing, improvisation-rich world of clinical medicine.


Second, the current status quo is highly profitable to the vendors, resulting in reactionary mindsets and conduct.


Third, there is a lack of appropriate intellectual capital in leadership roles in health IT. Medicine is a largely scientific field. Automation in the field is a highly scientific endeavor. Yet the field is dominated by business computing personnel (usually of a management information systems background), instead of true computer scientists and engineers, human factors scientists, information scientists, and the like.


If my observations of who went into science/medicine/perhaps law vs. who went into business-related fields when I graduated high school and college are indicators, the very best scientific minds are poorly represented in healthcare IT leadership circles.


In summary, both observations of health IT and research on health IT show many stumbling blocks towards achieving major, unequivocal, reproducible benefit with a robustness that could support spending tens or hundreds of billions of dollars on the technology in ambitious national-scale implementations.


I fear that under the current culture of IT, achieving effective national health IT is Mission Impossible.


Your mission, Mr. Phelps, should you choose to accept it...


A vast body of new, critical thinking about the technology and its industry is needed.


Aug. 30 note: As a result of a complaint to the blog owner (not to me, curiously) that I am not providing answers to the problems I write about...


... his long screed ends with "national Health IT is mission impossible"... and while he says "new, critical thinking is needed", he doesn't say what that is. His next post is even more out there.


... I will add this:


I do not have the answers (if I did, I'd sell them). This is a serious social problem. The best I can do, and have done, is point out what not to do - and even that has not had much of an effect despite ten years of public writing in numerous venues.


I do know that believing there's a major problem is the first and most important step to solving these vexing issues. We as a country and culture have not yet reached that point; irrational exuberance prevails. Thinking about health IT requires a reboot, but denial of the problems must stop before that can occur. I exert my efforts largely in an attempt to bring the health IT patient out of denial.


Final addendum 8/31:


On reconsideration, it's not as if I haven't offered advice on "what to do" as well; see for example my post "Science or Politics? The New England Journal and The 'Meaningful Use' Regulation for Electronic Health Records."


-- SS