Friday, March 18, 2005

AMIA on the JAMA CPOE article controversy

A reasonable assessment of the recent CPOE controversy by Don Detmer, MD, the President and CEO of the American Medical Informatics Association, is below. He writes that healthcare IT is of great value, is moving forward, and that those in clinical IT need to rigorously assess issues and problems and correct them iteratively, as in any evolving field. He is confident that AMIA and its researchers, clinicians, and educators are committed to these objectives and will continue to raise the performance bar for clinical IT. I heartily agree, as does Ross Koppel, the author of the AMIA article that caused the controvery on CPOE (see the end of this blog entry).

A Statement from the AMIA President and CEO
March 15, 2005

In politics and marketing it is often said, “all headlines are good headlines.” There certainly have been a lot of headlines lately about health care IT. The March 9 issue of the Journal of the American Medical Association included two articles on the state of computerized clinical decision support and computerized physician order entry systems and an editorial on the state of computer technology in clinical work more generally (Garg et al., 2005; Koppel et al., 2005; Wears and Berg, 2005). And this week, the British Medical Journal published the results of an extensive review of clinical decision support systems while AMIA announced the availability of a white paper on clinical decision support in electronic prescribing (Kawamoto et al., 2005; Joint Clinical Decision Support Workgroup, 2005).

The JAMA articles as well as those in the BMJ have generated a lot of interest and debate within the medical informatics community. A careful reading of these articles provides us -- the developers, purchasers, users, and advocates for these system -- with a fresh perspective on how to evaluate our past efforts and frame our future work. Some readers can miss what may be the deeper message of the recent JAMA articles and editorial. As a result, they may reach the conclusion that once again the information technology in health care has been oversold and that there is no reason to attend to or invest in the current generation of information technology. I would in particular like to take issue with Wears and Berg’s suggestion that computer technology and clinical work are like Estragon and Vladimir of Beckett’s Waiting for Godot. Although we are by no means where we want to be with respect to the development and implementation of electronic health records, clinical decision support, computer physician order entry systems or a national health information infrastructure, we are also by no means still in the same place as where we started.

Progress in any complex environment comes in steps, but progress has been achieved over the past two decades. The progress in computer-based health records is now sufficiently successful that newer versions of older legacy systems such as those reviewed in JAMA have left the centers of these early adopters and innovators – the enthusiasts – and are entering mainstream care. Meanwhile, some of the innovating institutions are just now upgrading from their legacy systems. Like the older systems, these newer ones also involve substantial transformations in practice behavior. And, like the older systems, the newer ones also come with human adaptation risks that sometimes create new problems as well as new solutions. Further, there is always pushback from those clinicians who simply don’t wish to change. Historically, most substantial changes to clinical practices that are resisted by clinicians appeal to a concern that the quality of patient care is being compromised. This was noted as long ago as the origins of the use of the stethoscope.

Not all change is progress, but some of it actually is. The informatics community does need to keep its focus not upon simply putting heretofore paper processes into a computer format but to redouble its resolve to create evidence-based applications that do indeed transform care by involving patients themselves in utilizing personal health records, by enhancing safety, by increasing the timeliness of care through use of the Internet, and by enhancing efficiency. Further we need to continue to reinforce the message so aptly stated by Wears and Berg, “any IT acquisition or implementation trajectory should, first and foremost, be an organizational change trajectory. This is true at the organizational level and the national level …” Thus, we must continue to advocate for other system changes that are needed for health care transformation – including the creation of appropriate incentives through reimbursement mechanisms and a workforce trained to use developing information technologies effectively.

While there is much work yet to be done, the optimism remains that the use of communications and information technology will greatly improve both care and human health. Emerson noted that in nature, nothing is free. While there are real prices to be paid as a result of experimentation, at time things actually do get better. What we need to do is to continue to assess rigorously the balance between the sunny and shady sides of IT innovation and ‘stick to our knitting’. I am confident that AMIA and its researchers, clinicians, and educators will remain committed to these objectives and will continue to raise the bar for our performance.

Don E. Detmer, MD, MA, FACMI
AMIA President and CEO

The only addition I'd make to Dr. Detmer's assessment would be to the statement "any IT acquisition or implementation trajectory should, first and foremost, be an organizational change trajectory. This is true at the organizational level and the national level." I would add that "organizational change" should include investigation into change in the leadership and management of clinical IT projects towards leadership by cross-disciplinary experts, in recognition that those solely of a Management Information Systems (MIS, or business computing) background often do not have adequate knowledge and expertise to effectively lead major change trajectories in scientific computing, of which clinical computing is a subset.

It has often seemed unreasonable to me that the leadership and process models of business computing could be deemed adequate for the needs of major change trajectories in clinical medicine. Via an internet collaboration several years ago, I and other medical informaticists documented our own examples where the unidisciplinary MIS leadership model has not only been inadequate, but actually harmful.

I think the Koppel JAMA article does reflect what happens in real life in today's medical centers. The UPenn authors advise appropriate caution in implementing any clinical IT. They state that "with any new technology, initial assessments may insufficiently consider risks and organizational accomodations." I agree. It then goes on to suggest how organizational and risk issues should be handled, including the "need for continuous revisions and quality improvement."

It's not simply a question of obsolete software versioning either, as some have stated. As Joan Ash, Ph.D. noted at Oregon Health & Science University in a 2003 study “Most hospitals don't use latest ordering technology” ( ), "patient care information systems like CPOE also can create unintended or "silent" errors, according to a separate study conducted by Ash and colleagues in The Netherlands and Australia. This study also was published in the online version of JAMIA.

The study's authors divide these silent errors into two main types: errors during data entry and retrieval, and errors in the communication and coordination process. Both types of errors occur because the systems simply don't take into account the work atmosphere most health care professionals experience, according to Ash, also lead author of this study ... "many information systems simply don't reflect the health care professional's hectic work environment with its all too frequent interruptions from phone calls, pages, colleagues and patients. Instead these are designed for people who work in calm and solitary environments … some patient care information systems require data entry that is so elaborate that time spent recording patient data is significantly greater than it was with its paper predecessors," the authors wrote. "What is worse, on several occasions during our studies, overly structured data entry led to a loss of cognitive focus by the clinician."

Attempts to develop a CPOE system internally, such as at Cedars-Sinai Medical Center in Los Angeles, have also not been entirely successful (AMA News, Feb. 2003, "Doctor Pull Plug on Paperless System"), subtitled "California's Cedars-Sinai turns off its computerized physician order entry system after physicians revolt, demonstrating that implementing new technology is easier said than done." I was not sanguine about the planned remediation attempts, either ("How to Avoid CPOE Failure a Second Time"), which seemed more focused on the traditional MIS "process analysis" than on the organizational and sociotechnical issues that likely led to unusability and revolt in the first place. As Berg noted, "The depictions of the formal workflow of medical work are often not realistic; for example, the task boundaries between doctors and nurses are not always tightly drawn. Translating professional knowledge and workflow processes into some automated record is not impossible, but much care and evaluation must be taken for it to not be hazardous to a healthcare environment." ("Considerations for sociotechnical design: experiences with an electronic patient record in a clinical context", Berg M. et al, International Journal of Medical Informatics, 1998;52:243-251)

An interesting collection of related articles on sociotechnical aspects of clinical IT can be seen via a PubMed search initiated by clicking on this link .

I think any controversy about the UPenn article is coming from emotions, not logic, as a result of the Penn authors having critiqued "established wisdom" (i.e., dogma) on CPOE, and from vendors who see it as putting even more doubt in the minds of capital-strapped hospital executives.

In other words, it is a political controversy, not a rational, scientific one.

The article "Experts Expect CPOE Adoption To Continue Despite Study on Errors" by Fanen Chiahemen, iHealthBeat staff writer, is also worthwhile. Of note are the following excerpts:

Ross Koppel, a sociologist at the University of Pennsylvania School Medicine and the study's lead author, conceded that there initially was some fear that the health care industry would now "go slow" on CPOE adoption, but he said "that would be like throwing the baby out with the bathwater." Koppel said that the study was not intended to discourage the adoption of CPOE and that the authors are in favor of the technology. "We're very much pro-CPOE, but we're against arrogant CPOE," Koppel said. "Do I want to go back to a system where somebody scribbles on a piece of paper? No," he added.

... Paul Tang, the chief medical information officer at the
Palo Alto Medical Foundation, said he hopes the study will increase awareness about what is required to implement IT in health care. "Successful implementation of CPOE and realization of the safety benefits from using the technology requires good software design, a clinically led implementation effort and good training," he said. "It would be a shame if we lost sight of the lifesaving potential of CPOE because of a report of one experience in which many of the critical success factors -- good product design, good implementation strategy and good training -- were not in place," he said.

Wears said training plays a role in enhancing the effectiveness of such systems, but he believes the real issue is better system design. "If you need a lot of training to take young, computer-savvy, highly motivated people to get them to use this thing to do their work, that should tell you something," he said. "These are not the kind of people that should normally take a lot of training if the tool actually helps them," he said.

Building on that notion, Suzanne Delbanco, CEO of the
Leapfrog Group, said the study could put a little bit more pressure on vendors to develop systems that have all the capabilities that benefit patients as much as possible.

However, [Upenn author] Koppel stressed that even new systems have to be thoroughly investigated and constantly evaluated because "every change has cascading changes." He said he has been looking at some new systems that have "tons of foolish programming and poor integration issues." Therefore, health care practices that choose to adopt CPOE not only have to closely monitor how physicians and nurses use the systems, but they also have to "be a lot more humble and a lot more vigilant," Koppel said.

I strongly agree with that assessment. I also believe Medical Informatics specialists (MD's, PhD's, nursing informaticists, etc.) need to take genuine leadership roles in clinical IT initiatives if these initiatives are to succeed in a time frame and at a cost acceptable to those who provide the capital and expense dollars. By leadership I mean substantive control of vendor product evaluations, project hiring (including personnel evaluations), major decisions affecting clinical care, and other factors that distinguish leaders from what I once termed "Directors of Nothing."

The MIS mantra "doctors don't do things with computers" may or may not have been appropriate in the days when the MIS department was referred to as "Data Processing", but as I pointed out in a 1998 webpage on stereotypes of medical informaticists, is entirely inappropriate today [note: I have not maintained that old site; some links will not work]. This is especially true now that the U.S. and other countries such as the UK are embarking on national EMR initiatives.

-- SS

No comments: