Another viewpoint: Cedars may be doomed to fail again

SAN FRANCISCO—After reporting in Thursday’s Health-IT World News that Cedars-Sinai Medical Center in Los Angeles had made changes in the workflow of its cath lab in the wake of the January 2003 demise of its computerized physician order entry system, I received an e-mail from a physician who says that Cedars has not fixed the fundamental problems that led to the failure.

With the permission of the writer, here goes.

While it is a good thing that Cedars-Sinai is trying to “learn from its CPOE mistakes to improve workflow,” I suggest that the likelihood of failing again may be too high for comfort for two major reasons.

I am uncomfortable that they may assume that a “failure analysis” need not include a comprehensive review of organizational and sociological issues leading to failure. Internal “politics”, territoriality, suitability of the intellectual capital in the hospital I.S. departments towards implementing systems that impinge on clinical practice directly, IT project leadership models, and related issues, are in my opinion of much greater importance than “work flow process analysis” in assuring success. The circumstances of the first failure suggest that these issues are not being managed well.

For more on these matters, see the introductory section of my web page “Common examples of healthcare IT failure” at I authored and edited this material via an internet collaboration with other Medical Informaticists several years ago. This intro was recently published in your competitor’s journal, Health IT Strategist, in August.

The other major reason the second CPOE effort might fail is due to the assumption that “work flow process analysis” is the correct process for a busy clinical environment. A paper written by one of the students here addressed this issue by citing a 1998 article, “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) as follows:

“… Social studies of professional work have show repeatedly that professional knowledge and information cannot be conceptualized as atomic bits and pieces that can be stored and retrieved at will, and that can be unequivocally mapped on simple, universal schemata.”

Professional knowledge is a complicated thing and must be dealt with case by case, it is never the same. 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 work flow 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.

Medicine is complex. Good relationships between the stakeholders (the executives, clinicians, and IT personnel) are the biggest key to success.

While I wish Cedars-Sinai the best of luck on attempt two, I am not optimistic, based upon the learnings you describe from attempt one.

Scot M. Silverstein, MD

Adjunct faculty

Drexel University

College of Information Science & Technology

Institute for Healthcare Informatics


Former director of scientific information resources at Merck & Co. Inc.