Thursday, October 22, 2009

NSW Nightmare and Overuse of Computers: Do We Really Need Full EHR's in ED's?

At "From Down Under: The Story of the Deployment of an ED Clinical Information System ‐ Systemic Failure or Bad Luck" I posted excerpts from a paper of the same name by an Australian medical informatics specialist, Professor Jon Patrick from the Health Information Technologies Research Laboratory (HITRL), University of Sydney, about what appears to be an Emergency Department (ED) nightmare. The paper is here. [10/24 note: not available at present at this link or from its author in Australia; I pulled it from the Drexel U. server in the U.S. at the author's request while his right to post it on his department's server in Australia is being discussed Down Under - ed.]

[2010 update: the paper is now at the following link (PDF): A Critical Essay on the Deployment of an ED Clinical Information System - Systemic Failure or Bad Luck, version 6. Prof. Jon Patrick, Health Information Technologies Research Laboratory, University of Sydney, Australia, Dec. 2009.]

Those excerpts should be frightening to anyone who ever gets ill and might need to visit an ED (meaning, all of us).

An ED electronic health record system (EHR) is being installed in an entire Australian state, New South Wales (NSW), with 200+ hospitals that apparently presents a mission hostile user experience and is causing great opposition by critical care physicians in a setting where death can occur - suddenly and irreversibly - in the flash of an eye.

His essay about the problems with complex EHR's (of American design and manufacture, no less) being installed in Australian ED's -- without doubt a highly expensive undertaking -- raises several questions about both the EHR industry and hospitals themselves.

When I was a CMIO at a 1400+ bed hospital system a decade ago, a regional center in a state with very few hospitals at all (Medical Center of Delaware, now Christiana Care Health System), I counseled that the best solution in my opinion for the very, very busy ED was document imaging of paper, supplemented by a nurse/intake triage system to rapidly record and/or confirm basics (e.g., meds/major problems/allergies/vitals) that was interfaced to the main EHR system.

I based this on the assessment that in the ED, a localized and "closed" environment, the incidence of charts getting lost or writing being illegible resulting in adverse outcomes was minimal. ED charts also did not get lost when patients moved to the floors and the information passed along on paper was adequate for quick transfer and acceptance. Therefore we felt images of past ED charts (of paper) would be satisfactory for assisting care in the ED, where time constraints and hectic pace made the type of EHR system and primary data entry described in the Australian paper disadvantageous (and for exactly the reasons described in the above linked essay from Down Under).

Document imaging is a proven technology that works well even in high volume settings. For example I managed a pharma departmental budget of $13 million, as did an entire pharma company, using an enterprise document imaging system.

I've also been startled by the ED EHR installed at the hospital where I take my mother, who unfortunately has needed far too many admissions in recent years than I care to see. Some of the ED staff were my former medical colleagues and even high school classmates. They've told me, in no uncertain terms, that they felt the system was terrible, again for many of the reasons cited in the Australian essay. My own views of it (albeit brief) showed what appeared to be a mission hostile user environment, including multiple very tiny pulldowns, cornucopia-like picklists, and screens.

Most recently that system did not prevent busy ED docs from almost giving my mother Levaquin after her telling the triage nurse it had caused tendon rupture in the past - and the data being entered. They actually brought in a bag of it to hang, and if I'd not been there as medical advocate for my mother they might have given it. Then when she got to the floor, the next day they almost gave it to her again, except by this time mom was her own medical advocate. I trained in that hospital, and as Admitting Officer for the ED held the record for the most number of admissions, ever, in one night (New Year's Eve 1986), when it was beyond crazy, starting out with one of our own physicians being brought in, shot in the chest, and dying after open heart massage just to set the mood.

The surgeons did the cutting. The heart massage was relegated to the physician present who didn't cut - the internal medicine representative - me.

We managed to treat hundreds and admit several dozens of sick patients using paper, and despite our gloomy emotional state not a single error occurred, to my knowledge.

With all this in mind, I raise these questions:

  • Can the EHR industry actually produce a competent ED EHR that can be used by ED physicians to enter detailed data in real time?
  • Do we really need full EHR's in the ED?
  • A related question: is the extent of ED adverse events related to lost or illegible ED charts (that could not be remediated with non-technology-based solutions far less expensive and troublesome than EHR's) known to a significant degree of confidence?
  • Is there trustworthy literature that shows that the time, expense, and resources for a full ED EHR are worth it in terms of clinical outcomes, ROI etc.? (By trustworthy, I mean scientific peer reviewed literature without author conflict of interest - common in biomedicine as readers of this and other healthcare blogs know, e.g., see this link - not glossy pseudo-P.R. health IT throwaway journals.)
  • Considering that the literature on benefits of EHR's in general is equivocal, I should also ask, is there literature that refutes the value of full ED EHR's, or shows it as possibly having a negative return?
Or:

  • Are we over-computerizing healthcare, even specialties and subspecialties where doing so might actually be deleterious due to the nature of their specific medical environments, just because "we can" (and because there is money to be made by some)?
  • Are we doing so based on irrational exuberance, leap of faith, hope, and/or an uncritical belief that if health IT provides benefits to medical domain "A0", it therefore must benefit medical domains and subdomains B1 to B99 through Z1 to Z99 as well?

I am concerned about the probability that the latter is the case.

I think it not unreasonable to ask these questions before we in the United States spend billions of dollars on our own ED's.

-- SS