Skip to content Skip to navigation

Evidence-based Medicine, and Less!

July 26, 2008
by Joe Bormel
| Reprints


Last week, I participated in the 17th annual AMDIS Physician Computer Connection symposium in Ojia, CA. As Larry Stofko suggested in his blog post on reporting relationships (), many of the other participants at the meeting were CMIOs, with a healthy balance of CIOs and other executives. Not surprisingly, the topic of Evidence-based Medicine (EbM) came up, again and again, as our industry moves from “how do we bring these systems live” to “how do we best use these systems to improve care and its delivery.”

There’s a useful earlier dialogue and presentation on EbM that I assembled two years ago (). The dialogue, then and at the AMDIS meeting last week, clarified that in medicine there is often the lack of scientific certainty (as well as social agreement) around what to do in many real situations. The consensus was, EbM is great when it exists. Something one notch less than EbM is even more valuable and practical in our real and imperfect world. This same point came up in Kate Huvane’s recent post (), described as “And when science fails, maybe art needs to step in.

Almost ten years ago there was a fabulous related JAMA article “Why Don’t Physicians Follow Clinical Practice Guidelines?” by Michael D. Cabana, MD, MPH, et al.

I have a friend, Vi Shaffer, who currently serves as a research VP with Gartner. At the time Cabana’s article was published, Vi commented that his approach to a serious subject was really quite profound. The reason, she said, was that he used a bit of humor to effectively describe why all of us, not just doctors, resist changing our behaviors even when the “evidence” indicates change would be a good thing for us personally. I strongly encourage everyone to read the article and to review the Figure at the bottom of page two, “Barriers to Physician Adherence to Practice Guidelines in Relation to Behavior Change.”

How do you discuss Evidence-based Medicine with your community? Where do you see the role of executives starting and stopping in regard to bringing EbM to our institutions?

[Flesch 47, Flesch-Kincaid 12 ... elaboration in an upcoming post! ]

/Media/BlogTopics/2005 Bormel - From Crisis to Confidence, Creating High Reliability in Healthcare.pdf

Topics

Comments

In the article: "Evidence based medicine: what it is and what it isn't: It's about integrating individual clinical expertise and the best external evidence"
Sackett DL, Rosenberg WMC, Gray JAM, Haynes RB, Richardson WS.
BMJ 1996312(7023):71-72.
The authors are saying: "Evidence based medicine is the conscientious, explicit, and judicious use of current best evidence in making decisions about the care of individual patients. The practice of evidence based medicine means integrating individual clinical expertise with the best available external clinical evidence from systematic research." The challenge is not about the use of guidelines, for me the challenge is harmonizing or integrating the individual clinical expertise with the best references available in the context of the unique situation of your patient here and now. The challenge is about balancing the expected outcomes with the specific value system and priorities of your patient, within the professional norms of your practice.

Addendum:

I just attached a file to the original post above. It was a HIMSS presentation I did in 2005. It provides a 'Systems' level, overarching framework, into which EBM is the glue at the insterstices.

The speakers note contained are pretty rich and worth reviewing.

The file name is "2005 Bormel - From Crisis to Confidence, Creating High Reliability in Healthcare.pdf"

Yes, thank you!

The issue is far broader than EbM—it is really all about standardized care. Namely, forming consensus on how to provide care when the data is not clear. In fact, where the evidence is clear, e.g. ASA for chest pain, and monitoring and incentives have been established, performance has vastly improved. Compare the rates of compliance now with 10 years ago—stunning changes. Not without grumbling, but change nonetheless.

However, those EbM practices realize only a tiny fraction of the potential for improvement in care available if we actually apply the principles of continuous quality improvement to how we provide care. To evaluate the utility of different approaches to care, each case can't be a "custom job." We need an agreed upon approach, or two, data collection, and comparison of outcomes. Of course, there will always be some outliers whose individual circumstances will require plan modifications, but that is true less than 20% of the time.

Variation in care is the big problem. If there are standard protocols and ongoing data collection, it is easy to introduce a change and see if things improve. With standard protocols, everyone—docs, nurses, pharmacists, patients—know what to expect, fewer things fall through the cracks, and patient-related variances become clear earlier.

I think we've focused on EbM in part because it seemed like low hanging fruit. It was, but time to get to the hard stuff. So, my answer to the question is "what areas of care best lend themselves to standardized protocols?"

I find your thoughts about EbM very interesting, and the comments to date helpful. In the long-term, Rich's comment about "standardized care" appears to be right on target.

But you need to start somewhere, and EbM appears to be the right place. That's where my organization sits at the present time - square one.

I'd like to help us move forward. To do so, I thought you might be able to provide me with some additional intel.

In particular, if you could expand a bit on your statement, "Something one notch less than EbM . . . " that would be very helpful. Thank you.

I am a project manager for a company that is implementing an EMR with a Health Consortium in Southeast Alaska. The organization has not designated one Provider as CMIO. Those duties are being performed by the Medical Staff Leadership participating in a Transformation Expert Group.

They seem intent on implementing EbM principles during their transformation. It appears that the process has been underway here for quite some time, well before this current transformation project. I am impressed. I see interesting things in the future here.

Recently, Shahid Shah's blog made reference to this topic and my name, here.

It reminded me that I never shared explicitly following, which I believe was intended to be funny:

 

Michael,

Thanks for chiming in. You provided the best articulation of a word we have not used here yet, 'personalized' medicine. It's definitely part of the BROAD EbM as described above. Yet, the use of the word is changing.

Gartner is now using the word 'personalized' medicine the way you describe it, i.e. treating patients as individuals with the best EbM.

In contrast, words like customization and personalization in the recent past of the evolution of EMRs has meant content specific for each provider. That can mean more inappropriate variation. It can also mean very appropriate user interface variation, i.e. terminology and workflow for individual or classes of providers. This can turn a hard to use system into a usable one.  For example, a nephrologist doesn't need an alert that his patient has renal failure or has a low albumin; most of their patients do, and the alert produces no value.  The nephrologist does need order sets that take this stuff into account, in ways that the larger organization often does not.  So, that's personalization for the provider.  (In distinction of personalizing care for a patient, taking into account clinical features, patient preferences, and relevant provider issues.)

So, Personalization-for-EbM, as you describe, is an important attribute of EMRs. EbM; More and Less!

EbM: Is it's coverage TINY or BROAD? Is it a Super-text or a Sub-text for improvement?

I'm thrilled with the comments. Rich, Francois, and David each have a decade or more of intense, healthcare informatics practice experience. They raised the highest-order issues around EbM.

Rich clearly brings out the perspective of EbM as a sharp, narrow-context instrument for driving measurable improvement. Rich also breaks the code on what great executives and managers really want when they use the term EbM. "EbM as a tiny fraction of practice." [... so choose that fraction wisely and recognize the bigger picture.]

Francois clearly brings out the broad challenge a more expansive than fraction goal (the hope that EbM will someday, somehow cover all decisions), and the even higher goal of using evidence wisely, including patient preferences in decision making. "EbM as a broad, high ideal." ["a" and not "the" highest goal, because it co-exists with choice and local realities, including economics.]

David brings out two humble observations: 1) EbM transcends HCIT the desired change(s) existed before the current project at every organization and will exist after any initiative and 2) change is often slow and is very sensitive to leadership practices. "EbM as an evolution in practice." [Leadership and learning are the supertext to the subtext of EbM.]

Jack please allow me to chime in on your point about and also to share my thoughs....
Today, the healthcare industry impact is as great as the Industrial Age years back or the emergence of the Internet. This time, one of the revolution concerns evidence-based medicine, or the ability to screen, diagnose, and treat patients as unique individuals, not statistics. Evidence -based medicine holds the promise to change medicine, and therefore human health, dramatically and forever.

The challenge that health systems faces is how to facilitate healthcare decisions by getting the right information in the right form to the right person at the right time. Infrastructure and process are key issues here.

In this country, the absence of health information infrastructure and serious deficiencies in existing infrastructures is a clear barrier. The support for evidence —based medicine revolves around system standards-based interoperability and the reengineering of processes that are inefficient and counterproductive, yet firmly entrenched. The technology exists to solve these problems, but the challenge becomes ever greater as information proliferates at unprecedented rates.

Thanks for asking, Jack.

The short answer [what is one notch less] is facilitated arguments.

So, suppose you know that treatment XYZ is proven to be safe, effective and appropriate for a patient who has issues A, B, and C (that is, ALL three of them). But, your patient has A, B, and does not have C. Where is the evidence based medicine that's indicated. Formally, there isn't any, in this example, as described.

"One notch less" would be a system that give you specific help in that scenario. There are a variety of approaches, brittle as well as more robust, that are designed to deal with this.

In one way or another, they argue with you, constructively. Aside from the direct EBM content, consultatively get you to your options and their limitations. The recommendations are less evidence-based, but are better than no alternatives when the EBM doesn't exist for a specific case.

These approaches aren't new, and they are proven to enrich the ability of CDS to find 'potentially' relevant EBM.

If you're of a theoretical bent, you can read about one specific approach here:

http://en.wikipedia.org/wiki/Dempster-Shafer_theory

When applied to CDS for EBM, it is used algorithmically to assume that some of your beliefs, i.e the patient has issues A, B and not C, are wrong. Using that approach, the algorithm would consider the patient had A, B, and C issues, find the relevant EBM, and then tell you why it doesn't apply. Obviously, this is a stronger search result than providing nothing.

Does that make sense and help?

Joe Bormel

Healthcare IT Consutant

Joe Bormel

@jbormel

...