Skip to content Skip to navigation

Part II: Setting Expectations Works For GPS, Why Not EMRs Under Meaningful Use?

September 21, 2011
by Joe Bormel
| Reprints
EMRs must have depth to create value and drive adoption

Part II: Setting Expectations Works For GPS, Why Not EMRs Under Meaningful Use?

(Part I here)

Going back to my analogy in Part I, having a GPS for guidance while driving, even when its recommendations are not followed or even needed, is a good thing. It may lead to the safest and best way to reach your destination, but at the very least, it provides choices to make a more informed decision. For example, a GPS can monitor and communicate traffic ahead en route, information that is often not available through other means.

This is certainly the case when the GPS identifies the optimal path or road, but does not actually meet the needs the driver has in mind. As we all know, GPS isn’t infallible.

Recently, I was driving to an unfamiliar destination. Before starting out, I checked a Smartphone app (Yelp) and found that several folks commented they ended up driving in circles trying to find the same location using their GPS.

It turned out that three separate roads seemed to converge on the address. However, that’s because GPS interprets locations in two dimensions. In reality, two of the routes placed drivers on roads that passed over their destination, but didn’t actually intersect it. If GPS worked in three dimensions, that would not have happened.

In working with my team on our EMR, I’m reminded that a “three dimensional approach” is a necessity. In simpler terms, the system must have depth to create value and drive adoption.

We need to create system expectations based on this type of mindset as it applies to clinical decision support. This is true for Stage One MU, and it's especially true given what we know about Stage Two and beyond, where the care plan, goals, and prevalent clinical conditions will take a more central role.

EMRs will not have adequate documentation for results of diagnostic testing to make fully informed recommendations in every instance. That’s a given. Furthermore, it's become increasingly clear that some electronic recommendations, offered in a broad brush fashion, based on incomplete (GPS two dimensional) knowledge – perhaps driven by a single problem on the problem list – will need to be interpreted by the clinical provider to achieve an adequate decision-making process.

Therefore, as I wrote earlier in this blog series, similar to GPS, the EMR and its recommendations are needed decision support for physicians who still make the final decision. This needs to be an expectation set for providers, as well as nonclinical senior executives, designers and architects of EMR systems and, perhaps, even patients.

In the era of Meaningful Use, we’ll need systems that, in a respectful manner (see my previous Butler Model post, are helpful, and that in some ways mirror but transcend the performance capabilities which previously existed in the paper world.

Although some share this expectation today, it’s all too easy to dismiss the EMR, clinical aids, and their workflow role in the era of Meaningful Use because of inappropriately set expectations. However, in the era of Meaningful Use, EMRs will need to be used universally. So as I return to work with my team on the evolution of EMR technology, let’s also commit to setting the proper expectations for all. Stage One awaits us.

What do you think?

Joe Bormel, M.D., MPH
CMO & VP, QuadraMed

Part I:
Part II:
Previous post:

Civilization advances
by extending the number of important operations
which we can perform without thinking of them.

 - Alfred North Whitehead




Doc Benjamin,

Thanks for your comments and kind words.

Your first question related to the bounds of tolerance and role of government mandates. My colleagues who are generation older say often and with confidence, the only changes that happen in healthcare are secondary and in response to payer mandates. There's a lot of evidence for that. In balance, there have been campaigns, like the IHI's 100 thousand lives that brought about rapid response teams, SBAR communications, and other badly needed standard practices. My direct observation is that those leaders were literally in pain from the status quo. They didn't see provider organizations or payers or pharma stepping up to enough major innovations that were meaningful to them.

I think it's important to appreciate HCIT adoption is an evolutionary path, with dependencies that keeps change slow. They include investment dollars. They include moving from handwriting on paper of unstructured and uncoded messages to an ill defined audience, to semantically interoperable communication built on a reference information and interaction model. That last step conservatively requires 25 to 50 years.

Thomas Friedman in The World is Flat reminded us that a century ago, when it became possible to use small electric motors in factories, rather than extensive belt systems, transitioning to the new technology took fifty years. Despite availability, lower cost, better flexibility, etc. My point: we can influence the pace of evolution, to do so, we need to understand and address the drivers. I need to pay off that gamma knife before we obsolete it, or be prepared to write it off.

On your second point, i.e. the impact of industry consolidation and how that relates to expectations, that's actually more chaotic than may appear. Consumer economics have historically not brought dramatic reductions in HCIT prices. Since the average selling price and costs of ownership for enterprise platforms are still and will continue to be in the millions of dollars, switching systems is still a huge barrier. HCIT platform and domain complexity, as well as long term uncertainties keep choices limited.

The bottom line for me is that healthcare and healthcare delivery systems don't behave economically like most other industries and businesses. Cost, quality, and outcomes expectations are starting to move more rapidly toward franchise business models, from the cottage industry model most of us grew up with. The nature and use of EMRs are changing accordingly. Meaningful Use measures capture those dimensions with striking clarity.

Setting and maintaining appropriate expectations of ourselves and those we serve will continue to be critically important work as we evolve.

Dr. Bormel,
This has been a good series. I thought Mr. Land made a solid point, as did you in the close of your reply to him where you stated, ". . . leadership must create awareness of the challenge and the need for tolerance of imperfect technologies."

That led me to two thoughts and questions. My first is, EMRs have been evolving, particularly in terms of CPOE, for quite a long time. Yet, there are still considerable obstacles to their full acceptance due in large part to their overall performance in terms of value for CDS to improve patient care and safety.

I know vendors have worked hard, especially in recent years, to meet what may not always be reasonable expectations but realistically, the systems still have a long way to go. That said, the pace of the evolution seems to be much too slow. So, also realistically, how long do leaders of provider organizations and the end users need to be tolerant before frustration overtakes an acceptable level of satisfaction considering the cost of these systems, and do you see mandated government standards as the ultimate motivator to resolve the challenges?

Second, although there are considerable opportunities for companies to compete for EMR business, the growing number of vendor mergers and acquisitions appear to be significantly reducing the system choices in the marketplace, which is contrary to what one would expect in such a business environment. Do you see this continuing, and will it force providers to simply accept "what they can get?" If so, that would seem to further complicate the issue of setting reasonable expectations.

Keep up the good work!

Doc Benjamin

Joe,  Interestingt post and very relevant.  Even useful!

The GPS:EHR analogy hits home. Not totally analogous, of course, but enough parallels to aid insight and appreciation.

Couple of thoughts. 

First, although the technology behind the GPS and the EHR are the sine qua non, neither instrument is really about technology at all.  Both are about getting you to a destination.

Second, the evolution of both technologies are similar.  Both need serious pieces of infrastructure in order to function, and that infrastructure is essentially invisible to the end user.  As the infrastructure grows in size, speed and complexity, the more value is obtainable from the tool.  Sometimes a leap forward in the infrastructure enables new functionality in the tool; sometimes known needs of the tool drives infrastructure development.


Thanks Frank for the kind words and your perspective.

You are spot on. I don't know what all the dimensions are but certainly the dimension of time is critical, as is the dimension of certainty. Both are pretty much fixed in a GPS.

The existential questions, "who/where am I?" and "where am I going?" are very relevant, as you point out. Related to your cognitive approach issue, I would add the notion of chaos theory. Sometimes knowing where you are and where you're going aren't terribly important if you know two other things:

1) what are the boundary conditions of the system, and

2) are you in a system that is fundamentally stable or unstable.

If you are far inside the boundaries of a stable system, you probably don't need to take heroic actions. Control charts are a great example of a display metaphor that communicates those attributes.  The medical acronym, OOC, which stands for "out of control" is used to modify a clinical condition, like diabetes out of control, or hypertension, OOC.  Again, we dont generally expect clinical decision support systems to reason in that space, but give them time!

Thanks again, Frank.

Dr. Joe,
Interesting analogy, I particularly like the part about willingness to try new things, but I have two quibbles:
1) As you so correctly point out GPS is two dimensional (although with an altimeter it can easily handle 3 dimensions.) I believe medicine is at least 3 dimensional and probably more like 4 or 5 dimensional if you consider the mental and other aspects.

2) More importantly the GPS assumes you know where you want to go, although you do not need to know where you are. It'll quickly figure that out for you. In cognitive medicine all we know is general state we'd like to be in, and for much of the time we sure do not know where we really are. As you know well, the basic process that underlies cognitive medicine is R/O or 'rule it out'. Now just try that approach with your GPS and you'll spend a lot of time driving in circles.

Keep up the good work,

Frank Poggio
The Kelzon Group

Thanks for your comment.  Your observation highlights one of the big misunderstandings in this field.  Not only do we need to set different expectations; there has to be a recognition that the technologies are evolving.  That requires use of imperfect solutions.

Here are some additional thoughts on that idea of use under imperfection:

1) Matt Bennardo in a Lexi-Comp blog recently wrote  ( ) that dose checking might be far more important than interaction checking, and implies that many of today's alerts are not "intelligent" enough, contributing to the alert fatigue problem.  He is, of course, exactly right.  

Systems to check and improve dosing aren't "good enough."  If their evolution requires tolerating and learning from imperfection along the way, then it's the path we must take.

2) From the 2005 Gandhi study:  "Prescribing errors occurred in 7.6% of outpatient prescriptions and many could have harmed patients. Basic computerized prescribing systems may not be adequate to reduce errors. More advanced systems with dose and frequency checking are likely needed to prevent potentially harmful errors."  Eliminating handwriting errors is helpful but far from sufficient to reduce the error rate.  Advances are needed.

3) Setting expectations for what happens when we don't use electronic aids:

    a) from Patient Safety Solutions:   "[turning off dose checking] … resulted in a massive overdose of sodium to a premature baby. That IV compounding machine apparently had alerts capabilities but these had not been activated."

    b) Leapfrog's Evaluation Simulator:  the safety issues of EMR CDS inactivation has been extensively studied, reviewed and discussed.  I've reviewed the issue and provided links here:  This work by Metzger, Welebob, Bates, Lipsitz and Classen is unequivocal.  As with the recent sodium overdose above, their evaluation work has documented that, in effect, people are carrying a metaphorical GPS and leaving it turned off.

Turning off decision support may result in one-in-a-thousand cases when it's needed, it's not there.  The difference between the GPS and EMR decision support is that lives are at stake.  And in many care settings, as documented by Metzger/Classen and Gandhi, the rates of harm are likely much, much higher.

My conclusion is simply this:  When we put computers into any process, there can and will be impact on accuracy and productivity.  To expect accuracy of a system to be nearly perfect and productivity to always be improved as a prerequisite to adoption is both common and short-sighted.  If we are to set appropriate expectations,  leadership must create awareness of the challenge and the need for tolerance of imperfect technologies (as opposed to turning them off).