Skip to content Skip to navigation

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

September 13, 2011
by Joe Bormel
| Reprints
Evolving the EMR in ways that make physician adoption easier by developing capabilities that create real value for improving patient care

I work with a team of very talented and dedicated people. We spend much of our time evolving and enhancing our EMR based upon a combination of advances in the technology, user-defined input, and anticipated user needs moving forward. One of the most important goals we have is to evolve the system in ways that make physician adoption and use easier by developing EMR capabilities that create real value for improving patient care and safety.

Occasionally, I find that stepping back from these challenges, to look at them from a different, often less complicated viewpoint, helps me gain a clearer perspective of what needs to be accomplished. It’s this thinking that led me to write the blog that follows.

I see numerous similarities between the effective use of GPS assisted navigation while driving and physician adoption of electronic medical records under Meaningful Use. To me, it’s a matter of setting the proper expectations.

Getting instructions from a GPS can be a little frustrating and even dangerous at times. For instance, while driving at highway speeds, with stressors like heavy traffic, or unfamiliar routes. Even getting directions when driving at 25 mph can present special challenges, such as when those directions conflict with your expectations, common sense, signage or your spouse's recommendations.

Those of us who use GPS on a regular basis know that the directions it provides are not always correct. Therefore, we set our expectations based upon this knowledge, and temper the results based upon our experience, the use of additional tools, and sometimes, intuition.

I’ve heard many reasons not to use GPS . . . it’s distracting, it’s too complicated, it’s time consuming, it pesters me constantly; if it’s not always right, I don’t need it, I know where I’m going, why use it? This is getting even more interesting as GPS assisted navigation has become more sophisticated in recent years, with Internet connectivity and with it, integrated access to up-to-the-moment traffic, nearby services, and communication. My conclusion is that in almost every instance, the right expectations were not set with detractors, so the technology was not accepted as beneficial. The GPS capabilities either under adopted or not adopted. And per the title graphic above, that may be a safety issue.

I find the same is true when it comes to EMRs. GPS does not make decisions for you. It provides decision support to help you make more informed decisions, and beyond what any one human user can know . . . just like an EMR.

Too big of a leap to relate GPS to EMR? Let’s take a closer look. First, a basic framing of their definitions. GPS is a technology that defines a driver’s current location and helps them find a path to their desired final destination, or at least as close as possible.

An EMR is technology that helps a clinician determine “the location” of the patient in terms of their available vitals, labs, and other inputs such as radiology reports, which in turn help to select a path to the desired final destination. That destination being curing the problem or coming as close to doing so as possible.

The technologies for GPS and the EMR have been evolving for decades. In recent years, both have been vastly improved, transcending merely creating clarity on the current location, to helping their users get to where they’re trying to go. Both now have the capability to guide, that is, point out an available course and its specifics. With GPS, that may be turn-by-turn directions. With the EMR, that may be a care pathway, articulated as an order set or sequence of order sets. And with both, each step may have additional details available, such as the traffic on a particular highway segment, or a clinical contradiction to a medication order.

So, is it too big a leap to relate the service of a GPS to an EMR? I don’t think so. When I take a step back from what’s being asked of clinical users in the era of Meaningful Use, the guidance needed at many discrete steps of a clinical process becomes even more recognizable as similar to the guidance provided by the modern GPS. As with all new technologies, there is some initial resistance by intended users. It was true with GPS, and it’s true with the EMR. But as these technologies evolve to create more perceived value, with expectations are properly defined, adoption increases.

This brings us to the challenge of how to drive physician adoption under the criteria of Meaningful Use; a relatively new challenge. It involves " problem-based and problem-list-based workflow" to ensure that appropriate order sets are used, the documentation templates and guidelines are available, quality measures are explicitly available in context, and complete handoff documentation is produced. The hand-off documentation refers to both clinical summaries for patients (an electronic copy of health information in MU context), as well as documents for subsequent clinical professionals to assure seamlessly coordinated care.

One of the interesting, albeit odd twists in all of this has been clinical decision support, and specifically alert fatigue and alert overload. Essentially, a provider organization needs only to include one clinical decision support rule in Stage One MU. In reality, however, problem-based workflow will increase the degree of passive prompting to ensure a faster, easier, and safer end-user experience than the alternative.




Contributing Editor to FierceEMR, Marla Durben Hirsch, responded to the original post here.   She made some intriguing observations.

The notion of an analogy between the GPS and the EMR and their respective forms of decision support is problematic because they are so fundamentally different:










basic data delivery


patient-related observations: labs, vitals, problems, allergies, medications, documents, ...

primary operations


assembly (e.g. face sheets, summaries)

network-delivered services


exchange of clinical summary / continuity of care documents (CCD)



local-based services (find and potentially call businesses such as restaurants, hotels, fuel, etc);


conjoinment with smartphones

recommendations based on CCDs pertaining to existing or proposed diagnoses, treatments, and ongoing management

domain complexity

two-dimensional, finite graphs

simple, complicated, complex, chaotic, and combinations thereof, with temporal evolution, uncertainty, and life-threatening consequences from both underlying diseases and from the treatments

(derived from Snowden and Boone)


obvious advantages to the absence of navigation services

more complicated

necessity at introduction

no (not if you knew a route)


necessity now

yes (even if you know the route)


I'll elaborate this a bit in Part 2, but the thinking is that traditional EMR decision support, like GPS at introduction is pretty basic.  It consists of drug-allergy checking, drug-drug interaction checking, passive decision support of thoughtful displays, flowsheets, templates, and rules.  That said, like the GPS, it already exceeds what can be reliably achieved by any unaided human brain.

As we'll see in Part 2, that does not mean that such EMRs can or should replace humans in the diagnostic or therapeutic process.  There's simply too much missing in terms of deep smarts, necessary data, and true, relevant wisdom.

Again, keep those comments coming!   (most are coming in private email at this point) 


Dr. Bormel,
Excellent analogy, I like it. You make some very solid points about EMR adoption that I think few if any of your knowledgeable readers will dispute.

I would like to take this topic of adoption through setting expectations on a slightly different, but converging path.

In talking with my peers, I find that hospitals have become much better at involving their physicians in the EMR process, but for the most part at the front end. That is, during the procurement cycle, physicians take an active role in system selection, and their buy-in to the project is generally considered to be a necessity.

However, as the implementation begins, the IS people take over with the vendor and consultants. In too many instances, physicians are then relegated to the background until they are expected to begin using the system. This is simply not acceptable, if the hospital intends to get the full support of its docs.

The fact is, buy-in for a large system such as an EMR is an ongoing process, and to exclude the end users from much, if not most of the initial install and implementation is not a good business practice. When this happens, it isn't that the physicians have the wrong expectations, in many cases they have none at all. That will obviously have a negative effect on adoption.

I've found those hospitals that have created a CMIO position appear to have better adoption rates over shorter periods of time. But still, I believe a small group of senior docs, representing all of the organization's physicians, should be an integral part of the entire process to better ensure its desired outcome.

Could you share your thoughts and perhaps, your real world experience with us? Thank you.

Doc Benjamin

Doc Benjamin,

Thanks for your comment. Your comment about the implementation process is certainly recognizable, although it's not just physicians who "are often relegated to the background." The same is too often true regarding nursing, or pharmacy, for the CDI folks, etc.

Another dimension related to physician involvement in building for adoption is not just getting the docs involved, but getting the right docs and setting the right vision.

Perhaps half of the time, especially in community hospitals, one of two unfortunate things happens. First, the physicians who get involved are those who have adopted other systems, are often open-minded, flexible, and willing to change their workflow. These physicians are willing to adopt at least once or twice a decade, although understandably not every year. Hold that thought for a moment.

The second thing that happens is a desire or perceived need to replicate the prior system. For example, physicians often want to access patient information through patient lists, and results in access that more-or-less exactly matches their current reports . As we've discussed previously in this blog, that approach may be discordant with problem-based workflow and other workflow models. This can make achieving meaningful use criteria cumbersome, navigationally too hard (too many non-intuitive clicks) and unsafe (critical results don't get in front of the physician reliably).

The net of all this is that it's critically important to roll out the system well one time, paying close attention to adoption needs for both immediate use, and immediate horizon use cases related to processes such as admissions, discharges, medication reconciliation, and documentation improvement initiatives.

My real world experience seems to match yours. Care providers, nursing, pharmacy, and all of the other departments and therapies need to be at the table throughout the implementation, with a commitment to future-state visioning. This takes an investment of people's time, as well as build resources and associated QA that collectively is extremely uncommon. Again, this has been chronicled by multiple editors and bloggers at Healthcare Informatics. There is an implementation resource commitment that is all-too-often under estimated. And, yes, a symptom can be physicians relegated to the background of an implementation.