UX Nuggets Thoughts and advice on usability and user experience

Designing EHRs Starts in the Field

July 29, 2010 |  Bob Schumacher, Gavin Lew

This is the first in a series of posts discussing electronic health record (EHR) systems.  Why a series?  There’s a lot to talk about, including many myths and misunderstandings about EHRs.   These are a few of our thoughts; do you agree?  Disagree? Have other ideas?  Let us know what you think. 

If electronic health record systems (EHR) are as helpful to medical practice as it seems they would be, why have such a small percentage of physicians adopted them?  And why has no single company emerged as a frontrunner in the EHR field?  There’s probably not one easy answer.  The more we thought about it, we came to believe that the problem might lie with the technology-centered view that has dominated the field.  

Most EHR applications are database-oriented.  It’s a linear process – symptom, diagnosis, treatment, and then billing.  The problem is that the practice of medicine isn’t clear-cut.  It’s messy.  Patient/physician encounters are filled with nuances and complexities. It’s not easy to document everything that occurs during an office visit by clicking boxes.  Even paper files themselves – age, thickness, notes -- convey clues that may, even subconsciously, direct a doctor’s understanding.  Subjective impressions can be difficult to capture in an electronic format.  But if EHRs can’t capture this information, they will only be helpful in a limited way.  They’ll handle billing and insurance, but won’t really help the doctors become better at what they do.  Distilling all the information from a patient visit into a database may, in fact, be taking valuable diagnostic tools away from physicians.

So the question becomes, how can we design a better mousetrap?  Thorough research is essential.  Before building an electronic health record system, we need to observe healthcare professionals as they work.  Understand the kinds of questions they ask and the reasons they ask them.  Watch for subjective cues that may inform a doctor’s decision-making process.  Review existing records and have medical professionals explain just how they use them.  Learn about gaps in knowledge.   This information isn’t easy to get at, but it’s essential if we want to avoid turning physicians into mere documenters.  While taking advantage of the benefits of the EHR, we need to preserve the richness and artistry of what it means to be a diagnostician.  The shortcoming in most current systems isn’t with the technology or the database, the missing piece is a real understanding of the physician’s experience and the clinical workflow.  Research is the key to gaining that insight.  

In future posts we will be highlighting how implementing a calculated research strategy (1) involving ethnographic and in situ research to learn about workflow, (2) understanding clinical needs and user needs to create an exceptional user experience, and (3) applying a robust user testing program that would result in a better, more adoptable EHR - one that would not just store data, but would add value to a doctor’s time and expertise.  It would extend the capabilities of medical professionals, help them work smarter, and give them the information they need to make the best decisions they can on behalf of their patients.   We feel that in the end, the system that emerges as a leader in the EHR field will be the one who thoroughly researches the full experience of care delivery, and is able to develop that experience into a usable interface. 


Robert Schumacher, Executive Vice President, has more than 25 years of professional experience in corporate and academic environments with expertise in areas such as global user research, health information technology and contact center applications. He holds a PhD in Cognitive and Experimental Psychology from the University of Illinois at Urbana-Champaign. 

Gavin Lew, Executive Vice President, has 20 years of experience in both corporate and academic environments, giving him a strong foundation in user-centered design and evaluation, with particular expertise in mobile technology, healthcare, and global research projects. He has a MA in Experimental Psychology from Loyola University. 

Contact Us for Answers

Send Message | +1.630.320.3900 | Contact Info

November 11, 2010 - 11:35 Anonymous said:

As someone who reads medical reports generated by physician EHR's on a daily basis, I can assure you there is plenty of room for major innovation, not just in data input, but also in internal data processing and external presentation.

For example, data input needs serious help. And by help I do not mean copy and paste last patient visit notes. Why do so many physician visit note reports appear to be word-for-word the same from visit to visit? Copy-and-paste cloned content is nearly meaningless because it is biased by ease of entry rather than true current observations, which may vary from visit to visit. I would suggest that this makes for less observant clinician as well. Did he really ask the patient all those questions again, yet not have time to enter the data? I tend to ignore general statements that are identical from visit to visit, especially when inconsistent with parts of the EHR which are obviously new (such as "patient will start drug X this week....") Calling the physician sometimes results in a conversation where he or she admits that these glaring errors were just accidental carry-over of information from the previous visit. One wonders how many less glaring errors or missed clinical findings are not detectable. Why is this allowed to happen?

Secondly, why is it that many EHR generated reports provide the name (date and signature) of the dictating physician on the LAST page of the report? Is that the customary way to communicate information? No, it is important to know the name of the person generating the report and the date of the report to place it in context while reading. I would recommend an open source style agreement on formatting of basic document identifying information: patient name, birthdate, age, date, record number. treater name, date and location of service, etc

November 11, 2010 - 22:59 Kathleen said:

Usability for healthcare software is a huge new field. Existing software is mostly pathetic, and healthcare professionals have resisted computers because their software is so un-usable! If you have expereince and skills in UE, and an interest in healthcare, you are sorely needed! Many companies that develop healthcare software don't even know they need you, so you might need to apply for a job like "software analyst" or "IT consultant." Once you're in the door, you will be able to add value if you can help improve usability.

November 16, 2010 - 18:07 Michael Milne -... said:

While I agree with you - a lot of people are shouting it is all about the UI and recognizing with a Medscribbler demo how great it is but few like the authors here even bother to do simple research on the availability of different forms of user input into an EMR other than keyboard.

Try a search for handwriting EMR or Tablet EMR and Medscribbler comes up either directly or in an article. Keyboards can NOT be used effectively by doctors in day to day medicine because they are ^^&^%&^% tools for the job.

December 7, 2010 - 12:22 Gavin Lew said:

This is not a forum for specific vendor support or rejection, especially because the article that stresses the complexity of the patient encounter and how EHRs tend to have a database-oriented view rather than physician focused. It also urges taking the time to get out of the developer cubicles and design using insight from the field.

The authors focus was on the design, not technical features that might improve design. User Centric does plan on conducting user research on technological features that might show promise through our involvement in the Chicago REC (see www.CHITREC.org for more information).

December 23, 2011 - 00:47 Shree said:

I'll never understand how the EMR systems developed are less flexible than their predecessor: paper. There are mixed reviews (because of course, how much staff chooses to deviate from the predetermined values is another thing), but the only consistently good one is perhaps the VA CPRS, the most simple looking but the most flexible. But I think it and other interfaces still need to be introduced to better ucd practices (as you outline steps later) but also better ecological interface design to reach full potential. More prediction and visual analysis running to tell doctors "there is this pattern you may not have noticed" in a neat, non-cluttered way.

Drupal site hosting and development by SiteBasin and Alpha Cube, Inc.