Re: "People preferred the [patient] portal over the telephone for getting test results, updating personal information, getting medical records copies,…
EHR Design Talk with Dr. Rick: Keep or Replace VistA? An Open Letter to the VA 4/24/17
Mr. Rob C. Thomas II
Acting Assistant Secretary & Chief Information Officer
US Department of Veterans Affairs
Dear Mr. Thomas:
The decision whether to bring state-of-the-art innovations to the VistA electronic health record (EHR) system or to replace it with a commercial EHR such as Cerner, Allscripts, or Epic will have far-reaching and long-term repercussions, not just for the VA, but for the entire country’s healthcare system.
Several years ago, when Farzad Mostashari was head of ONC, I attended a conference (see post) where he stated that when talking with clinicians across the country, the number one issue he heard was that their EHR was unusable, that "the system is driving me nuts." After his presentation, we had the opportunity to talk. I asked him, given the dominant market share (nearly monopolistic for hospital-based EHRs) that a handful of EHR vendors were in the process of acquiring, where would innovations in usability come from? His answer was that they would come from new “front ends” for existing systems.
In your deliberations, I would urge you to consider how innovative front end EHR user interfaces, based on the science of Information Visualization, could improve our country’s healthcare system. The field of Information Visualization systematically designs interactive software based on our knowledge of how our high-bandwidth, parallel-processing visual system best perceives, processes, and stores information. Stephen Few describes the process as translating “abstract information [e.g., EHR data] into visual representations [color, length, size, shape, etc.] that can be easily, efficiently, accurately, and meaningfully decoded.”
Sadly, while EHR technology has almost totally replaced paper charting over the past decade, not much has changed in EHR user interface design. For a number of reasons, the major EHR vendors have not made it a priority to develop better front ends based on principles of Information Visualization. The adverse consequences for physicians and other healthcare providers, for patients, and for our entire healthcare system are immeasurable. An Institute of Medicine Report found that current EHR implementations “provide little support for the cognitive tasks of clinicians . . .[and] do not take advantage of human-computer interaction principles, leading to poor designs that can increase the chance of error, add to rather than reduce workflow, and compound the frustrations of doing the required tasks.”
A well-known example of an EHR user interface design contributing to a medical error is the 2014 case of Mr. Thomas Eric Duncan at Texas Health Presbyterian Hospital, where there was a critical delay in the diagnosis and management of Ebola Virus. No doubt, this case is just the tip of a very large iceberg because most major EHRs use similar design paradigms (and because many medical errors are never reported or even recognized, and even when reported, are rarely available to the public). In the most comprehensive study to date of EHR-related errors, the most common type of error was due the user interface design: there was a poor fit between the information needs and tasks of the user and the way the information was displayed.
Furthermore, current EHR user interfaces add to physician workflow. A recent study found that nearly half of the physicians surveyed spent at least one extra hour beyond each scheduled half-day clinic completing EHR documentation. In addition, current EHR user interfaces frequently fail to provide cognitive support to the physician.
Innovative EHR user interfaces, based on principles of Information Visualization, are the last free lunch in our country’s healthcare. EHR usability issues are becoming increasingly recognized as a major barrier to achieving the Triple Aim of enhancing patient experience (including quality and satisfaction), improving the health of populations, and reducing per capita costs. Well-constructed EHR user interfaces have the potential to improve the quality and decrease the cost of healthcare while improving the day-to-day lives of physicians. In my opinion, a well-designed EHR user interface would easily increase physician productivity by more than 10 percent, probably by much more, while reducing physician stress and burnout.
On the design front, innovative EHR front end designs, based on principles of Information Visualization, are already being created by a number of research groups, including Jeff Belden’s team at the University of Missouri (Inspired EHRs). See also my design for presenting the patient’s medical record chronologically using a dynamic, interactive timeline.
In addition, technological advances in computer processing speed and programming language paradigms now support the development of a comprehensive, open source library of interactive, dynamic Information Visualization tools. In this regard, see the work of Georges Grinstein and colleagues at the Institute for Visualization and Perception Research at UMass Lowell.
The beauty of building new front ends on top of existing EHR data bases is that the underlying data structure remains the same. This makes the design much easier to implement than if the underlying data base structure and software code had to be rewritten. Fortunately, all of the EHR systems being considered by the VA, including VistA, have excellent and robust underlying data base structure and organization.
The question then becomes, which EHR system is most likely to embrace intuitive visually-based user interface designs and make these designs widely available? In my view, the clear winner is VistA, for the following reasons:
- VistA, unlike the other for-profit vendors, is government owned. Its goal can be to improve the VA’s and the country’s healthcare system.
- VistA became a world-class EHR through its now famous open source model of distributed development, incremental improvement, and rapid development cycles. Using this same model, visually-based cognitive tools for the EHR could be rapidly created, developed, tested, and implemented. Commercial EHRs do not use the same development model and their development cycles are typically much longer.
- VistA is the only EHR in contention which is open source. Any innovative user interface designs developed in VistA would be freely available to commercial EHR vendors and third-party developers and would thereby benefit our entire healthcare system.
- A major federal health IT goal is for EHRs to “be person-centered,” permitting patients to aggregate, organize, and control their own medical records, regardless of the sources. Innovative user interface designs developed in VistA could, with modification, serve as the basis for an intuitive, open source patient-centered medical record.
If the VA’s goal in selecting an EHR, both for the VA and for the country as a whole, is to improve health outcomes, reduce costs and errors, and improve physician satisfaction, then VistA is the clear choice. Any other choice will set our country’s healthcare system back decades.
Rick Weinhaus, MD practiced clinical ophthalmology in the Boston Area until 2016. He writes on how to design simple, powerful, elegant user interfaces for electronic health records (EHRs) by applying our understanding of human perception and cognition. He welcomes your comments and thoughts on this post and on EHR usability issues. E-mail Dr. Rick.
There is so much that is right about this post. So why then do I find myself disagreeing with the message?
Highlighting EHR Usability? Good.
Farzad Mostashari quote? Good.
Stephen Few quote? Good.
If all of that is true, and if the matter of EHR usability is a general problem, then why is this open letter addressed to “Acting Assistant Secretary & Chief Information Officer, US Department of Veterans Affairs”? And why is VistA singled out?
I submit that this is little more than a rearguard action to attempt to save VistA. The future of VistA is clearly in doubt and this is an effort to save VistA by piling on the issue of EHR usability.
What may be worst about this open letter, to me anyway. There is an implied message that VistA may have the same EHR usability issues as all those nasty, bad, “dirty” closed source EHRs. Which merely reveals the open source issue to be the canard deployed here. If VistA has EHR usability issues, that is a liability, not an asset!!
Our expectation must be that ALL leading EHRs improve in usability. Those that do not should be kicked to the curb.
Dr. Weinhaus’ conclusion is as flawed as his proffered evidence. The Texas Health issue with Ebola was tracked back to medical error on the physician’s part.
The EHR showed the doctor the travel status at least seven times on a clean and uncluttered screen if a brand new patient. That doc failed in diagnosing the patient ordering a wide variety of tests including a cat scan.
The real evidence provided under oath to the US Senate conclusively proves that as did THR’s subsequent mea culpas.
Rick, just because you graduated med school doesn’t make you an expert at UI. If you were you would have researched this issue thoroughly. You didn’t.
We’d all agree that narcissism is not a great design process by the VA a vendor or yourself. Vendors however stand judgement by their customers every day. I’d take that a million times over your confused ramblings.
To Fake News and W:
Thank you for your comments.
My point was that current EHR user interface design is a major contributor to medical errors. The paper I cited, which stated that EHR usability issues contributed to the delay in diagnosis of Ebola virus, was written by Divvy K. Upadhyay, Dean F. Sittig, and Hardeep Singh, three leading researchers in the field of Medical Informatics. In the paper, they found that “current EHRs lack the innovations needed to prevent misdiagnosis.”
Can you provide us with a link or a reference to the screenshots you refer to? There is no question that the travel history was documented in the nurse’ s note. The question is whether the EHR user interface made it hard for the physician to find the nurse’s note and travel history.
To “Fake News”…
If I knew who you were, I’d consider shoving a bottle of my mother’s heart medication that Bad Health IT led to be erroneously discontinued, resulting in my mother’s severe injury and then death a year later, up your distal gastrointestinal orifice … the same orifice you apparently speak from.
The majority of medical societies in the USA have written HHS in no uncertain terms about Bad Health IT. http://mb.cision.com/Public/373/9710840/9053557230dbb768.pdf
It’s tiresome to see Health IT amateurs hyping or making excuses for Bad Health IT, and attacking those who call out the industry for its antediluvian practices as Dr Rick has done.
Re: “just because you graduated med school doesn’t make you an expert at UI”
By the way, I believe a Medical Informatics postdoctoral fellowship and faculty time in informatics at Yale School of Medicine may give me more gravitas to speak on such issues than Dr. Rick.
“Fake news”, did YOU graduate from medical school?
I thought not.
Why wouldn’t Cerner, or Epic, or any other commercial vendor be able to innovate as fast or faster? I’m guessing Vista in it’s time also cost more than the market capitalization of some of these commercial companies.
Actually curious since I don’t know, did Vista need to get certified in meaningful use?
Dr Rick is just repeating and recycling the same old stuff.
Hi Echo chamber:
Thank you for your comments.
The history of how VistA was rapidly developed in adversity and at very little cost is recounted in the book, Best Care Anywhere, by Phillip Longman.
At this point in time, I do not know what the relative cost is for developing innovative user interfaces in VistA compared to the commercial EHR products.
Going forward, in order for VistA to rapidly and economically develop new user interfaces, I would suggest that the VA muster the resources of the open source community. For instance, there could be a competition among engineering and design schools to design new user interfaces, with the best design ideas prototyped and then rapidly tested with a small group of users.
The version of VistA used by the VA is not certified for Meaningful Use. Other versions of VistA, such as the one used by the Indian Health Service, are certified for Meaningful Use.
I agree with a lot of what Dr. Rick has said. However – one point of disagreement. Dr. Rick advocates letting the open source and student community design the UI of VistA (presumably on a volunteer basis).
What would happen if we took that suggestion and applied it in another arena: I don’t think it would be popular to advocate people obtaining cheaper/free medical care from unpaid med students. Wouldn’t be fair to the medical community, right? Doctors deserve to get paid. Furthermore, an experienced doctor is often a better doctor.
At some point, we need to get back to the idea that our current societal model involves paying money for goods and services, which allows the providers of those good and services to:
1. support their families, and
2. focus on their trade – conferring insight, skill and discipline (best case) – and rewarding those qualities.
If you don’t like UIs these days, fine. Indeed it sounds like there is much room for improvement and Dr. Rick has some great suggestions. But let’s not throw the baby (capitalism) out with the bathwater by trying to get our UI design done for free by an army/committee of unpaid newcomers. IMO, this would cause more problems than it would solve.
Agreed. As a veteran and long-time member of our industry, I agree with Dr. Rick.
My VA providers have always said they feel the ability and intuitive nature of Vista adds much to the patient experience as well as the ability to share data across the whole VA System.
As I travel I can walk into any VA in the US give them my ID # and other brief information and get vital Rx filled within 30 minutes.
Save your breath, the beltway companies have been embedded in VA leadership over the last decade and the billions are lined up to be distributed. Look up the last 2 CIOs for VA and see who they went to work for right after leaving. Then look up how much money that company was awarded in no bid contracts after they hired the last CIO!
It is a rigged game! The top EHR in the world is the VA’s VISTA. The top commercial EHRs are a FOIA copy of the VA’s VistA. They are just stripped down and have a different GUI. Now the “VA Leadership” wants to scrap their version so the can buy it back from a for-profit company. I wonder were that leadership will be working in a few years when they leave the VA!
Agree with Fake News. The linked article is a “hypothesis” and IIRC, screenshots showed that the travel history was clearly documented and visible. The physician just didn’t check it.
If Vista was so good, then why hasn’t a commercial company been able to deploy it with wide success, at much lower cost than alternatives?
Ben, I text “refill” to Walgreens and it takes less than ten seconds.
Vista’s architecture is a disaster (sad) and its front end is good, though built for purpose at an unconscionable cost.
Replies sound very much like Commercial Vendor audience.
TO Fake New – Rick, just because you graduated med school doesn’t make you an expert at UI. If you were you would have researched this issue thoroughly. You didn’t.
Is this the best you have – pretty sad.
I guess Marc Probst CIO for Intermountain Healthcare and Larry Wolf of Kindred Healthcare who criticized the EHR interfaces as adding cognitive load, difficult to navigate and are a source of frustration to users were wrong too.
Or maybe Dr. Steven Stack when he spoke of “poorly-designed EHRs have physicians suffering –
Maybe you should stop being lazy and rather criticize you do a little research Mr. Epic or Ms. Cerner or whomever you have aligned yourself with.
Way to go Dr. Rick –
Dr. Weinhaus, the VA has made some great strides in clinical usability and information visualization with their Enterprise Health Management Platform (eHMP) which overlays their VistA system. I would encourage you to look at this great presentation of the system on the OSEHRA web site. Citation below:
McNamee, Shane, MD. “Recording – OSEHRA AWG Presentation – EHMP Activity Management, Shane McNamee, VHA.” Recording – OSEHRA AWG Presentation – EHMP Activity Management, Shane McNamee, VHA. OSEHRA, 23 Aug. 2016. Web. 09 Dec. 2016. .
Also review the powerpoint presentation from John Short and David Waltman at this past year’s HIMSS Conference. Citation below:
Short, John, and David Waltman. “VA’s Data Sharing and Interoperability Strategies March 2, 2016.” VA’s Data Sharing and Interoperability Strategies March 2, 2016 (n.d.): n. pag. 2 Mar. 2016. Web. 9 Dec. 2016. .
Thank you for the information and I followed the suggested links.
First let me state that I want the best system for the vets and the clinicians that serve them. I am not a hater.
I followed the links you provided to the OSEHRA web site and went to the two demo sites provided with access codes and verify codes to see what you were so excited about.
As expected, one of the links did not work and the second one took me to an interface that was a cluster of text combined with blocks that was a pathetic version of visualization. I can see why you would feel it was progress but it does not follow basic rules of visualization but a lazy excuse for a GUI.
In addition, I went to the PowerPoint HIMSS presentation to see the future vision as you suggested. I was very impressed with the vision and would love to see it implemented as soon as possible.
My chief complaint and what I think Dr. Rick was eluding to is we live in a world that I can go to any ATM machine in the world and withdraw money from my bank – true connectivity, however, my 94 year old WW2 veteran father fell at XMAS and wound up in the emergency room for four hours. In March, when he visits his VA doctor for a checkup has to tell the doctor about the fall and what happened because the VA doctor does not and cannot see the event on his healthcare continuum. This sends a big red flag up to me.
Do we really want my dad who cannot remember what he ate for breakfast by noon being the record source for an event at an Emergency Department 3 months earlier?
I do not care who provides the solution my only request – make it work, make it effective, make it efficient and let’s stop giving lip service to what we are going to do and spend the funding correctly.
I think we are all saying the same thing and need to unite.
Please look at this link if you are having trouble with the main OSEHRA links:
This is a direct link to the MP4 file.
I think you will find that this visualization and demonstration of eHMP – which begins at 25:00 in the video – is fairly robust. Particularly the way that a clinician can chose the applets that they want to view as well as the data filtering and graphing criteria.
Hi Jack Shaffer,
Thanks for your comment.
I found Dr. McNamee’s presentation on YouTube. I was very impressed with the flexibility of eHMP, and especially the filtering function, the graphing function, and the ability to do a word search on the entire medical record.
The U.S. government doesn’t care about getting to patient-centered or user-friendly for the VA EHR. Just follow the money. That is what this is all about.
Hi everyone, thanks so much for your comments.
It will take me a little while to respond to all of them, but let me start with Brian Too’s comment:
Thanks for your reply. I am a little puzzled by the points you make.
My letter was written to the Acting Assistant Secretary & Chief Information Officer at the VA precisely because the future of VistA is in doubt. I wanted to weigh in on the matter before the decision is announced on July 1 or before. For me, it is a pro-active letter as opposed to a rearguard action.
I do believe that EHR usability issues should be a fundamental part of the discussion in the decision-making process of whether to keep or replace VistA.
You are correct that I feel that the current version of VistA shares many of the same EHR usability issues as the systems of the major commercial EHR vendors. My point was not to claim that VistA has a more usable product at this time. Rather, I proposed that VistA has more potential to develop innovative EHR user interfaces by using an open source development model that includes distributed development, incremental improvement, and short development cycles.
I do not in any way feel that the EHRs of the major commercial vendors are “nasty, bad, or dirty,” as you imply. Their source code is certainly closed. Also, they are for-profit entities. From a business point of view, they may have to focus on issues other than EHR usability. For instance, meeting Meaningful Use requirements and other federal mandates has been a requisite of staying in business, but it has also drained their resources for many years.
Of course VistA has usability issues, and of course they are a liability, not an asset. I don’t understand your point.
While we have an expectation that all leading EHRs must improve their usability, open market forces should be the mechanism allowing the more usable systems to survive. Removing VistA, the only major nonprofit open source system from the mix will make it much more difficult for open market forces to operate.
Does anyone know what the VA or related federal bodies spend annually to maintain and enhance VistA? Capital and operational expenditures? Different vendors have reputations for front of back end costs, and I wonder how that compares to employing people to maintain open source.
Our government through its very patriotic contractors who donate huge amounts to politicians and who provide immense salaries to federal and military “retirees” have billed US Taxpayers blind.
Add up the costs.
Do the side by side on architecture, feature, function and safety.
Vista no longer stacks up and they avoid direct comparisons at all costs.
More has been spent on VA and DoD EHRs than on all EHR vendors profits combined over the last decade.
In order for VistA to e ready for prime time it must publish a taxonomy and a data dictionary. 130 VistAs are incompatible in one way or another at the data level and therefore communications within and certainly outside is excruciatingly difficult. The DHA’s CHCS at least has solved the data standardization problem years ago. Ask the VA how many update to VistA are made when a major patch is sent? DHA sends just 1.
Hi Jim Lawerence,
Thanks for your comment. I don’t know the details, but it is clear that a universal taxonomy and data dictionary are a necessity.
Hi Ben, Fedupvet, Ben there, Big House, Tonya Wills, VFJ, Sam Lawrence and everyone else who posted a comment.
Thanks so much for your thoughts and insights.
I think there’s a LOT to learn from VistA on many fronts.
I used the book “Medical Informatics 20/20: Quality And Electronic Health Records Through Collaboration, Open Solutions, And Innovation” by VA EHR experts Douglas Goldstein, Peter J. Groen, Suniti Ponkshe, and Marc Wine in many courses on how to do HIT “right” and avoid Bad Health IT.
The HIT industry could learn much from this book, which, in fact, was published a decade ago. However, being the HIT industry, locked into their Big IT paradigms, this book is perhaps too unsettling for them. Even getting past the title may require Safe Spaces.
The “problem” is the “business”… It’s the way we pay for healthcare. Everything evolves around that in the “commercial” space… Until providers and patient care actually “matter” again, none of these things will happen (one of the reasons Rick is on to something…)