Q&A: ONC Deputy Nat'l Coordinator on EHR Usability

Jacob Reider, MD, discusses the usability of electronic health record systems, which was recently lambasted by the AMA. He agrees usability can be improved. At issue is "the progression from where we are to where we need to get to," he says.

The AMA made its latest broadside last week against the appalling state of user interfaces in EHRs. Barely usable EHRs are arguably the biggest impediment to the promised productivity improvements that EHRs were supposed to deliver.

In the wake of the AMA's condemnation, I sat down last Friday with Jacob Reider, MD, deputy national coordinator, during his visit to a patient experience hackathon at Stanford University. Although we also spoke at length about interoperability, this week I will share the usability portion of our discussion.

HealthLeaders: Everything seems pivoting towards interoperability. Is this a blind spot in the current ONC strategy, or do you have any particular sage words of reaction to what the AMA has been saying this week about usability?

Reider: I would first direct you to the blog post I wrote about usability. It is a pretty good expression of ONC's position on usability. Is this a blind spot? Absolutely not. It is probably one of the two reasons that I left industry to come to ONC. Usability is the single most important thing that I have worked on in my career. That attention has not faded.

We applaud the AMA. Steve Stack [president elect of the AMA] and I talk frequently about this topic, so we are on the same page as the AMA. Now, [as for] the cause of some of these usability challenges, we may not agree one to one.

We do feel very strongly that it's a problem that needs to be solved. The question is, how should it be solved, and what's the progression from where we are to where we need to get to be with the user experience activities.

EHR Systems 'Immature, Costly,' AMA Says

I think of user experience as a continuum. It's not something is usable or not usable. It starts out at functional. So the bottom [part of the pyramid] is functional. The Model T Ford was functional. It did it. But it didn't do it every time, because it wasn't reliable.

So the next level is, the system needs to be reliable. The Model T Ford wasn't reliable, but let's say a Chevy Nova is reliable. And yet it wasn't usable in the sense that it anticipated my needs. And there's that little bit of happiness you get when this anticipates your needs, or when something happens that you say, wow, whoever designed this thought about me when they designed it.

Steve Jobs is famous for saying, 'my customers never told me to make an iPhone.' Because good designers give users what they need, which is different from what they ask for. It's a very different thing. And if we really do it well, it's meaningful… It's valuable to my life. It changes my life in some way.

We saw the people on University Avenue standing in line [today for the iPhone 6]. Who's been standing in line to purchase an electronic health record? Nobody, because we're between reliable and usable.

Some of these systems anticipate the user's needs, and we know that clinicians in fact, they expect clinical decisions support to remind them when things that they did are unsafe. So they're now so accustomed to that anticipation of their needs that they rely on clinical decision support to catch their allergies. They're not thinking as hard. So I tie usability to safety.

Simple EHR Function May Trigger Audits, Hospitalist Cautions

HealthLeaders: I have heard conversations recorded at a meeting of physicians who made a really good point. They find it so frustrating that they're in an EHR, in this case it happened to be Epic, and they're having to enter a diagnosis before they're ready to enter a diagnosis for some particular encounter, and they can't continue until they put something in, and that's where the safety issue pops up, because the something they put in could be the wrong thing, and bad decisions could be made later if that something isn't corrected, and there's even some issues about the process by which you go back and correct. That's a kind of problem that Steve Jobs never had to deal with.

Reider: I agree that there are anecdotes like that that are evidence of bad implementation of software into the clinical workflow, and when I hear an anecdote like that, I always peel the onion seven times. I always ask the five whys, if we're going to be lean about it. Because what I find is sometimes it's the design of the software. Sometimes it's the customer. It might be the healthcare facility that these doctors work for, that defines a process that then the software instantiates.

HealthLeaders: In the case of Epic, they may customize to please the customer.

Reider: Or the customer may have customized it locally. So we've seen great examples of products that you might not say are usable or easy to use when you look at them at first. It might look like an old, clunky screen. The old character-based screens in many cases didn't look usable, but once you got [in they were] good, man.

HealthLeaders: I remember at the airline counter what those reservation agents could do with those character-based screens.

Reider: They could do things without even looking at it. So usability has lots and lots of factors, and as you know, there's a science to this. There are people who spend their careers studying this and defining it.

To circle back to your question, what we did in Stage 2 was we required that the health IT developers use some of the best practices that other industries have incorporated for decades: user-centered design. We required that they do it, and that they document it, and that they test their products, and that they post the results of those tests on our Web site.

So you can log into our Web site, and you can look at the usability testing that was done on a set of products, so now, as a customer, you can say, gosh, this doesn't make any sense. That was our first step toward bringing this industry to where other industries have been for many years.

It wasn't the last step, and it's not a complete solution. We haven't fixed it yet.

HealthLeaders: I've heard the NIST usability guidelines may be part of the next step. How does that map to addressing the AMA's concerns this week?

Reider: It's the first step. The first step that we took was requiring [developers] to do things the way that other industries do them, so that the products are safer and easier to use… We didn't tell them what their products needed to look like. We didn't say, here's a picture of a usable product. Make it like this. Because obviously, that's not our job.

But we did say, use the right processes. That's what FDA says when you're making sausage or drugs. They say, use the right process. So we said, use the right process. That was our first step.

And I can't talk about our next step, because we're writing our next step. So we applaud what the AMA is doing. We are right on board with them. The industry acknowledges: We know we're not where we need to be, and we need to do better here.