Robert Wachter, M.D., author of Digital Doctor: Hope, Hype, and Harm at the Dawn of Medicine's Computer Age, recently spoke with Healthcare Informatics' Senior Editor Gabriel Perna at length about the book in a two-part interview series.
Wachter is a professor and the associate chair of the Department of Medicine at the University of California at San Francisco (UCSF).
Part one of the interview featured Wachter talking about how the book came to be, how he was able to get so many interviews, and some of the people he enjoyed talking with while writing it. In part two of this interview, Wachter addresses the meat of the book: the critical chapters on alert fatigue, Aetna vs. Epic, and healthcare IT policy. Below are excerpts from the interview.
The story that got me interested in reading the entire book was the tale of Pablo Garcia and his 38-fold overdose of a drug called Septra, thanks to issues with alert fatigue. Obviously, that happened in UCSF, what fascinated me was that your organization allowed you to write such an in-depth tale of a major screw up. How did you convince them to do that?
It never would have worked if I were younger, newer, and hadn’t been hanging around here for 20 years. Doing stuff that's pushed the envelope has mostly worked out. I've had series in medical literature and wrote a book on patient safety where we talk about medical mistakes and did them anonymously, paid attention to HIPAA protections, paid attention to the right people. The organization believes in me and believes when I choose to do this my motives are in the right place, and I'll do this fairly and carefully. UCSF believes that’s been a good thing for the organization, for our patients, for the people that work here. I come into it with a little bit of street cred, which helps.
When I heard about that case, I was sitting at the root cause analysis meeting, I was flabbergasted. I thought, "My God this is the perfect case to illustrate how our hope that computers would make plow through medical mistakes has been partly realized but has created a new classes of errors." I loved how the case fit in the dominant paradigm of patient safety, the Swiss cheese model. It's never one thing, it’s 10 things. It’s the technology plus the people, the human factors, the culture.
What was so beautiful about this was that the part we thought were fixes created new problems. "The computer can alert you," doesn't work when you get two million a month. Replacing a human with a robot is fine for medication distribution, but when it’s told to do something crazy, it doesn't object. You had a young nurse on an unfamiliar floor seeing an order that is in retrospect completely ludicrous, and she knows that, but she figures to get to her it had to go through a robust safety system. She checks with the barcode scanner, which signals that the order is correct. So that extinguishes her concerns over the bizarreness of the order. It was so beautiful in every dimension, I couldn't have made it up. So when we're at this meeting, the head of risk assessment, the lawyer whose job is to prevent us from getting sued, is sitting next to me and I said to her, "This case is perfect, I need to write about it." She quite appropriately said, "Oh my God really?"
She told me I'd need to anonymize it but she didn’t reject it out of hand. She told me to get permission from the CMO, who told me I needed permission from the CEO. I asked the CEO about it and he was mulling it over. He and I were at a meeting discussing the case and it became clear that the case was fabulous in spelling out the challenges in technology, the human/technology interface, the culture, all that. He looked up at me. I saw I had gotten an email from him and he said, "We have to publish this."
I think it’s a remarkable act of organizational bravery. There are many, many places in the country that would not have let me do it. Part of it was me because I had done this sort of thing before and it worked out. The fact the kid didn’t die, had he died I don't think I could have gotten permission. That was dumb luck. He could have easily died but it turned out this antibiotic is not the most dangerous one in the world. They all told me it would be great if I could get permission from the patient, which I did. There was a lot of bravery. The mother allowed me to interview her and her son. The most the brave people were the doctor, nurse, and pharmacists...they asked me to change their names...I could see them never wanting to talk about it again. They all said, if anyone could learn from this experience and decrease the chances of it happening again, then it's worth doing. I'm proud of them and proud of my organization for letting me doing this. Not a lot places would have.
What do you think Pablo’s tale tells you about health IT systems?
I think it says if you approach IT as a technology project, you're getting it wrong. We have the best computer system that money could buy. We did not fully appreciate the degree to which these are real live humans put up against this technology. If you don't understand the way they think and how they communicate...you’ll get it wrong. I didn’t understand that fully until I saw that case. You have to put yourself in the provider's shoes. If you are an engineer at a vendor designing alerts, you don't get that. The only way you do is if you get down to the factory floor and see how your tools are being used. This issue of automation complacency is one that every industry deals with. It prevents all sorts of errors but it causes us to turn our brains off and not trust our own instinct. This was complete predictable. In retrospect, as I said in the book, we were bound to be disappointed because we're used to pulling out an iPhone and downloading an app and you're good. We thought it'd be that easy but that's the furthest thing from the truth. It's transformed the way people think and the way they work. If you don’t understand that at a deep level and reimagine the work at a deep level, you get stuff wrong. That’s what that case taught me. It's not an easy lesson. We spend our $300 million, Epic is put on your computer, and you want to believe you’re done. But you just started.
Speaking of Epic, Epic and athenahealth was another interesting chapter. I liked how you contrasted the two companies. You talked about there being room for both in the industry. Are you worried that if Epic continues to dominate like it has, the industry will suffer from lack of innovation?
That may have been the most fun chapter to write. They were such vibrant individual and corporate personalities. Sometimes you write through contrasts. I did the same thing with Abraham Verghese and Vinod Khosla. Having two personalities be the spokespeople for a particular point of view. The chapter [with Epic and athena] was the same thing. The old way of approaching things and the more modern way of approaching things. The fact that they don’t like each other, and Jonathan Bush has this vibrant, over-the-top personality, it made it immensely interesting to think about and write. I came out of it deeply respecting both companies. Both produce the best products out there. I chose them because I consider them the leaders. I didn't want people to come out of it and say, "Epic sucks, athena sucks." Those are the two best products out there for the two things they do. And they do very different things. It’s an open question who the winner is. At the end of the day, if you look at the history of technology, you'd be betting on athena or an athena-like company that had a more modern sensibility. Ones that think about openness, thinking about the cloud. The history of technology you'd say is Microsoft gives way to Google gives way to Facebook gives way to Snapchat. It moves in the direction of modern. On the other hand, I came to understand that the act of creating an app store for your iPhone and the act of wiring a 700-bed academic medical center are like putting together a paper airplane and flying to the moon. They are so different. The reason Epic is winning right now is their corporate philosophy, which is 'We control everything and will build a wraparound product that does everything you need to do.' It's a winning formula for the next 5-10 years. Beyond that, who knows. The issue of vendor lock is a problem because we do want competition. I can tell you now that we have put in Epic at my place, a competitive doesn’t need to be 50 percent better before we switch, it needs to be 2000 percent better. It’s expensive.
Epic was the best product out there when the $30 billion became available. It wasn’t a conspiracy; they have a business model that works well. Where I worry is if that model turns out to not be the right model for the future. There are so many places that are stuck on an old model and can’t get out. That's harmful. You need a level of nimbleness so when new products come along that are better, they win. There is a level of obstacle there because of the high-cost of switching.
You were keen on the government moving beyond meaningful use. Where do you see the government’s role in regulating health IT systems in the future?
I think the $30 billion turbocharged the level of government engagement and it somewhat was artificial, although artificial in a way that's understandable and wise. They were worried they'd give out $30 billion and it would be abused. Had they not done that it would have been responsible. But now the money is spent. It is 2015 and the technology has moved on and it's a different world.
I have this thought experiment, let’s say there was never a $30 billion and you were starting today. What would you want the government to do and what would you want them not to do? When you look at meaningful use Stage 2 in particular, I came to believe that’s too deeply in the weeds. It's too much government control of the nature of the technology tools, what they do, what look like, etc. There were no terrible intentions but we are where we are. The question we need to ask is what is the unique role of government? When only government can do something that creates a public good, you want them to you want to do it. When the market can do this though, the market needs to take over. I think that's where meaningful use needs to recalibrated.
I think interoperability doesn’t happen unless the government makes it happen. It won’t happen naturally. They need to have a role there. They need to absolutely have a role in privacy standards. But in terms of patient engagement tools and what your EHRs do and what they don’t do, it needs to have a light touch. If patient engagement is a good thing, and it is, and if health systems are being incentivized to provide high-value care, which they are, then they’ll figure out we need patient engagement tools. You want them to figure it out and to have the tools evolve organically.
Your vision for health IT was interesting. What do you want people to get out of your book when it comes to the future of IT tools being used to aid healthcare?
I want people to understand that implementing IT tools in healthcare is complex. In order to reach the potential of what we want to do in healthcare, we have to think about these tools in a new way. We can't just think of them as technology tools, but as tools that transform the nature of work we do and the relationships we form. To get the most of these tools isn’t just a technology improvement, it's an act of re-envisioning the entire healthcare system. My effort with the book was to generate a conversation about why these tools aren't reaching the potential we had for them, but to do that in a way that isn’t a total downer. We have to go forward, but unless we go forward in a new way and think more deeply about the nature of the work and what we're trying to accomplish, I think we'll get it wrong.