The Digital Doctor – The Review

Digital Doctor

Bob Wachter has been about as influential an academic doctor as there’s been in recent years. He more or less invented the concept of the hospitalist, he’s been a leader in patient safety, and even dressed up and sang as Elton John at the conference he runs! (He’s also pissed off lots of doctors by being a recent one year chair of the newly controversial and perhaps scandalous ABIM). But for the last 2 years he’s been touring the good and the great of health care and IT to try to figure out what the recent introduction of EMRs at scale has meant and will mean. The resulting book The Digital Doctor is one of this year’s “must reads” and yes we will have Bob as the keynote at this Fall’s Health 2.0 Conference.

The immersion research he conducted was fantastic. Bob interviewed just about anyone you’ve ever heard of and a few you wish you hadn’t (more on that later). And in fact he’s been running interviews on THCB and elsewhere sharing some of the stuff that didn’t get in the book. But I’m still wrestling a little with what I think about the book itself. And I think it’s because I largely agree with him and his angst.

There is lots of wonderful stuff in this book. The change in the role of radiology post PACS, how patients are using open notes, whether Vinod Khosla agrees with Vinod Khosla about algorithms replacing doctors–all this and much more are here. But the book is largely about the introduction of the current generation of EMRs into the everyday practice of ordinary clinicians. There are by and large three camps of opinions about what’s happened.

One is that the EMR is a pox visited on physicians that costs a fortune, has worsened quality, heightened medical errors, blown up successful processes, and ruined the lives of doctors–unless they were given scribes. The second is that because of the “rush to judgement” caused by the HITECH Act and Meaningful Use, we put in EMRs that were based on 1990s client-server technology but they were the only ones mature enough for the job. Most of this camp thinks that they were way better than paper, will slowly improve, and that doctors and patients will find that these technologies will soon integrate with easy to use iPhone-like apps as their APIs open up–and that if we hadn’t mandated EMRs when the great recession gave us the chance, nothing would have happened. The third camp agrees that EMRs are better than paper but felt that the way HITECH was rolled out kept a bunch of dinosaurs in business, and is preventing the health IT equivalent of Salesforce displacing Siebel (or Slack displacing email).

Wachter interviews most of the former ONC Directors and he puts Blumenthal in the second camp and Brailer in the third. He also puts athenahealth’s Jonathan Bush in the third camp, whereas Epic’s Carl Dvorak & Judy Faulkner are (optimistic) members of the second camp.  I think Bob himself is straddling camps two and three. UX expert Ross Koppel and a few other “camp oners”  are interviewed about how EMR vendors know nothing and care less about the user experience, other than they try to prevent anyone from improving it using a variety of dubious legal blocks. But it’s the major story in the book, excerpted in Medium and featured in The New York Times op-ed Wachter wrote, which is squarely in camp one. It’s the harrowing five chapter recounting of how an adolescent at UCSF–Wachter’s own hospital–choked down 38 pills when he was supposed to get two and ended up in a coma. In short the Epic system was trying to handle pediatric and adult meds at the same time. The user interface was too confusing, and the alarms so frequent that physician, pharmacist and nurse all cancelled them out multiple times. Even the (relatively experienced) patient gamely assumed that it had to be right as it was in the system. In fact, Wachter went to Boeing to see how they treat alarms and found that they very rarely interrupt a pilot and when they do, it’s really serious. There’s a telling moment in The Digital Doctor when a nurse knows something is wrong in the ICU because she can’t hear an alarm.

There’s already a mini-meme online beating up Wachter and UCSF (and treating them as one) for poor implementation (see HISTalk here and here) but that doesn’t let Epic or really any of their competitors off the hook. The “camp three-ers” all decry the lack of UX testing and the “camp oners” claim conspiracy–particularly contracts that prevent screenshots or public criticism of technology. I know that at the HxRefactored developer & design conference Health 2.0 ran last year all the major EMR companies present claimed they were in the process of increasing their UX team by a factor of ten. And Wachter did have to get permission from Epic & UCSF to share the screenshots of the drug error (and in fact to publish about it at all).

Despite all this, and several other complaints–like the community clinic doc who can’t hand out her patient education materials that work but aren’t MU certified–Wachter doesn’t want to go back. You get the feeling that he wants to jump to camp three–let’s get it all in the cloud and get rid of the huge client server monoliths we’re putting in. But that would suggest that the EMR cloud was ready in 2008-9 which it probably wasn’t, at least not for complex academic centers–although with the Halamka-Bush alliance it might be soon. But I sense that overall Bob is a little tortured. He (like me) is a believer, but the current pain we have put doctors through is too hard to bear.

Like any tortured soul he flees at the end to Silicon Valley. To be precise he visits the two major centers of “new tech in health care” both within a stone’s throw of UCSF’s new multi-billion dollar Mission Bay campus. And in order to guarantee a lousy review from me, he gave Health 2.0 and me half a paragraph while lavishing 3 pages and a full color spread on Nate Gross at Rock Health. (Heaven knows how much ink Halle Tecco would have got had she been around that day, but for his next book I’ll make Indu Subaiya fly up from Los Angeles!) I did joke with Bob that my review would mostly be about Eric Topol’s new book instead…

All kidding/personal grudges aside, I think that the vision piece at the end (post Silicon Valley chapter) is the most troubling part of the book. As it seems disconnected from the rest and perhaps most unrealistic, given all the issues Wachter already raised. It’s just hard to connect the future we hope to see with the present, especially when the future may be the present, only longer.

But it’s a deeply informative book and it’s destined to be a classic. I just don’t quite know what to think about it, As the subtitle says, there is hope, there is harm and there is hype, and throughout the book it’s hard to tell one from the other. Which–and this is where I agree with Bob–is the truth in health IT as a whole.


Categories: Uncategorized

Tagged as: , ,

20 replies »

  1. You know, the issue of cloud vs client/server has ABSOLUTELY NOTHING to do with the quality of the products, nor the possibility of sharing data, nor anything else discussed. That was a complete non-sequitur.

  2. I think we are going to need the EHR and the EMR for Artificial Intelligence to work, but surprisingly not for what most of its proponents claim are its advantages today. We are going to need it to help in getting the correct diagnosis, especially in our next era of precision medicine and molecular pathology: too much information out there for one human brain.

  3. “People quite senior in the administration – even the White House – knew of safety incidents involving “Certified” EHRs and chose to downplay them in the name of “progress”

    Document that. Name names. Cite/link actual statements.

  4. This is an important conversation to have within our industry, we need to face our challenges head on and solve these problems for the greater good. Given the fact that entrenched vendors are not going to be displaced any time soon, their willingness to make adjustments is critical and I hope this book helps the process.

  5. Quite a nice review of a book that I must read!

    I have to say that I’m a little in all three of Mr. Holt’s camps:

    – HITECH was sold to Congress based on lies (the now disproven Rand report) and the environment of the time (new presidency, ACA battles, focus on stimulus, need for “success” of HITECH) led to some short-sighted and politically motivated decision-making and steamrolling of doubters (i.e. people raising concerns about safety, interoperability, usability, fitness of vendors, etc.) People quite senior in the administration – even the White House – knew of safety incidents involving “Certified” EHRs and chose to downplay them in the name of “progress” (however illusory it was).

    – As stimulus HITECH had to be based on the legacy products available in the marketplace (as bad as they were/are) at the time. Vendors lobbied HARD to ensure that they didn’t need to make fundamental changes to inherently proprietary, non-interoperable products. Again, “progress” depended on it. ONC’s FACA’s chose to focus on the trivial many for Certification vs. the really important few (e.g. a standard data model, foundational interoperability standards for critical use cases, usability testing / use of standard colors and icons in products). Will incremental progress take place? Sure. Will the real breakthroughs to say, 21st century technology platforms, truly usable, interoperable systems that docs and nurses don’t hate happen? Not if the oligarchs (Epic, Cerner) can help it.

    The real question is: What is the path forward?


  6. Granpappy Yokum is right….the key to optimizing innovation is this:

    “everyone voluntarily and gladly used it.”

    For EHR’s to progress at the optimal rate, all the mandates, incentives, subsidies, rewards and penalties….all administered by bureaucrats seeking to coerce implementation for their own ends (not the ends patients and doctors want)…..should be lifted. Then the focus of all investment would shift to creating systems that patients and doctors would clamor for.

  7. “I do worry that docs are badmouthing (or worse) EHRs, creating poor public profile and opinion.”

    Agree. Docs should just do what they’re told and keep their mouths shut. Who told them they were allowed to have opinions?

    “before there was the iPad, there was the Newton; before there was ubiquitous wireless connectivity, we suffered the screeches of the dial-up modem”

    Terrible, terrible analogies. It was obvious that the Newton was a lemon and, like EMRs, nobody bought them. There was no government subsidy or requirement to use them, so they quickly disappeared. And, at the time, we thought dial-up was fantastic. It looks bad in retrospect, but it was an undisputed improvement on what we previously had, and everyone voluntarily and gladly used it.

  8. Thanks, Steven. I agree with your point. Here’s how I addressed it in the book…. After talking about the new pressures for value, quality measurement, increasing regulatory burden, etc, I wrote:

    “Clinicians’ unhappiness with the current state of health IT needs to be viewed against the backdrop of these wrenching, but largely needed, changes. They are changes that will transform the doctor-patient relationship, profoundly alter the roles of physicians and other health professionals, and ultimately result in some, perhaps many, people losing their jobs, income, and prestige. Our electronic health records could be vastly better, no doubt, but we also yell at them the way we sometimes yell at our kids after a bad day at work. They’re caught in the crossfire of a profession being reshaped.

    That certainly doesn’t mean that the EHRs are blameless. As you’ve seen, today’s systems are clunky and irritating on a good day, maddening and dangerous on a bad one. But perhaps it was naive of us to believe that the first versions of these technologies would be perfect; that in a field as complex as healthcare, the technology Hype Cycle could be abolished or the productivity paradox sidestepped. After all, before there was the iPad, there was the Newton; before there was ubiquitous wireless connectivity, we suffered the screeches of the dial-up modem, to say nothing of its molasses-like speed. We’re simply not smart enough to make it from A to Z without going through the rest of the alphabet first.”

  9. Just downloaded it and look forward to a good read on this ever-fascinating issue. Thanks Bob for your hard work and contribution, and Matthew for your review. I’d welcome more comments/discussion of the consumer interface issue, and how portals linked to EHRs are and are not working. I do worry that docs are badmouthing (or worse) EHRs, creating poor public profile and opinion. I had dinner with friends recently (people in their 60s and 70s) where 1/2 hour discussion took place about how EHRs have screwed up doc’s world and the doc-patient relationship. Two of the attendees said their docs and doc relatives had complained to them.

  10. Matthew — thanks so much for this generous review. I agree with pretty much all of it (particularly the “must read” and “destined to be a classic” parts) 😉

    Regarding the optimistic chapters near the end: well, you’ve captured my angst and ambivalence nicely. I do believe that HITECH was right and wise, in that it allowed healthcare to traverse a major tipping point. In 2008, 10% of doctors’ offices and hospitals had EHRs; today that figure is about 70%. We have finally, reluctantly, become a digital industry. I don’t think this would have happened without the $30 billion in federal fiscal fertilizer.

    But the enterprise products bought with the subsidies are suboptimal, to say the least, and, as importantly, we have committed the sin of turning them on without deeply appreciating their impact on the nature of the work. Until we do, we won’t get this right. That is the lesson of IT implementation from other industries: to reach the full potential, you need to have good tech, but then you have to completely reimagine the work.

    One of the positive side effects of HITECH is that it awakened Silicon Valley to the fact that 18% of our economy was largely unwired, and that the existing systems were immature. It was like discovering a new oil field the size of Canada, and it has had the desired effect (which you have helped lead); namely, start-ups and consumer-facing IT companies (Apple, Google, Salesforce….) are jumping into the mix. That’s good.

    It feels a bit like the building of the transcontinental railroad. Coming from one side, you now have a vibrant market for enterprise EHRs, which Epic is winning (whatever you think of them, they’re winning because they have the best system, at least for the moment). This is no longer simply a doctor/hospital play — today, more patients than clinicians will use an Epic EHR (via their portal). But, clearly, these systems were not designed to be consumer-facing.

    And then, largely coming from the left coast, you have the building out of the Health 2.0 track, one characterized by great exuberance and can-do optimism, with apps and wristwatches and sensors and whatnot designed to solve the world’s health and healthcare problems. Some of these products are fabulous and useful, some are trivial and wildly overhyped. But the market will speak and winners will emerge, and I expect that we’ll see some of the magic we’ve become accustomed to in the rest of our digital lives.

    And that’s where the transcontinental railroad analogy comes in: how do these two tracks come together? And where? Building an Epic takes unbelievable knowledge of the complexities (clinical, regulatory, billing) of delivering healthcare in a big system. I don’t see anyone — yes, including in Silicon Valley — displacing that anytime soon, or even trying to. On the other hand, I don’t see an Epic or a Cerner or a GE building the kinds of consumer-facing tools that your pals are working on in the Valley. And the usefulness of those tools will be stunted if they can’t share data with the enterprise tools your doctor is using. How does that all get stitched together? ONC? The market? Who knows?

    As far as my “camp” — I think you’ve pegged me right. I believe the argument that HITECH was a mistake because we should have waited for better tools is wrongheaded. We never would have gotten to better tools without the subsidies (that’s not to say that I’m a fan of Meaningful Use, which I see as stunting innovation via overregulation).

    And, despite the problems I’ve articulated in the book, I have zero doubt that care is better and safer with the today’s tools — massively imperfect as they are — than without them. Today’s chaos is unsettling but, I suspect, this is what progress looks like. To get this right, to reach that optimistic future state, we need healthcare delivery organizations, vendors, policymakers, researchers, and patients to make better choices. I’m hoping my book increases the odds that they will.

  11. Here is the problem – formalization certainly reduces quackery but then formalization becomes a quackery of sorts. I’ve seen this in peer review – where the reviewers comments are nonsensical monologues that express nothing.

    I wonder if the ONC could have designed an I pad.

  12. To that point, see “Wikipedia vs Quackery – Standards vs Chaos”

    “Wikipedia, an online open-source encyclopedia, can boast 470 million visitors each month, making it one of the most popular websites on the internet. It is an incredibly useful resource – I think it’s fair to say it is the online reference of record. For that reason people care how topics important to them are represented in Wikipedia.

    Wikipedia, in fact, has become no less than a battleground over certain controversial topics. In essence people generally want Wikipedia to reflect their opinions on controversial topics, and if it doesn’t then there must be something wrong with Wikipedia (rather than there being something wrong with their opinions). I don’t mean to imply that Wikipedia always gets it right – it is a crowdsourced reference and the content is only as good as the editors. But at least they make honest efforts to be neutral and to have standards.

    Those standards are the real conflict here, and it is part of a broader conflict over standards. In medicine there is a standard of care, which in turn is based on an underlying system of professional and scientific standards. Medical education is standardized, students have to pass standardized exams, post-graduate clinical training is standardized, there are standardized exams for specialty certification, there are ethical standards enforced by institutions, hospitals, professional organizations, and the state boards of health, and peer-reviewed journals have standards.

    This is all meant to ensure that individual patients receive the highest quality of care…”


  13. I’m now determined to read Wachter’s book in the next ten days, along with a thousand other things pending.

    But there is one thing that strikes me about EMR, as someone who knows nothing about IT.

    That is the problem lies not with EMR but with formalization (i.e. a central systematic way of collecting knowledge, assessing value and obtaining approval). To be more specific, the problem lies not with bad formalization but with formalization in and of itself.

    It’s like the problem with peer review. Peer review certainly curbs the proliferation of BS in journals. But it also removes things that the reviewer simply does not understand.

    EMR, to put simply, has not been subjected to the wisdom of the crowds. Rather, it seems to have been stunted by the biases of the wise.

  14. “preventing the health IT equivalent of Salesforce displacing Siebel (or Slack displacing email).”

    I had to use Salesforce while working in MU at my REC. What a total PoS. ONC made us use it, they owned the licenses. It was all about “tracking OUR every move,” not helping us get our work done. Sound familiar?

    We ended up doing “double entry,” i.e., we set up a mirror system in Microsoft CRM (another clunky PoS) in order to capture and retain the “business intelligence” we gathered so that our data would survive for our further use the short-lived REC program.

  15. The Medium excerpt as referenced is a terrifying and sobering story. Should be at the very top of meeting agendas for every single healthcare entity across the country immediately.

  16. I’m reading it at the moment (along with several others that are equally pressing). I will cite your review on my blog.

    I’ve just experienced Epic from the PoV of an ER gurney and MedSurg floor bed. The user opinions were uniformly negative. Frustrated, resigned, wary of saying too much.

    Q: “So, what do you think of Epic?”
    A: “It’s a great tool — for tracking our every move.”


    Re “tracking our every move” — see some of my other reads: ” The Future of Violence,” “Data and Goliath,” and “Terms of Service,” to cite just three.