The Electronic Medical Mess


I posted a quick tweet about this a few months ago:

Over the years, I’ve worked with at least half a dozen projects where earnest, intelligent, diligent folks have tried to unlock the potential stored in mid to large scale batches of electronic medical records. In every case, without exception, we have wound up tearing our hair and rending our garments over the abysmal state of the data and the challenges in getting access to it at all. It is discordant, incomplete, and frequently just plain-old incorrect.

I claim that this is the result of structural incentives in the business of medicine.

What is a Medical Record?

Years ago the medical record was how physicians communicated amongst themselves. The “clinical narrative” was a series of notes written by a primary care physician, punctuated by requests for information and answers from specialists. Physicians operated with an assumption of privacy in these notes, since patients didn’t generally ask to see them. Of course they were still careful with what they wrote. If things went sideways, those notes might wind up being read aloud in front of a judge and jury.

In the 80’s, electronic medical records (EMRs) added a new dimension to this conversation. EMRs were built, in large part, to support accurate and timely information exchange between health care organizations and “payers” including both corporate and government insurance. EMRs digitized the old clinical narrative basically unchanged. They sometimes allowed in-house lab values to be transferred as data rather than text, though in many cases that sort of feature came much later. Most of the engineering effort went into building a framework for billing and payment.

The savvy reader will note that neither of these is a particularly good way to build a system for the collection of patient data.  Instead, we’re dealing with risk avoidance.

A Question of Risk and Cost

Being the Chief Information Officer (CIO) of a health care system or a hospital is a hard, stressful, and frequently thankless job. Information Technology (IT) is usually seen as a cost center and an expense rather than as a driver of revenue. A savvy CIO is always looking for ways to reduce costs and allow their institution to put more dollars directly into the health care mission. Successful hospital CIOs spend a lot of time thinking about risk. There are operational risks from attacks like ransomware, compliance risks, risks that the hospital will expose patient data inappropriately, financial risks from lost revenue, legal risks from failing to meet standards of care, and many more.

These pressures lead to a very sensible and consistent mindset among hospital CIOs: They have a healthy skepticism of the “new shiny,” an aversion to change, and a visceral awareness of their responsibility to consistent and compliant operations

So physicians are incentivized to avoid litigation, hospital information systems are incentivized to reduce exposure, and the core software we use for the whole mess is written primarily to support financial transactions.

Every single person I’ve ever met in the business and practice of health care, without exception, wants to improve patient lives. This is not a case where we need to find the bad, the malevolent, or the incompetent people and replace them. Instead, it’s one of those situations where good, smart hardworking people are stuck with a system that we all know needs a solid shake-up.

That means that when someone (like me) shows up and proposes that we change a bunch of hospital practices (including modifying that damn EMR software) so that we can gather better data, it falls a bit flat. If I reveal my grand plan to take the data and use it for some off-label purpose like improving the standard of care globally, I am usually politely but firmly shown the door.

But it gets worse.

Homemade Is Best

Back in the bad old days, it was possible to convince ourselves that observations made by physicians were the best and only data that should be used in the diagnosis of disease. That’s demonstrably untrue in the age of internet connected scales and wearable pulse and sleep monitors. I’ve written before about the reaction I receive when I show up to my doctor as a Patient With A Printout (PWP). Even here in 2019, there are not many primary care physicians who are willing to look at data from a direct to consumer genetics or wellness company.

The above isn’t strictly true. I know lots of physicians who have a very modern approach to data when we talk over coffee or dinner. However, at work, they have to do a job. The way they are allowed to do that job is defined by CIOs and hospital Risk Officers who grow nervous when we try to introduce outside data sources in the clinical context. What assertions do we have that these wearable devices meet anystandards of clinical care? Who, they might ask, will be be legally responsible if a diagnosis is missed or an incorrect treatment applied?

So we’re left with a population health mindset that says “never order a test unless you know what you’re going to do with the result,” except that in this case it’s “don’t look at a test that was already done, you might wind up with an inconvenient incidental finding, and then we’ll have to talk to legal.”

Health systems incentivize risk avoidance above more accurate or timely data. They do this because they are smart, and because they want to stay in business.

So we collect information with a system tuned for billing, run by people whose focus is on risk avoidance. Is it any wonder that when we extract that data, what we find is a conflicting and occasionally self-contradictory mess?

There’s no incentive to have it any other way,

A Better Way

Here in 2019, most people who pay attention to such things believe that data driven health insights will lead to better clinical outcomes, better quality of life, lower overall costs for health care, and many other benefits.


One ray of hope comes from the online communities that spring up to connect people with rare and terrible diseases. These folks share information amongst friends, family, researchers, and physicians as they search desperately for any hope of a cure. Along the way, they create and curate incredibly valuable data resources. The difference between these patient centric repositories and any extraction that we might get from an EMR is simply night and day.

A former colleague was fond of saying, “a diagnosis of cancer really clarifies your thinking about the relative importance of data privacy.”

Put another way: If we put the patient at the center of the data equation, rather than payment, we’re really not that very far from a much better world – and all those wonderful technologies I mentioned will suddenly be quite useful.

Unfortunately, that’s a political question these days:

So where do we go from here? I’m not sure.

I do know for certain that -merely- flinging the messy pile of junk against the latest Machine Learning / Artificial Intelligence / Natural Language Processing software, without addressing the underlying data quality, is unlikely to yield durable and useful results.

Garbage in, garbage out – as the saying goes.

I would love to hear your thoughts.

Chris Dwan is an independent consultant serving clients in biotech, pharma, and health care. His unique experience spans high performance computing, bioinformatics, and genomics. Prior to starting his consulting practice, he led the research computing teams at the New York Genome Center and the Broad Institute. He was the first employee of the bespoke consulting firm BioTeam. Chris’s graduate studies were in artificial intelligence.

He tweets regularly (@fdmts), and blogs occasionally (

47 thoughts on “The Electronic Medical Mess

  1. I have my own theories about what’s wrong with electronic health records and why they aren’t yet doing the miraculous things they’re supposed to.

    I’m long on Artificial Intelligence / Big Data and blockchain but can’t help but wonder if the problem is more fundamental. It seems obvious to me that at their core, electronic health records are a communications & information-sharing tool.

    As many observers have noted, the current generation of the technology has been optimized for billing and administration. That’s a problem. If they’re doing their job right, EHRs should allow information to be communicated quickly and efficiently. Based on the success of the web as a blazingly fast information sharing tool, they should work incredibly well.

    It’s no secret that the current generation doesn’t even come close to meeting that need.

    It’s as if somebody made an iPhone that only allowed outgoing calls by forcing the user through an elaborate phone tree every time a call is placed, mindlessly cloned incoming messages and dumped them in users’ inboxes and only authorized information-sharing with Apple’s corporate partners.

    And we’re surprised that users are complaining.

    1. Doctors are not the reason for useless EMRs, software companies are. Doctor and even hospital workflow were not considered in original designs. They were designed for hospital administrators and sold to purchasing agents. Providers were blindsided and early systems like EPIC and GE Centricity were difficult to train on.

      Later updates created supplemental modules designed purposely NOT to communicate with other software. It was a case of two or three fiefdoms locking out Innovation.

      Anyone remember Practice Fusion?

      1. One tiny correction. A hospital purchasing agent does NOT buy a billion dollar EHR. It’s the CFO and CEO. 🙂 But agree that MDs were not major players then. Also agree we should point to vendors first

      2. I don’t know, Ross…. I don’t know that we should point to vendors first. What follows below is first hand experience. For many years, I’ve worked for both independent vendors and hospitals engaged in building their own solutions.

        Most product development was geared at pleasing doctors. Most requests for enhancements came from practicing doctors. The highest priorities were for requests from doctors. Nobody gave a rat’s ass about what the back office wanted, or the front office wanted. It was always about the docs, because their time is the only billable time in the organization, especially in outpatient settings, and because, well, they are doctors and they know best.
        I know people who spent Christmas making sure that the pathology reports start 1/4 from the top of the page and end (exactly) 1/8 of an inch from the bottom because the chief threw a huge temper tantrum. Nobody else cared about that. He was the only one. I got an example of the “perfect” report, and spent all night laying printouts over it and looking against the light until the damn things matched exactly for every type of report. On Christmas night (I didn’t really care, obviously, but he couldn’t have known that). That was in a hospital system.
        Also every time we designed something new, we’d try to get the docs to try it out and give us “feedback”. We had to cater breakfasts (richly) to have anyone show up, and even then the nurses got to take home most of the food because one or two bothered to come down to the basement for 5 minutes.
        As far as vendors are concerned, I know the ambulatory segment really well and I know lots of people working for some big inpatient products, including plenty of practicing docs. Yes, practicing. And academics that do all 3 things. Nobody gives a damn about billing. It’s all about what the physicians would like best. Billing is boring. The best developers clamor to work on clinicals. That’s where the action is. That’s where you can do fun and challenging stuff. That’s where the best of the best end up.

        So who’s to blame then? I place the blame in its entirety on government intervention. Their stupid, stupid, data collection paradigm ruined everything. It stripped the elegance from almost every screen. It clogged the enhancement lists, so what doctors wanted had to be bumped down and, the horror, scrapped altogether because it didn’t jive with the latest BS requirement from first the PBMs (yes, the big PBMs had to approve the screens), then from CCHIT and then from ONC.

        Everything got ruined by this brutal and rude intervention into an industry that was still trying to find its legs. An intervention that brought with it billions and billions of bribe money and corrupted everything and everybody, including doctors who sold their soul to MU for a few thousand dollars, and also because of sheer terror about their future.

        So here we are today, with Dr. Topol making a video to promote his new book, saying that AI is going to free doctors from being data entry clerks…. I guess he forgot who turned them into data entry clerks, and also why….. and, no it wasn’t billing. It was “research” grants from NIH and all other “innovative” junk science peddled by HHS with full bipartisan support, from Bush to Obama to Trump, and pushed hard by Silicon Valley Big Tech that had its hands in all policy recommendations from PCAST to all government consultants and distinguished panels, where the longer your COI list is, the more of an authority on everything you are.

        Disgusting all around…. There. I’m done now 🙂

      3. Margalit,

        Your raise essential and very valid points. It sure as heck ain’t black and white. Vendors are trying to make a living, and certainly not actively trying to kill patients or destroy healthcare or enrage doctors….although they do a good job of mucking up the field something awful. But, alas, I’m told I have an interview tomorrow morning on the role of AI in healthcare with the university press folk, so I’m delaying my response to your thoughtful comment while I try to read Topol et al on AI. (To be fair, I’ve been working for several years with computer scientists on AI for cybersecurity and for healthcare… so I’m not a total moron on the topic…but gotta read up anyway.)

        Response delayed…but not topic not ignored.


      4. This is a great discussion. I’m now into Topol’s book a bit further, and his central hypothesis is that AI will give us virtual personal medical assistants that have all our information and will allow the doctor visit to consist of 13 minutes of care and 2 minutes of gathering information, rather than today’s 13 minutes of information gathering and two minutes of talking to the patient.

        I have such a difficult time envisioning that. He says he believes AI will end medical errors, misdiagnoses, and attention to screens during the patient visit. Someone tell me how that happens from the current context. I have sensors and monitors all over my body now, I listen to every medical podcast I can, I’m a good researchers on healthcare, and I still want someone to explain things to me when I’m sick. Topol says AI will facilitate that, but in the current “let’s automate and televise and read data to lower costs” environment, I suspect AI will just make sure I never see a doctor again, just a technician or a “physician extender.”

      5. Thank you for your comments. I agree with them completely. Dr. bashing is so 90s.

        Scientific evaluation of sweeping health policy change would have been ideal, even the 1970s, but it remains largely absent even today.

        There will always be healthcare policy salesman and guys hawking the latest health gadget. Because the public believes they can do what physicians do without the requisite 12 years of education. That’s not going to change anytime soon.

      6. Francine-
        Practice fusion is the EMR part of Office Ally, the computer billing software system my office uses.

        I don’t use practice fusion but if I had to switch off my beloved paper that is probably what I would choose.

      7. Yes, Niran, but I remember when it was a precedent setting cloud-based EMR for small practices. But it couldn’t stand alone.

      8. Practice Fusion was a free EMR funded heavily by venture capital. It was one of the crappiest EMRs I’ve ever seen. It made whatever money it made from selling data to the highest bidder. It was sold last year to Allscripts (for pennies on the dollar), which is continuing to operate the darn thing, although most stuff is not so free anymore.

        The more serious “cloud-based” “innovation” came from athenahealth. They had a very good run (IPO and everything) and the EMR is not too bad (and not too good either). athena was also sold last year…. to private equity.

        Health tech is an unmitigated mess now.

      9. An attempt at balance on EHRs. Response to Margalit

        I agree it’s too facile to blame only the vendors for the lousy state of EHRs. Although the vendors were instrumental in passing HITECH (that forced everyone to buy the EHRs and that subsidized the purchases), the vendors were aided by true believers in the technology–many of whom were clueless about it– and by the need for shovel ready projects. The industry association (HIMSS) and the tech true believers (e.g., AMIA) were ready with plans had been in waiting for decades.

        The vendors wanted market share and rushed their software into the hospitals and offices. Clearly it was not ready for prime time. The ONC acted as sales agent…and certainly not as protector of patient safety or rational workflow. Doctors became typists and didn’t like it. Everyone (ONC and health authorities) also saw the EHR as an opportunity to add a question or task (quit smoking, do you feel safe) which only exacerbated the rage. No one gave a damn about usability except a few academics who were viewed with the warmth and kindness usually accorded ants at a picnic. (Skunks?). The ONC and the vendors claimed usability was “subjective, determined by the implementation process, due to individual experience, too theoretic to consider, unmeasurable, and just needed a bit more training.” Absurd all, but that was the predominant mantra until recently. Now, all claim to care deeply about usability…although many are offering semi-workarounds via NLP, scribes and AI assisted dictation.

        Re: the role of doctors. All of the EHR firms hire MDs. But in my experience, the MDs quickly become co-opted to the marketing folks. CMIOs are MDs too. But if you’ve told your hospital that it should spend one or two billion dollars and 4 years on installing an EHR, cognitive dissonance and your job does not make you the voice of careful reason about problems. You want to say you were wise. And, you’ll be spending most of your time negotiating among dept heads, the CFO, the vendor, the vendor’s 50 reps (with a mean age of 21 years old), and the many consultants who are using up a lot of those funds. There is little room for buyer’s remorse…and even less reason to express that to anyone but your dog.

        No one wants to harm patients. But the work of making an EHR user friendly and safe is unending, hard, constantly changing in response to new drugs, treatments, demographics, technologies, etc. Also, a vendor has to think about the many customers/implementations that exist and are planned–including the trillions of dollars of legacy technology. You may want something good for your hospital, but the vendor must think more about the total market.

        MDs are important, but so is coordinating/linking the EHR with the 10 inside labs, the 5 outside labs, the pharmacy IT, the inventory, the nursing software, the eMAR, the radiology folk, the outside clinics and practices, etc etc.

        And then there’s the basic DNA or bones of much of the software: Billing …and the 34 insurance systems involved in that. All must be obedient to that critical function, even the MDs.


      10. I dug me a wee bit of a hole and the way out is to make the hole deeper??? Irish saying?

        Love ya man but this is not fixable; only way out is not use them and construct a record between you and your patient. Population based trials will then develop from targeted data collection with $ to patients not patents.

      11. Dream on, but better pick better dreams. Nonsense in; nonsense out. Need a new unifying theory of care we all can understand and appreciate; then we can reconstruct the idea of how to maximize the care of individuals. Great discussion.

      12. As a primary care doctor for 30 years, I think that EMR based healthcare represents a substantial improvement over “the good old days” but it depends on a unified system shared by all providers.

        When I see my patients now I can view all the consultant notes, the social worker notes, the labs that all the specialists ordered, etc. We all work from the same med list, so I’m not ordering meds that conflict with those the specialist ordered. My notes and others’ notes are now legible and available across the system, whether the patient is in the Emergency Room or elsewhere. I can filter labs going back years to find the ones I need at the time I need them. I no longer have to wait until a patient comes to the office to find out that they’ve not gotten their screening colonoscopy or mammogram. EMR generates a list for me. I don’t miss health maintenance issues when a patient has a visit because EMR provides alerts on each issue when patients are due. If the patient checks on their “patient portal” they can find out when they are due. When we’ve had outbreaks of illnesses, we’ve been able to notify those at risk electronically. If they have a question, they can e mail me in a privacy protected manner via their chart.

        The coordination and delivery of care in our system is far, far better than it ever was thanks in part to the EMR. Was it a struggle? Of course. But, the system continues to improve.

        Two decades ago I heard Robert Brooks, healthcare researcher at UCLA/Rand, tell public health students that if a doctor received a chem panel on his patient and put the report in his pocket, he was more likely to harm the patient by having it fall out of his pocket than by making the wrong decision in dealing with it. That’s a paradigm for medical harm. Accidents and the issues that fall between the cracks largely responsible for the fact, cited by the Institute of Medicine back in 2000, that as many as 98,000 people die annually from medical errors. Many of those issues can be prevented by effective use of an EMR.

        Getting EMR to work has been a major struggle. In our system it’s paying off. I benefit from it every day as do my patients.

      13. See and I can do everything you are talking about way, way, way better on paper. I’ve only been in practice since 2001, but I rarely need specialist except with occasional heart defects. Where I am located, there are no pediatric specialists. When the specialist to make a change, they call me and talk to me over the phone, because due to distance restrictions I do most of the follow up for the specialists.

        My patients have ONE Doctor, me. My notes are transcribed so they are always legible.

        The way I look at it is this: your inability to make it work on paper shouldn’t be my problem. (I’m not saying this personally to Dr. Stone rather more broadly) I’m advocating for choice and freedom to document the most efficient way for me and my patients

        I have 3000 patients I’m responsible for. I have to see 25+ sometimes in a four hour day.
        If I have to use an EMR, 1500 children will have to find a new physician in a health professional shortage area.

        How is this better exactly?

      14. Thank God for a voice of reason. The doctor provides good care, notthe damn computer. 🤮

      15. I will say it — EMRs are flat out dangerous and should not be required unless they can be scientifically proven to save lives.

        If the surgery center of OK can do well on paper, so can I.

      16. If you’re the only doctor taking care of a patient and can call all of the patients there’s not much benefit to an EMR.

        I practice internal medicine and geriatrics in a busy metropolitan area. My patients have many medical problems and see multiple consultants. They get their in-patient care from hospitalists. If I had to call all those physicians I’d do nothing but that. In the “good old days” the consultants left illegible notes in hospital charts. When they saw patients in their office and changed the patients’ meds, it would take two weeks to get the consult note, if I ever did.

        In a setting like mine, we have evolved from “one patient one doctor” to teams of providers taking care of complex patients. So, having an electronic platform ensures coordination of care by making everyone’s decision making transparent by maintaining single problem lists and medication lists.

        Giving patients the ability to see their system med list, their upcoming appointments and what’s due for health maintenance is also extremely valuable. And, having non-simultaneous communication via the patient portal keeps them connected to their doctor.

        Paper worked well for 19th Century medicine. All the (legitimate) anger and frustration about the roll-out doesn’t change the fact that complex teams providing high level care need the coordination and data collecting/organizing ability of an EMR. It works well in our system.

        EMR doesn’t “take care of the patient.” It’s a tool to assist care delivery. Most medical systems don’t use the EMR to their advantage or to the patient’s advantage. We might ask ourselves, if a carpenter isn’t using a hammer correctly, is the fault with the carpenter or the hammer?

      17. Nine years ago
        and we are still arguing about this…..

        It’s not the idea of EHR and it’s not the fact that computers are used. It’s the hijacking of the primary purpose by greedy bastards.

        You are very lucky, Daniel. I know other docs that work in large systems and are equally happy with their Epic or Cerner, and even that old VA EMR. For some folks, it just so happens that the EHR is precisely how they like to work. Maybe they helped influence the installation and configuration so it fits their personal style, and maybe they just got lucky. My experience has been that there are huge differences in preferences and the best way to accommodate that is to be as non-prescriptive as possible. That however is not allowed when every tiny data element is specified from above, usually by people that have other goals in mind.

        I’m sure some folks absolutely hated Plummer’s pulleys and tubes and lamented the loss of their beloved ledgers. The difference is that back then nobody forced physicians practicing in a remote rural practice to purchase and install pneumatic tubes to nowhere.

        Regulators should have stayed out of it. Nobody hated EMRs before the onslaught of government intervention. My guess is that we would be much further down the road to real time communications without the strong arm tactics, brought to us by various “data” seekers.

      18. If EHRs didn’t exist, I’d be begging for them. Their advantages are massive and many. They do amazing things! And if I one had to build them from scratch, the process would be mind boggling. The problem with them, alas, is that the available ones ain’t very good. The usability sucks. Their (non)integration with workflow is too often perverse. Dan Stone illustrates what EHRs can do. Not magic, but wonderful. They have massive benefits…along with their failures.

        Unlike Margalit, I don’t only blame the gov’t. Sure, the gov’t is guilty of forcing everyone to buy them when they were not ready. But the gov’t did what it does when it’s under the power of rich and well connected vendors, unexamined assumptions, hoped for technology, ideologs who had connections and acted at the right moment, the need for shovel-ready projects, and everyone’s hatred of medical errors (which EHRs were supposed to stop…HAHA).

        So, they turned from dream technology to — not the sorcerers apprentice, but certainly something getting closer to Charlie Chaplin’s “modern times.”

        Unrelated: Adrian: What makes you think I don’t promote open source EHRs. For years, I’ve been referring folk to Hamish Fraiser (now at Brown) and to vistA.

      19. Ross, when I say government, I mean government writ large. We are now governed by Robber Barons. The intermediaries, i.e. elected and appointed officials, are there exclusively for mass entertainment purposes. Rock stars….

    2. Agree with Bob McNutt that Topol is way, wwwwwaaaaayyyy too optimistic. AI can help. A lot. AI can suggest alternative DX, can scan 10000000 eyeballs for diabetic retinopathy, can help radiologists, etc. But it ain’t magic. And the AI ain’t gonna catch the slight delay in your voice when you explain your stomach problems to your doc. And unlike Topol’s hype, I don’t think discussing my childhood with a screen is going to be very efficacious in helping me resolve my emotional struggles.

      But now we return to Bob McNutt’s concern about the horrors of the EHR — the fragmented, twisted, money focused hole they have helped dig.

      Yes, they are user-hostile, but I don’t see how the requested one-on-one relationship with my “doctor” is going to help. I’ve got a GP, an eye doctor, my back surgeon (one time), my audiologist, my dermatologist, dentist, periodontist, my various allied professionals (a PT to help when my running hurt me), two pharmacies (because of legacy and price), and recently I found that IBX keeps a PHR on me.

      So what’s the one on one relationship that’s going to help? There are about 8 or 9.

      Now, Ideally, an EHR would capture that all of the data and pull it together. . I can assure you — as someone who knows a lot more than the average bear about these things — that 8 EHRs don’t add up to anything approximating reality, Not individually or collectively. I’m downloading all of my 8ish PHRs. 3 are from the same hospital system and 3 are from another hospital system– two systems blocks from each other. Ergo, you’d think I could have 2 EHRs that combined 3 docs and 3 docs….plus some ectopic EHRs (eg, Rx, IBX….) But no! Even from the same hospital system, the histories, meds, problems etc differ wildly. I’m unrecognizable, even to me. And I am not able to get them fixed. There are spaced for me to note errors but the box says that no one reads them. (Note: I’m professor of medical informatics at two major universities and some of my doctors know who I am. Nevertheless, nothing gets fixed.)

      Bottom line: I want technology to help. I want what Dr. Stone says he has…a full record of his patients’ labs, meds, etc. My docs don’t have that and I can’t get it unless I cut, paste, edit and send posters to my various providers….whom I rather sure would ignore my missives.

    3. So, dear Ross, love your note. You see, Ross, you probably don’t need to see many of those you are seeing, and if you would have come to me with your, Bob’s EMR, you would not need to worry about their data.

      And, if, per chance, you did get hoodwinked into seeing those folks, you could open Bob’s EMR and enter information; if you need to know what data to enter on your own, if you will, I will help you. For dentist, don’t enter a thing. For audiology, just the objective data. Why the heck are you seeing PT, not a single trial says they benefit you and I will tell you how to fix your running as I run too so their data collection need is gone. I don’t give a rat’s hind end how many pharmacies you have, write down your med or just bring it to a visit so I can cancel them.

      What is driving me nuts about this conversation is that EMRs, Digital data or paper have nothing to do with practicing well. It is not what we have, it is how we use it. I know that many are smitten by this topic but it is a waste. Let’s talk about what care is worthwhile and why, talk about how to get better information focused on specific disease topics, pertinent context specific patient oriented data collection,and how to incorporate a patient’s input rather than the damn practice of medicine’s wishes and needs. The video sent by Adrian was cool. Let’s democratize medicine by getting out of it – as presently constructed.

    4. Dear Ross, Bob, Francine, and all

      For me the “good old days” were the 5 years B.E. (Before Epic.)

      One of my Rheumatology Fellows was extremely bright, an excellent pianist, and a good hockey player.. He had impressive mathematical aptitudes and an affinity for the life of the binary. He discovered that his love of patient care had limits, but the lure of administration had none. So he positioned himself in the administration of UNC Hospitals during its adolescent growth spurt and bullied his way into becoming the leader of a team devoted to digitizing our medical records, which were still heavy binders of sheets of lab data and hand written notes as well as typewritten discharge summaries and op notes.

      [Interlude in this narrative: We had a record room staff devoted to keeping the binders together and delivering them to the appropriate clinical venue in a timely fashion. All of us grew up knowing how to dig into these idiosyncratic archives. All of us knew how to review the drugs and referral documents. All of us knew the record room staff because we were called down to complete or sign off on discharge summaries. It was sloppy, inelegant, but we knew it and tried hard to deal with it for each patient. It was my custom, for decades, to write my clinic notes on double=sheet paper that automatically made a copy on the under-sheet. I kept these and the voluminous material mailed in by referring physicians in files in my office – files which I brought to that patient’s next clinic visit. I did this because mailed-in materials seemed to disappear in the miasma for great periods before finding their way into the “record.”]

      Back to the summary and the machinations of my former fellow: He also worked his way into some sort of collaboration with programmers down the street where IBM has a large facility. Together they created an EHR, called WebCis, for exclusive use in UNC Hospitals and Clinics. We all grew familiar with WebCis and facile with it. The learning curve was steep, smooth and entirely relevant to our need to care for patients. It was not codified for billing or for ready accumulation of large data sets. But it made it easy to care for Mrs Jones. By the way, most of us typed our noted directly into WebCis. Some, like me, would dictate a letter to referring doctors following clinic. My letters were valued because they responded directly to the queries of the referring doctor and were specific to the particular patient’s circumstance, including reference and relevant digressions. I had a “secretary” (that’s a human being whose name I knew, who knew my idioms, and who was a partner in communicating). The letters were so letter perfect that editing was superfluous (I never could train my EPIC dragon and most such notes by those who think they can are replete with blanks and errors.) The letters ended up in the WebCis chart and were mailed as soon as I signed off on them.

      WebCis was INEXPENSIVE to adapt and maintain. It has been replaced by EPIC for an obscene initial investment, great maintenance cost, and predictable doctor “burnout.”

      The inmates have taken over the asylum.

    5. Bob, If I lived in Chicago, you would be my doctor. Absolutely.

      To All: Bob wrote a wonderful book for laymen/women on how to work with one’s doctor and how to understand medical nonsense and sense. The book is:
      Your Health, Your Decisions
      How to Work with Your Doctor to Become a Knowledge-Powered Patient
      By Robert Alan McNutt

      I recommend it enthusiastically.


      PS: I’m downloading my Portals and PHRs because I’m going to write an article about the absurdity of the myriad contradictory lists, meds, histories, etc. According to my PHRs and portals, I’ve been dead for about 14 years because they have me on about 10 PPIs simultaneously. Apparently when I switched from Rx to OTC, and then lowered my OTC dose, nothing ever stopped. I’ve digested no food for 14 years and yet sill want to drop 3 lbs. All meds are in their forever. Every typo or miss-click is my enduring problem (err… on my problem list). An eye ointment prescribed for 3 days is still active 12 years later. My god, how that little tiny teeny tube lasted for over a decade. This is a 2 gram tube of biblical generosity.

      Praise pfizer or merc or J&J or….


    6. Hey, Ross

      Bob’s faculty-level peregrinations commenced as junior faculty in the Department of Medicine at UNC. I was also junior faculty at the time (I joined the UNC faculty in 1973). Not only is Bob a truly fine physician, but he’s an outstanding teacher. Several of his students when he was at UNC are now senior faculty and consider the experience of having been mentored by Bob McNutt to be pivotal in their career development.

      As I said to John Irvine before, they broke the mold.

    7. Yup-

      Basically you will see a physician extender, they will miss something and then when you are really in trouble, the doctor will see you.
      Your explanation of AI is exactly why it won’t work. Taking s history from a patient is how I make the proper diagnosis. It is literally that 10 minutes of time that I observe them while they talk to me. How are they breathing? What is their appearance? Are they sick, dehydrated,” sinusy” sounding. Are they mouth breathing? And my favorite question while observing: What are they not telling me?
      I can’t do my job if I don’t take the history. Having AI do it for me defeats the purpose entirely.
      This is the entire problem with electronic records. They never knew what information is important so the physician spends that 13 minutes gathering useless information, still needs to talk to the patient about what’s important ignoring the template entirely, and then coming up with a diagnosis thereafter. AI can’t replace that 13 minutes. It’s ridiculous to think that would make us more efficient. But again the systems are being designed by non-physicians.

    8. This has been a fascinating discussion. Regarding Topol and the promise of AI, I’ve written before that the hype of the product is far ahead of the reality. Right now, it’s a great diagnostic tool when fed necessary proper information but, like Ross said, it has little ability to evaluate inputs or engage the patient and suss that out. AI proponents tend to underestimate how difficult it is to get relevant information and how to use it to generate fresh analysis and new relevant questions.

      1. +1 AI proponents tend to underestimate how difficult it is to get relevant information and how to use it to generate fresh analysis and new relevant questions.

  2. Don’t think we got what intended. Easy access to OBJECTIVE data, graphs of run-time of data, collection of important clinical/personal factors, decision discussions, and notes reminding to talk about other family members. Don’t have EHR we conceptualized; got a billing machine?

  3. IMHO, risk is a minor part of the problem. Sure, there’s a risk of having to follow-up an abnormal result that may be incidental but the reason EHRs suck as a clinical tool is because they are designed as a strategic tool for the business-of-medicine. Health records are, first and foremost, a way to control the practitioners and the patients. This is how administrators can grow by 3,200% over a period when physicians grew by 150%.

    Patient-centered longitudinal health records can be had but they would have to be institution-agnostic, standards-based, and open source like the rest of clinical medicine. Where’s the profit in that?

    1. I can’t figure out what Chris is talking about. This essay reads like a reverie, not a useful comment on EMRs.

      I understand the problem he’s trying to solve — useless EMRs — and I gather he thinks doctors stuck in the past who go to work fearful of lawsuits are the problem, but I can’t figure out what he’s proposing to do to solve the problem. (As I believe you all know, I’m not a doctor.)

      My primary criticism of his essay is the high ratio of abstract and questionable claims to precise and evidence-based statements. There are way too many of the former, and none of the latter. Here are several examples of the former:

      Back in the bad old days, it was possible to convince ourselves that observations made by physicians were the best and only data that should be used in the diagnosis of disease. That’s demonstrably untrue in the age of internet connected scales and wearable pulse and sleep monitors.

      “Bad old days?” “Observations … by physicians … the best and only data?” “Demonstrably?” I doubt “wearables” as a class of gizmos are all that useful, but if some are, what’s keeping patients from telling their doctors about what the gizmos are telling them? Chris seems to argue we should just pour all wearables data willy-nilly into the patient’s EMR, but I’m just guessing because this sentence is so unclear.

      If I reveal my grand plan to take the data and use it for some off-label purpose like improving the standard of care globally….

      Could Chris elaborate on his “grand plan”? What does “improving the standard of care globally” mean, as opposed to, say, “improving the standard of care”? And why improve the “standard”? Why not improve “care”? He refers to “the data.” What data might he be referring to that he would “take and use” that could have such a marvelous outcome?

      Even here in 2019, there are not many primary care physicians who are willing to look at data from a direct to consumer genetics or wellness company.

      Is this true, or is this just Chris’s opinion? If it’s true, are the doctors in a better position to opine about “genetics or wellness company” data, or is Chris? Rather than stereotype all “genetics and wellness” data as useful, and all doctors as Luddites, might it not be more helpful to have a discussion about the difference between junk data from wearables and useful data and who in society is in the best position to make that distinction?

      So we’re left with a population health mindset…. “Mindset” is abstract enough. What is a “population health mindset,” and how did we “get left” with it?

      Here in 2019, most people who pay attention to such things believe that data driven health insights will lead to better clinical outcomes,…

      This is circular reasoning. Yes, of course, if you say an idea is an “insight,” then it follows it will lead to “better” something, in this case, “better clinical outcomes.” It doesn’t make the claim any less circular or tautological to put “data driven” in front of “insight.” And, BTW, who are “most people who pay attention to such things?” And how would we know that “most” of these people agree with Chris? I think I’m not one of them, but I can’t be sure, because I don’t know if I’ve been paying attention to “these things.”

      Put another way: If we put the patient at the center of the data equation, rather than payment,….

      Putting patients “at the center” of something has been all the rage since the early 2000s. What does “the center of the equation” mean? What “equation” are we talking about? How would “we” accomplish this?

      Sorry to be so critical, but the American health policy debate suffers from way too much of this wishful thinking, this voice-of-God opining about the latest fads without evidence. We need to call it out when we see it

      1. Hey Kip, thanks for the comment.

        I assume that you’re the Kip Sullivan who wrote “The Health Care Mess: How We Got Into It and How We’ll Get Out of It.” I haven’t read it, but I assuming it’s great and proposes the kind of specific, data driven, actionable plans that you found so lacking in my blog post.

        If that’s the case, I’m at a significant disadvantage here. I wrote a blog post, you wrote a book. I had a cup of coffee and worked for a couple of hours. You researched and edited for months or years.

        I’m thrilled that you took the time to read and respond!

        In my defense, I -have- written a couple of other posts describing my experiences with direct to consumer health services and how my primary care physician has rejected them out of hand. As above, it’s just me, anecdotally describing my experiences. Each post is absolutely vulnerable to your critiques above.

        I share them, not in rebuttal, but to continue the conversation:

        Also, while it’s still blog-scale and not book-scale, I did write a follow up piece sketching the beginnings of a path forward. Shared here in the same humility as above:

        Thanks again for your detailed comments, and keep up the good work.

    2. Re: response to a comment from Adrian.

      Dear Adrian, we (Steve and Ross) know you care deeply about patient safety, and probably your statement was part of a larger thought …that is not intended as presented. But we wanted to respond. That is, Steve Soumerai and I noticed your comment about minor risk from EHRs. Are you really saying that the tens or hundreds of thousands of patients harmed by bad EHR designs and implementations “is a minor part of the problem”? EHR induced errors are massive, frightful and all too common. Gordy Schiff et al published examples of thousands of REPORTED errors linked to EHRs. Koppel’s original JAMA piece linked 22 EHR design failures to errors. Han et al found a 3-fold increase in mortality when his hospital installed an EHR in the pediatric unit. Jon Nebeker found similar problems with the VA’s EHR. Also, a team from 5 hospitals (Penn, Harvard, U of Ill, Montefiore, U of Ill, and Kaiser) and reflecting 5 EHRs, studied the errors generated by the interfaces. We documented 53 screens that were often lethal, e.g., warnings for overdoses of warfarin that don’t appear if you order coumadin–even in the same system and with the same patient. It’s a free download at (and then type CPOE or EHR). Almost last, remember that the IOM issued “Health IT and Patient Safety: Building Safer Systems for Better Care” which pointed out the several harms caused by EHR . Last, two days ago, ECRI came out with their annual list of the top ten patient safety dangers. EHRs topped the list for the 5th time in a row.

      We’re not saying EHRs are inherently dangerous (although some would say that flat out) but rather the current crop of EHRs just don’t work anywhere as well as they should or could.

      Ross and Steve.

    3. I did not mean “that” risk that you and Steve write about. I have no claim to insight at that level and have never commented on your work on risk as minor. I read the original post in terms of risk of malpractice or business risk. Maybe I read too fast, but that’s what came across to me.

      Nevertheless, now that you and Steve raise the risk of EHR design, what you refer to as “problems”, it gives me an opportunity to once again ask why you are not more supportive of open source clinical EHR software as a way to bypass gag orders, promote public bug tracking, enable fixes at the local level, pay primarily for software maintenance rather than licensing, remove barriers to competition in software development and software maintenance, and promote transparency in medical practice in general? What have you written about open source clinical software?

  4. Thanks for replying, Chris.

    I have written many short essays for blogs and newspapers, and only one book. I understand and chafe at the limitations of 800- and 2,000-word articles. Often I feel I’m forced to oversimplify, and to leave unmentioned much of the research that supports what I’m saying. I definitely did not expect you to write a book.

    I do expect people who claim to have expertise in health policy to write clearly, and either to refrain from making sweeping judgements or to make some attempt to illustrate or document such judgements. Even a link or two to good research is all I ask for, not a literature review. Your statement that data derived from wearables and genetic testing will “lead to better clinical outcomes…” is an example of a sweeping judgement you must document some way. We hear that claim constantly from the EMR industry and DNA testing companies and their allies. When you repeat it without qualification and without evidence, you sound like a salesman.

    You got a full blast of irritation that has built up in me over decades that in hindsight I wish I had toned down. For decades I have had to listen to and read stuff written by advocates of one fatuous, evidence-free health policy after another — advocates who had no evidence to back up their claims and who persisted in promoting their pet project (HMOs, “integrated delivery systems,” report cards, “medical homes,” ACOs, EMRs, P4P, MACRA, HSAs) long after evidence had piled up indicating the project was not working, or the harm it caused exceeded its benefits.

    Often these speeches or articles also indulged in demeaning stereotypes of doctors and hospital executives.

    I wrote an article with Ted Marmor five years ago criticizing the sloppy habits of managed care advocates. Ted also has enormous BS antennae and has written other articles about BS from the health policy aristocracy.

    Ted and I characterized these habits as disdain for evidence, and use of abstract and manipulative labels. (you might start at p. 162). A “manipulative” label is one that presumes something that hasn’t been demonstrated. “Health maintenance organization” is an example. When Paul Ellwood cooked up that phrase in 1970, he had no evidence to support his claim that these new-fangled entities were good at “health maintenance.” He just said it, and powerful pols and business people swallowed it whole.

    The EMR industry and their allies often reveal the same bad habits.

    These sloppy habits now characterize much of what passes for learned commentary and peer-reviewed research.

    I’m very interested in knowing what doctors think about having their already cluttered EMRs cluttered up even more with reams of data from wearables. But I don’t want to read sweeping judgements. If you want to tell me what your primary care doctor said in the course of rejecting your ideas, do that. Tell us the idea, and tell us what your doctor said. That anecdote could be interesting. But if you issue sweeping judgements about all doctors based on one or two conversations you had, you damage your own credibility.


  5. Kip,

    Thanks again for engaging so thoughtfully. It’s reassuring to me that the strength of your response was inspired by a larger and more pervasive problem than my little blog post.

    I hear you about the dangers of argument by anecdote. I hope that you’ll forgive me for not citing the literature when I say that I think we agree: Data driven, evidence based medicine is vastly more effective than the alternative.

    I also agree with you that the world is full of snake-oil salesmen, and that their pitches do tend to lack rigor. There is also a pernicious culture of pitch-men and women who wrap themselves in off-brand certifications and titles and relentlessly cite each other in order to achieve a bit of cargo-cult style credibility among people who can’t tell the difference.

    Perhaps a bit of the disconnect lies in the idea that I’m trying to pass myself off as an expert in health care policy. I’m not. When you say that I’m damaging my own credibility, I’m flattered that you granted me any in the first place.

    I am curious though: The way I read your responses – it sounds like you think the would would have been a better place without this particular blog post in it. Would that have been better, in your opinion?



Leave a Reply