Electronic records don’t tell us stories that make cognitive sense

One morning recently, I found another physician standing morosely at one of the mobile computer terminals we refer to as “cows”—computers on wheels—that are everywhere now in our hospital. I asked what was the matter.  “Oh nothing, really,” she said.  “It’s just that I don’t feel I know the patients as well as I used to.”

I knew exactly what she meant.  Things are different now that we have the EMR—the electronic medical record.  After two months of use, we’ve learned to our sorrow that these records don’t tell us stories that make cognitive sense.  Instead they offer data in endless lists.

Before the written word, people told stories.  In every culture, around hearths and on journeys, they remembered and retold tales of great deeds, romance, and tragedy.  When we were medical students, we learned to present each case on rounds by telling the patient’s story.  The story had well-defined elements:  the current complaint, the background of genetics or misfortune that led up to the present, the investigation that might clinch the diagnosis, and the plan of action.

The best stories almost told themselves.  The business executive fresh from a transatlantic flight presented with shortness of breath; VQ scan revealed a pulmonary embolism.  The young woman with Marfan’s syndrome began exercising one morning and developed severe chest pain radiating to her back; the echo demonstrated aortic dissection.

Now, however, we have lists.

One list will give us the medical history.  In no particular order of priority, it includes one-word problems such as osteoarthritis or hypertension that have nothing to do with the patient’s current admission for acute pancreatitis.  The relevant history of alcohol abuse may be found elsewhere, in the list under “social history”.  Our “social history” includes a field that will tell you whether or not the patient chews tobacco, which is so seldom helpful in southern California.  The complaint of abdominal pain won’t be found anywhere near the list of laboratory values with the important amylase and lipase levels.

If you’re a consultant trying to make sense of the patient’s case, you can find yourself frustrated and stymied at the difficulty of getting the big picture.  If you’re lucky, you can find a human who knows something about the patient, and get him or her to tell you the story.  You can bet that this won’t be the resident, who has just come on the service, didn’t admit the patient, won’t be following the patient, and will have to lie down for a nap soon.  But with perseverance you may find an attending physician who has no duty hour restrictions and actually knows what’s going on with the patient.

If finding a human fails, your second hope is to find a narrative note by a physician who is in the old-school habit of dictating an organized history and physical.  This is the pot of gold in the EMR, but you may have to sift through pages of notes on the computer before you find one.  Sometimes, just for fun, I print it out so I can refer back to it without logging on to anything.

The use of all the “smart fields” in the EMR looks appealing at first until you realize that they propagate themselves endlessly, like tribbles.  The same “past medical history” will appear as an identical list in note after note, because it’s so easy to type “.pmh” instead of summarizing the patient’s problems as a narrative.  If an error of any kind is made, it will continue until someone notices and takes the trouble to delete it.  If “Lasix” instead of “latex” is entered as an allergy, it may be listed that way indefinitely.  You’re much more likely to click on the wrong line of a list than you are to write down the wrong information in a handwritten note.

With the billions of dollars that are being spent on EMRs, and the Obama administration’s keen interest in their implementation, no one wants to hear about the problems they cause.  But the truth is that it’s much harder for physicians and everyone else in the hospital to learn and remember what they need to know about their patients from reading electronic records.  Human beings don’t learn best by memorizing disconnected lists.  From fairy tales to patients’ histories, we’re hard-wired to remember stories.

Karen S. Sibert is an Associate Professor of Anesthesiology, Cedars-Sinai Medical Center.  She blogs at A Penned Point.

Submit a guest post and be heard on social media’s leading physician voice.

Comments are moderated before they are published. Please read the comment policy.

  • http://twitter.com/drtwillett TheresaWillett MDPhD

    I love your tribbles comment! I remember working many hours with other interested residents and IT folks trying to generate a single ‘big picture’ page in my former hospital’s EHR. The real narrative part was still missing, but at least the active medical issue snapshot was there. Unfortunately, time-pressed interns and residents on the floors, and time-pressed and techno-wary attendings never used it. The ever present hamster-wheel feel of medicine allows companies to get away with poorly designed products, because none of us as physicians really hold them accountable.  There are ways of having smart fields, discreet data AND space for narrative, but alas not in most of what is available. The last field on earth that should be relying on flimsy IT and workflow work-arounds is medicine. 
    Thanks for spilling the beans so eloquently :-)

    • http://onhealthtech.blogspot.com Margalit Gur-Arie

      This has very little to do with flimsy software. Most decent EMRs and certainly those used in large hospitals have room for structured fields, discreet data and narrative. The problem is that those who dictate a narrative are not inclined to do double data entry and also click on boxes, or type in fields. Those who do document in the structured fields are not going to start dictating and very few are good typists.

      The technology to process narratives and extract structured data is in its infancy and not good enough for clinical decision making. The technology for creating narrative from discreet data is, by definition, deplorable and in my opinion should not be used.

      The preference for structured data is dictated by the government, and by the transition from individual patient-care to population management, which has no concern with any particular narrative.

      • http://dinosaurmusings.wordpress.com/ #1 Dinosaur

        Alert: Syntex Nazi attack:

        Both of you mean “discrete” data (separately identifiable), not “discreet” (as in “discretion”), which could humorously be thought to refer to HIPAA. 

        • http://onhealthtech.blogspot.com Margalit Gur-Arie

           With the proposed collection of gender definition, sexual preference and disability status, these may very well turn into discreet data elements in the near future… :-)

      • MarylandMD

        Let me get this straight: Poor documentation has very little to do with EMR software, except for the facts that narrative requires good typing, double data entry, and much better technology (either for processing discrete date into narrative or for extracting discrete data from narrative).  So the EMR makes it easy to produce narrative except for the fact that it makes it hard.  So it has very little to do with EMR except for the fact that it has a lot to do with the EMR.

        • http://onhealthtech.blogspot.com Margalit Gur-Arie

          There is no technology today to parse out spoken narrative into structured data, in a reliable manner. There are many companies working on it, but we are nowhere near a good solution.

          EMR software is allowing you to dictate as you always have. EMR software is also allowing typing and hand writing and the scanning of paper and in some instances the attachment of recorded sound files. EMR software is giving you all the options you had without EMR to create narratives, and then some.
          EMR software vendors are not the ones imposing the regulations that force you to record structured data. These same regulations are forced on EMR software vendors.

          The clunkiness of EMRs is due in large part to the implementation of clunky regulations, be it for revenue purposes, or more recently for meaningful use incentives purposes. If you want smooth and sleek software, perhaps we should get rid of the regulatory micromanagement of documentation, and health care delivery in general.

          • MarylandMD

            I have **never** dictated notes, so I don’t know how you can say that I can dictate notes as I always have!  Not all physicians have access to dictation, and many groups do not allow scanning of written notes except in unusual circumstances (power outages, etc.).  And anyway, if all you are going to do is dictate or hand write the note, what is the point of an EMR?

            While some of the clunkiness is due to regulatory and billing requirements, that doesn’t explain it all.  Many physicians find EMRs to be difficult to work with due to poor interface design (e.g., the excessively modal approach of EMRs like GEs Centricity, which requires you to only see and do what is in the current template and nothing else until that template is closed).  What makes sense to a computer programmer does not usually make sense to a physician.

            In any event, to argue that EMRs are fine, it is the lack of certain technologies (e.g., parsing spoken narrative) that is the problem seems wrong to me and a bit like splitting hairs.  It is the whole package that the physician is presented with: the EMR AND its underlying technology.  I think it is correct to critique the package as a whole as inadequate at this time.

          • http://onhealthtech.blogspot.com Margalit Gur-Arie

            Generally speaking, I do agree with your “diagnosis”. There are of course better EMRs and worse EMRs out there, but the cliche that they are designed by programmers who don’t understand medicine is not entirely correct. GE for example employs boatloads of physicians who have a lot to do with product design, and so do all other respectable EMRs. Programmers execute on those designs.

            Just to go back to the original premise of this post, that the narrative is somehow lost because of the introduction of EMRs, I was just arguing that the narrative is lost due to regulations specifically requiring that documentation not be in narrative form, and that EMRs although allowing you to record the narrative, must present documentation modalities that are compliant with regulations. They have as little choice as you do, and probably much less.

            Can these things be done better, even in the short term? Absolutely.

          • MarylandMD

            Just because GE employs boatloads of physicians doesn’t mean the EMR is physician friendly. My colleagues and I who struggle with the poor user interface of GE’s Centricity will share a big laugh over your comments.  We dream of the chance to have a “chat” with these physicians, asking questions like “What on earth were you thinking?!?”, “You call this intuitive?”, and “Do you actually use this software?”.

            Centricity looks like it was designed by computer programmers and physicians were only consulted after the initial critical structural decisions were made.  Its excessively modal approach is the opposite of how many physicians work.  It is clear that Centricity has a lot of extra features added at the request of physicians, but it suffers from the classic fatal flaw of bloatware–adding feature on top of feature on top of feature without making sure it all works well as a coherent whole.  It is the EMR equivalent of Microsoft Word, and that is not a compliment.

          • http://onhealthtech.blogspot.com Margalit Gur-Arie

            You may be surprised, but I actually agree with you in principle, The docs that work for (or own) EMR companies, have long since stopped practicing, or practice medicine as a hobby seeing a couple of patients a week. The EMR is designed by committees of people with all sorts of skills ranging from MBAs to JD to engineering in addition to the informatics docs.

            This is probably the reason why EMRs built in-house by hospitals, one piece at a time, seem to be liked more than off the shelf products. VistA, is one example and there are others. I know of one GE installation (ambulatory) where the medical group poured several millions into customizations and everybody loves the system now. Perhaps there is a lesson to be learned here.

            I love your MS Word analogy. That’s exactly what EMRs are, and what all enterprise software is, and we all use MS Word, moaning and groaning…. but we use it every day. You should try MS Project if you think Word is bad.

            P.S. I aim to entertain so I am glad you guys will have a good laugh sharing these comments :-)

  • http://twitter.com/rvpmd Rushil Patel

    How true – at times I even feel that the patient presentations juiced with jargon eliminates the patient altogether. Still, the oral presentation does far better than the list of data in the realm of maintaining a human connection. Thanks for sharing your insights! 

  • SaraJMD

    For the most part, EMRs were not truly developed with the patient encounter in mind, rather the data-mining potential. Not to say that data mining can’t help improve care, but it’s not the point for most of the people interacting with the systems. The organizations that largely fund EMRs (such as large hospital systems, government, academic institutions) are seduced by a product that can generate lots of reports, and the microcosm of how the system affects actually dealing with patients gets lost in the shuffle. It’s very frustrating.

    • http://www.facebook.com/profile.php?id=1508554960 Diana McCoy

      You have a gift for clearly and concisely identifying “the problem”.

      • SaraJMD

         Why thank you for the compliment! I’m gratified if I can help add clarity in any way to our immensely complicated mess of a health care system.

  • southerndoc1

    The horse is out of the barn on this one. The opportunity to develop EMRs that benefit patients and physicians is long gone. For the rest of the professional lives of most physicians currently practicing, EMRs will remain an obstacle to be worked around. Sad.

  • acaffaratti

    I was feeling like this too, but recently found a shortcut. It may be a lousy note, but I get to jot some personal stuff and put my actual thoughts into the plan. I can live with that which must be done for coding and keep it minimal and talk to patients. It took me four months to figure it out though.

  • http://www.facebook.com/profile.php?id=639156794 facebook-639156794

    Amen. EMR is the dark force in the universe. When I make emergency department observation unit rounds, I begin with reading the EMR(our format is Epic). Then I go back in and interview the patient, who invariably tells me “It’s all in my medical record….didn’t you read it?” I say, “yes, I read it. I need to know your story. Would you mind telling it to me?” Much more informative.
    When I do my EMR, I try (within time constraints) to make it as informative as the dictated notes I did in the past. Which means I usually am late getting out of every shift. But I feel the story is paramount. And each patient deserves to have his/her story told.
    I don’t know what the answer is. Our hospital aims to eliminate the option to dictate entirely due to cost constraints. I only hope that when that happens, and I am a patient, that I am awake and coherent enough to tell my story to each health care worker, so I don’t end up being allergic to “Lasix” instead of “Latex” and that my pertinent negatives don’t include a vasectomy….;(which I actually saw recorded on a female pt.)

  • Victor Trismegistus LA

    To be sure, I think we shouldn’t blame the technology, but instead think about what use (and misuse) we men make of it.

    I’m Dr. Victor Araujo, MD, IFCAP –  working as a Clinical Pathologist / Hemotherapist / Infection Control Professional at Salvador City, BA, Brazil. I suggest a manner of thinking:

    Back in 1988, when I finished Med, we used to talk and touch the patient.

    Nowadays I am a Dinosaur: Whenever some person asks me about this or that medical condition/exams, I ask in turn: “What medical specialist didi you visit? How long did he/she talk to you? DId he/she touch you? Examine your mouth, thorax, abdomen etc.”?

    Mostly, I get the answers that the consultation didn’t last ore than 15 to 30 min (though somewhat more time was spent waiting, as The Doctor is Very Important), made a brief examination (mostly directed to the main symptoms of the patient), and called upon these or those “clinico-epidemiologic data” to reach some conclusion. 

    Mainly, all of them do ask for these or those “complemantary” exams – which were the reason that brought my (now) patient to ask my “professional opinion” – After all, I am a Clinical Pathologist. Mark that often the exams are radiologic, or other kid of “non-laboratory” stuff, which should not be included in Laboratory Medicine specialty (After all, I am a Physician and should be kept aware of all the most modern medical tech), Note also that often I myself disagree as to the said “clinico-epidemiologic data”, and even as to the medical indication of such exams. Not, finally, that, IMHO, “complementary exams” should be used as complementary, meaning, as some additional instrument in the diagnosis – not as the providers of diagnosis themselves. Meaning that when you ask them, you should have some diagnostic notion (derived from History Taking ans Patient Examination), and would like to either prove it or disprove it, or else to have some idea as to its evolution, etc.

    What I mostly do in these case is: I take these (human) persons to some place like my room or some examination room in the hospital, and honestly talk to them. Sometimes this means a brief History Taking, even some brief items in the Clinical Examination. Most times, I do not require much time to conclude they have other problems then the primary stated clinical condition. Mostly that means, also, that  I disagree either as to the diagnosis and/or also as to the treatment.

    Now, when I state that to that persons, either I am turned into some species of DiagnoSaurus (no pun intended) or the said/sad “patient” disagrees with me ant goes back to his original Physician. They often prefer to have Hitech Diagnosis, Last Generation Antibiotic, and such.

    I do like Hitech and Last Generaton, but – I think, also, that they should be put in their due places – both by patients and by doctors!

    Yes, we have Electronic Records, we have Hightech, wue have Last Generation – we have bananas!. But… where is old fashioned  “talk with the patient”, “build a story”, “make a diagnostic hypothesis”, “work it out with complementary examinations”, “use cheapest, simplest and most tested clinical measures and medicines”? 

    — I tell you: NOBODY WANTS IT!!!

    Some of these days, I’ll let damn robots do all my lab work, computers manage through all the quality control data, Internet manage reporting and interpreting results, and my IPod generally manage Medical-Patient relationships. Who cares??


    All il all, I think patients (and physicians) should read les Science Fiction novels, and instead try to think about rewriting all that stuff – Huxlley’s Brave new World, Orwell’s 1984, Asimov’s Foundations, Dr Spock and such. 

    By the way, I love Science Fiction. But I am a fan of Raymond Bradbury too. Somethng Wicked This Way comes!

  • http://www.facebook.com/profile.php?id=1597478062 Nikki Campas Burdick

    I am a trained, certified Medical Transcriptionist. I put the title in Initial Caps so you would remember it. We have been transcribing the patient “story” for many moons, yet we are considered on our way out. And so we ask each other: who will care about the accuracy of the “story” – whether it was the right or left arm – and who will double-check the drug values & dosages, and who will, essentially double-check the physician (and, believe me, the ones I know rely on me to do this). In short, who will be the lone individual who cares whether the patient’s story gets told in a real, truthful fashion. Not the dropdown menus, not the EMR companies, not the greedy hospital administrator(s). Who will be left to tell the patient’s story?

    • Victor Trismegistus LA

      I think it is up to all medical professionals to show empathy and compassion for the patient, honestly hear him/her, truly examine him/her, ellaborate a honest description of the “clinical picture”, test and evaluate hypothesis through scientific methodology and complementary exams, treat the patient, monitor results, and register everything. That’s what we all were taught to do. Of course, we shall have the help of the other medical professionals. and we should make the best use of current technology. If needed, we shall have the help of technologists and informatics personnel.

      What many professionals do (and shouldn’t) is to stay behind any kind of technology; neither should we let technology block the “old ways” of hearing and examining, or of humanely thinking about our patients.

      Technology is great, but it is also man-made and therefore subject to failure if not adequately used.

      Patients and Physicians are both human, as well as technologists. None of them are to be put, individually, to blame for what is indeed a process (and a very human process at that) we’re all suffering from.

      This is Victor Araujo, MD, IFCAP, Clinical Pathologist / Hemotherapist / Infection Control Professional, working at Salvador City / BA, Brazil

  • MarylandMD

    An EMR is a tool, and, as with any tool, it can be used properly or it can be used improperly.  I have seen many, many physicians use the EMR improperly.  The problem is, once we are out in practice, we are all very unwilling to correct each others’ work, and so bad EMR habits do not get corrected.

    That being said, it is also true that some tools are very user-friendly and some are not.  EMRs are extremely user-hostile.  While it can be done, it isn’t easy to make sure your EMR note is relatively easy to read and tells a coherent story.

    Keep in mind, even in the old pen and paper days, a lot of physicians weren’t very good at telling stories that make cognitive sense.  When documentation is viewed as just an obligation that needs to be dispensed with as quickly as possible with the minimum necessary for billing purposes, then the note suffers.  Even before EMR, I have thought that many of my colleagues would do well to be sent back for some retraining in documentation and communication.  But I admit that the EMR clutters up the screen with additional (often pointless) information that makes a fair or poor note that much harder to read.

    • MarylandMD

      To be more specific regarding one example you cite in the article above: any physician who puts alcohol abuse just in the social history and doesn’t enter it as a diagnosis so that it appears under medical history AND doesn’t include it in the HPI (for a patient who was admitted for pancreatitis) is being an idiot, plain and simple.  That isn’t the EMR’s fault.

  • Sathyadeepak Ramesh

    Graduating med student here. I don’t understand the problem — it’s relatively easy to type an H&P “like the good old days” into a blank note. In fact, templates make it such that it’s hard to forget any part of the H&P, as you have to fill in all the sections. There’s plenty of room to free-type any sort of story that you would like, and it’s also extremely easy to sort notes by note type (e.g., “H&P” or “Progress Note” in EPIC) to find your “gold mines” of information. Blaming the EMR is non-sensical — the blame is on the physician who types a silly note with lots of irrelevant “data” pulled in.

    Also, let’s not forget the chief failing of paper notes — how can the patient’s story be told if no one can read it?

  • gerridoc

    In my current job, which involves a great deal of chart review and written reports, I have come to the conclusion that many physicians are terrible writers. I had to make an effort to improve my skills.  The ability to write a narrative or construct an argument is something that is not developed or valued in medical training. Physicians often wrote “SOAP” notes that were merely phrases, and were not coherent. (I was just as guilty as anyone else.) There was a time when consultants would dictate lengthy articulate letters to referring physicians. Given today’s pressures to see as many patients as possible, I think that the EMR has hastened the demise of meaningful communication between clinicians. We were already sliding down the slippery slope.

  • http://www.facebook.com/profile.php?id=1508554960 Diana McCoy

    Thank you, Dr. Sibert. I saw two related issues embedded in your comments. The first / most important is the idea that it is important to know our patients in context…something about the context of their lives and who what is important to them. Maybe eveno (archaic idea) develop a relationship with them based on the notion that we may be providing care to them over an extended period of time (unless their insurance changes and determines they must be treated elsewhere). 
    As for the electronic medical records…although enthusiastic about their promise initially, I also soon realized that they are a  mixed blessing. While very helpful in terms of most recent factual data, it can be difficult and time consuming to sift through old entries in an attempt to find relevant history / treatment approaches / responses and more. 

  • JenKove

    I get that the EMR requires some documentation and it can be translate to a lot of fields and lists, but there are open ended note fields in the ones I’ve seen for text like a story. Why does the patient/doctor dialog need to suffer based on how it is recorded? (enjoyed the Tribble reference!)

  • http://mtinnercircle.com Kathy Nicholls

    As the family patient advocate for my grandmother because of my understanding of medical things and my background in healthcare documentation, I experienced this first hand two months ago when she was admitted to two different hospitals, placed on a ventilator, and eventually died. The frustration of the staff and consultants not seeming to know her story or history for me meant less time spent with her in the last week of her life because so much time was spent correcting errors to be sure they really understood who she was. In addition, what we learned the day she was removed from life support was appalling when it comes to patient care and concern. She was removed from her Fentanyl pump (nobody wants to keep a dying patient in the ICU so she was moved to a floor, which wasn’t set up for “handling” a pump; okay I get that part for finances). As they readied her for transfer, she was told by the nurse to “just ring your button” if you have pain when you get to your new room. At 97? With dementia? Oh yes and let’s not forget she’s dying. We found the doctor, he ordered a standing order for morphine, and off she went. We took a break so they could get her settled in the new room. By the time that was all done, I asked them to give her the first dose of morphine because it had now been almost an hour off the pump and I wanted to make sure she suffered as little as possible. She went to get it, returned and said “I can’t order it yet, she’s not been transferred in the system.” That took another hour before the pain medication was given to her.

    On speaking to the family physician the next morning, his response was “that’s how it is with this new system. We can’t keep those meds on the floor like we used to, it all has to go through the CPOE system and come up to us when we need it.” He’s quite frustrated by it. And while I feel for his frustration as a patient, I have to wonder who is thinking about the patient in building these systems? Patients need their information to be readily available AND accurate. We seem to have lost so much in those two arenas and yet I hear very little in stories about how this impacts the patient.

    As for “entering things twice” in an EHR, I have to wonder how that can possibly be considered efficient or cost effective? Yes, it can be done, but at what cost to the patient and the time that physicians might spend really treating them?

  • http://www.facebook.com/people/Kathryn-Biggs/1133005692 Kathryn Biggs

    As the CEO of a small transcription company who is transitioning into working with EMRs, I too find this an issue.  I realize my opinion is biased but as my company works within EMRs for our clients I am amazed at how difficult the process is of finding information, how many errors are made with the push of a button and how seemingly detached the records seem.  When did the collection and analysis of data become more important that the patient?  Kathryn Biggs, Mercury Transcription

  • davemills555

    No question, EMRs have flaws. They’re will always be bugs. However, the flaws are outweighed by significant benefits. The more access patients have to their electronic health records and the more access patients have to the true cost of procedures, the quicker we will see costs come down and quality go up. Take for example Colorado’s new price list for procedures as reported by KHN. The article says, “…an MRI can cost as little as $450 to as much as $3,500, with no obvious distinction in terms of quality.” How can that be? I’ll tell you how! Greed! That’s how! The sooner our federal, state and local governments allow patients access to this kind of data, the sooner we will see the shysters of the health care industry simply dry up and blow away!

Most Popular