EMR doesn’t help much

So says Dr. Michael Wilkes in today’s Sacramento Bee:

Inside Medicine: So far, electronic records don’t help patients much

By Dr. Michael Wilkes –

Published 12:00 am PST Saturday, January 20, 2007
Story appeared in SCENE section, Page K1

Not long ago, I regularly coached students and residents that during their patient encounters, they should put their papers and notes aside and listen to the person — look them in the eyes. There was so much more to be learned by having doctors use their eyes as well as their ears.

Today, a flat-screen computer sits between the doctor and the patient — just as a fence divides two neighbors. My students and residents — like doctors around the country — are slaves to the computer and electronic medical records. If you’ve not had the experience of sitting across from your doctor as she or he types your medical history into the computer, then just wait, as it is likely coming. One health-care expert calls the push for the electronic medical record (or EMR, as it is called in medical circles) the search for medicine’s holy grail….

I like the article’s conclusion, which I’ll encourage you to read.


  1. I disagree with the conclusions of this article. While it is true that the current state of EMRs is sub-optimal, the principle is valid.

    The article states, “In medicine, our fundamental activity is not stocking warehouse shelves or ordering merchandise from vendors. Medicine involves one human listening to, talking with and examining another.”

    Medicine also involves doctors attempting to recognize patterns within the prose of a patient’s history, and the course of a patient’s symptoms. Physicians look for order in the chaos of a patient’s life in order to determine the diagnosis and appropriate treatment. Think about how we organize our notes and thinking.

    An EMR that follows a national standard allow for us to perform large-scale outcome studies, it would allow for us to save money by identifying therapies that don’t work. It would allow providers to avoid repetitive tests and treatments. It would accomodate the fragementation within our current healthcare system, with the same patient being evaluated by their PMD as well as multiple consultants, hospitalists, etc. It would save time by avoiding the repetative taking of the same history, and by avoiding repetition, it would avoid errors from being propagated.

    The bottom line, however, is that until EMRs follow a national standard and until the architects of these symptoms realize that a bottom-up approach (a system that addresses the fundamental necessities first) will be much easier than a top-down (a system that tries to accomplish everything all at once), we will continue to struggle and spend/waste money.

  2. Lots of principles are potentially valid but don’t work in practice (see also, communism, socialized medicine, resident work hour restrictions etc.). I enjoy tech, I truly enjoy gadgets, but having been to one physician as a patient and talking to their back while my information was being input into the computer (which was clearly more interesting than me, the patient) I have strong opinions on the subject.

    I also have the input of one of my colleagues who is deeply involved in the EMR for our parent hospital corporation, and his opinion regarding the EMR implementation at one of our sister ED’s. He tells me their efficiency fell off 30-40%, and that’s with scribes doing the data entry. He figures for our volume, which we can currently keep up with well using a paper chart and one scribe, we’re going to have to have two scribes, and do a leap-frogging arrangement.

    That kind of efficiency I can do without. But apparently, we can’t as this is being done ‘because we need to’, without any actual science to demonstrate if the principle actually works.

  3. Our ER went to an EMR a year ago and we have found that it slows down our processes about 20%. It is great at gathering data but the data is flawed on a number of fronts, garbage in garbage out. The charting for the nurses consists of ridiculous canned entries and the poor doctors have to type their whole exam, plan etc. I can see where there will be a place for this technology but we are definitely not there yet. Oh and yes, the patients biggest complaints are that they don’t feel like anyone looks at them

  4. I spent some time with the electronic mumbo-jumbo the other day and I can tell you the following:

    There is no electronic mumbo-jumbo if triage, or at least SOME one, hasn’t entered the patient arriving by EMS into the system. You can’t get into it. The patient has to be loaded into the system. In cases like that, I had to go back to the old style sheet of paper and write on it until the system was updated with patient info. Later, I had to enter everything into the system. Perhaps this is only relevant to the particular system I was using.
    Technology glitches while getting a history can be a pain and the system required a password each time you blinked.

    The system I was on had every doctor’s name in the US. Though a minor thing, I had to check and make sure it was the correct Dr. Ron Smith by knowing which city Dr. Smith practiced out of.

    Order tracking was nice. It’ll tell you how many labs are pending, how many completed, and if you put the mouse over it, it’ll tell you what was ordered and what is back. You can simply copy and past lab results into the chart rather than writing them (in my little world, lab results are written on the T-sheet). Patient privacy can be protected fairly well since an email can be sent to administration should someone unauthorized look at a chart.

    Healthcare providers who don’t have anyone to do the documentation for them, but want to provide eye contact, are simply going to have to develop good memories. Navigating the system is more time consuming than a circle or slash on a T-sheet. The T-sheet can at least jog your memory about something on ROS or the physical. The electronic system I used would only give you a blinking cursor.

    I’d imagine to the patient the typing would be as irritating as listening to the clack-clack of a keyboard while speaking to a company rep over the phone.
    “So you have problems urinating?” [clickity-clack x’s eleven million]. “Would you repeat that? I wasn’t on the right screen.”
    My experience left me with the impression that in some ways it’s faster, in other ways it’s slower and it works out to not really making things any quicker in general. Perhaps slower. I’m not sure it’s very well suited for emergency departments. Most certainly it’s less reliable.

    At least paper won’t give you a Blue Screen of Death.

  5. There are good EMRs and bad EMRS, EMRs configured for Billing and EMRs configured for checking the 5 rights of med administration, allergy checking etc. Learning how to use a computer with a patient takes practice, trying showing them the screen and walking through your actions, most people today would be fascinated and listen intently if involved in the process.

    Long term EMRs will change from frustrating, limited vendor solutions to Open Source systems with which physicians can get directly involved and customize them to exactly what is best for them and best for the patient.

  6. I agree, current EMR implementations are particularly awful for the general practice and emergency medicine settings, and rather poor in many other areas. To be honest, they work really well in ICU’s, at least the ones I’ve seen. Patients tend to stay in one spot, and having a computer terminal there hooked up to all the monitors is actually very handy.

    I have two points to add. The first is that pen and paper wasn’t necessarily all that great either in terms of encouraging communication. I’ve been to many doctors who spent the entire time behind a desk staring down at their pad and scribbling furiously. It can form just as great a barrier, and carries the added burden of illegibility!

    Second point is that current EMR systems are a necessary evil before we get to the good stuff. I’m in the “it’s crap, but let’s do it anyway” camp when it comes to electronic data entry and retrieval in healthcare… I think as we go along our technology will become much better. The most important thing to make sure of is that our clinical information systems adapt to us, and not the other way around.

    Easier said than done when large hardware and software companies are involved.

  7. Of course the early phase implementation of EMRs is fraught with problems, especially if physicians haven’t been given any training in, for example, how to interact with the patient while entering data. (I too have been on the receiving end as a patient, and in fact it’s actually more participative when done right than watching your silent doc scratching out god knows what kind of ominous comments about you.) But note that every one of the country’s premier physician-owned health systems has long since addressed those issues and swears by the pixel over the pen, as do their docs: Mayo, Cleveland Clinic, Geisinger, Palo Alto… or at least so they assure me. So that would seem to invalidate the contention that the EMR doesn’t work in practice. Washing your hands between patients seemed like a well meant but hopelessly inefficient nicety at one time.

  8. TheNewGuy says:

    In the opinion of this physician/geek, it’s all about UI. If the interface sucks, the actual data entry becomes excruciatingly painful, and the entire process slows to a crawl; everybody suffers. Once you have the patient data in a usable format (where it can be parsed, analyzed, pattern analysis done, etc), you’re golden. Getting it there will always be the problem.

    Right now, we’re brute-forcing the process by using dedicated scribes, but a sufficiently elegant UI could make that step unnecessary.

    It’s the eternal computer geek struggle of brute force vs elegant solution.

  9. Gruntdoc,

    You have just opened Pandora’s Box. We have just intiated an EMR. So far the results have been that the hospital can bill several million $ more per year in facility and nursing charges that weren’t captured before. It has crushed MD efficiency. At best we can see about 2 pts per hour in an ED where we were seeing much more than that with the dictated charts. We have yet to determine if MD reimbusement will go up. We were promised a 10-15% increase. I really worry about the blanket statements of examination. I think they invite a less than accurate chart of what “really” was done. I refuse to chart while talking to the pt. I think it is rude and unprofessional. Most ED docs only spend several minutes with each patient. You should at least make eye contact while you are in the room. We are trying to set up a scribe service to help with the data entry. It is amazing to sit back and watch the department function now that we have EMR. The vast majority of the shift all of the nurses and physicians are sitting at a station staring at a computer screen. Some of them (both nurses and MD’s) won’t get up for over an hour entering all of the ever so important data.

    I am not impressed

  10. Okay Gunner, enquiring minds and all that: what product / company? And, do you think a scribe (or two) to do data entry would help your throughput?

    I salute your not doing the computer part during the patient encounter, by the way.

    Now, what benefits have you found? Does POE help? Better turn-around between labs being done and notification?

  11. We are using Ibex.
    I think a scribe would really help. The days we have med students that do the data entry for us we can move the pts through. We (MD’s) have to do the order entry, work excuses, school excuses etc. Very time consuming. The chart itself takes 3-4 times longer than dictating. The lab system is somewhat better than before (printer that kicked it out when labs resulted). Overall it is just akward to navigate the system, doesn’t flow well. I am sure there are better systems out there. I am not sure we looked at enough systems before we committed. It is still too early to see if the billing will increase.