EMR: Round peg, Square Hole

A 57-year-old doctor I know is retiring to teach at a local junior college.  He is respected, enjoys practicing medicine and is beloved by his patients; therefore, I was surprised. While he is frustrated by the complexity of health insurance, tired by the long hours and angered by defensive medicine, the final straw is that he can not stand the world of the EMR.

As an Electronic Medical Record junkie, I would quit if I had to practice without a computerized information system.  These programs are a dramatic improvement over the paper and pen way of keeping records.  Still, I understand the onerous problems.  Data entry is clumsy, painful and takes hours.  Information is stored in a nearly random manner, not much better than papers tossed into a cardboard box.  Every EMR program is different and none share vital patient data.  Training is lousy, access is non-intuitive, support is spotty, costs are high and any gains seem to be countered by poorly timed system crashes.

Unhappy to lose a physician from our medical community, I find myself musing about what has gone wrong with a critical technology that has such shining potential.  Computer systems fly giant aircraft around the world without incident, handle trillions of dollars of financial trade without a penny lost and allow hundreds of millions to tweet, Facebook or blog.  Why is medical IT so bad?

The major problem with EMRs, as they are conceived and as they presently exist, is that they are round pegs in square holes.  They are designed to gather and store information; shiny electronic file cabinets, and they are built around the primary function of billing; grinding out ICD-9 and CPT codes.  That would be fine if that was what doctors actually do with their time and if making money was the primary goal of practicing medicine.  However, surprise, surprise, what doctors really do is treat patients.  EMRs often hinder, not assist, the giving of medical care.

A physician’s normal function is to interface between objective biology and the complexity of each human life. Often called “the art” of medicine, it is the act of bridging science to individual reality.  Ask questions; test; collect information.  Attempt to organize by creating of a list of possibilities, a differential diagnosis.  Assimilate, screen and sift that data until you reach a final diagnosis.  Then, implement therapy using science and the results of research, with compassion, patience and the skill of a teacher.

A functional electronic health delivery system would assist in this systematic decision process, actively participating in the query and analysis, adding scientific knowledge and observations based on state-of-the art recommendations.  Help the doctor build the differential.  Recommend testing or therapeutic alternatives.  The EMR should be aligned with the doctor’s goals, which are the patient’s health.

The GPS in my car is first rate.  Data input is verbal and flawless.  It tells speed, direction, and continuously adjusts recommendations based on my progress and traffic impediments.  It even throws in alerts about the weather.  In other words, the GPS not only stores data, it tells me what to do with it, and is constantly updated by events far beyond my windshield, which I have not yet considered. Someday soon, that GPS will actually drive my car.

A health computational system should have, at a minimum, the functionality of that GPS.  Easy data entry and access. Flawless expanding storage.  Clear output.  Actionable recommendations and observations, based not only on the patient, but on the science of medicine.  An EMR should be updated continuously by clinical information such as labs, vital signs and tests, as well as the most recent scientific discoveries, even if they are made halfway around the world, delivering at the bedside the vast resources of Big Data.  Help me care for the patient by complementing my work.

As the practice of medicine becomes logarithmically more complex with the expanding potential of genomic or “Personalized Medicine,” advanced information technology will be vital.  No doctor will be able to assimilate an individual patient’s genome and thousands of actionable variables into a differential diagnosis or comprehensive treatment.  The key will be real-time EMR support.

To date no one has taken the potential or complexity of EMRs seriously.  The assumption is that these systems can be built by cottage industries, with the result that there are hundreds of rudimentary programs, all grossly inadequate. The average GPS is far more functional.

This slowly expanding area of IT research is called translational bioinformatics, but there have been relatively few dollars invested by the NIH in the basic science.  Data input remains primitive.  We have no backbone on which to create a national network to maintain and track individual records.  There is no integration with decision making software or connection to research troves.  Medicine relies on the doctor to connect the myriad dots, even as he or she is up at midnight, typing elementary progress notes into elementary office systems.

Doctors need and desire help in taking care of their patients, but instead they have a tool designed for secretaries and insurance auditors.  We must re-address the goals of clinical IT to improve, empower and give medical care. The future of our patients and the future of health, depend on it.  No amount of frustration and burned out physicians will force patient lives into slots built for dollars.


  • Liz
    Hope that local MD has a boatload of patience for his new set of problems that are endemic at many junior and community colleges (and some 4 year for that matter). It is very possible (and likely) to spend 60+ hours a week (more when you are grading papers) on teaching, get frustrated with students who cheat, don't do their work, try to make their bad grades your fault, etc… and in many instances faculty get to deal with an equally primitive online platform often required to be used for everything including grading, uploading syllabi, papers, etc. (it is faster in my opinion to hand write comments on a student's paper than to deal with many of the online grading platforms)… The problem with change for the medical industry (and many other industries) is that the sunk costs and costs of change are too high. When all the records are in a system that is not compatible with a new, better system then how the heck do you cheaply and easily port them over? This is like the problem (which I am in the middle of) porting over a pile of files from an old computer where the new MS office does not read the old files (many of which are in macwrite and writenow, not MS word) and so I have to open each of them multiple times with older, but increasingly newer, versions of these programs, saving and resaving until finally I reach a version that the new office opens. Saving in rich text format doesn't solve that as I often lose much of the formatting. Changing everything to pdf files could work except there is nothing I can find that works on a program that old. I could pay about $1000 to get someone else so to it but I don't have that kind of money… Medical records online with a zillion different programs in use has that same cost of change. As a result I don't see this problem will ever go away any time soon unless the government legislates portability of records between programs (sort of like demanding internet compatibility to a set of standards and then it is the problem of the writer or the browser to be compatible). It sounds like by the time a solution is developed for medical records the cost of and disruption of change is going to be tremendous. Unless it is legislated, I'd gamble that the financial cost of change is going to be so high that few will be able to afford it. You see that all the time with college registration and transcript programs using 30 year old programs (heck in 2010 I worked for a place that used DOS to enroll students!!! - changed over to something new in 2011 and for about 6 months we had a big mess and for another year bugs were still being dealt with in the system and during all this it took about 3 times longer to do something than before); other businesses that are using outdated programs and systems because changing them would be too disruptive to their daily business operations, would be too expensive (both in electronic upgrades, hardware upgrades, human "downtime" learning the new programs and switching everything over…). The consequences are of this kind of disruption can be catastrophic for businesses operating with slim margins. Yet with change they have more efficiencies which brings down costs, increases competitive advantages, access to better data mining which would give them a better competitive advantage in the long run (presuming they knew what to do with the data), etc. And while no one is likely to die, businesses may go under and so many people would no longer have jobs, products people own no longer supported... The medical profession is experiencing the downside to computer age just as businesses have. When they first introduced computers to secretaries they sat on many desks like expensive paperweights as the training wasn't there for people to learn how to use them, the learning curve was steep, and meanwhile their work load didn't decrease while they learned how to do their job on the new toy on their desk. That their job, in the end, would be easier was not relevant to them at the time because the rest of the world didn't come to a halt just because they had to learn a new way of doing things. Sounds like that MD is/was caught in the same catch 22. Like some of the secretaries, including some very good ones. who changed jobs, retired, quit… so has he. Hope he doesn't discover the "grass is always greener elsewhere".
  • Good post, good comparison to GPS performance. My concern has been that in EMRs as in building cars, we need things, and do not have people skilled enough to make them. A maker can recall a car and try to fix the error. A hospital-load of patient records is tough to recall if an unskilled, lazy, or careless system designer doesn't demand to know what doctors and nurses (and we patients) need.
  • me
    I wish your doctor friend the best in his new career. I suspect there are more reasons for his leaving medical practice than his difficulty dealing with EMR. As for the 'new problems' he will encounter in community college teaching they will be real but what also will be real is the feeling of accomplishment when he is able to help his students move on. Having spent 31 years in community college teaching I can agree that the daily job took many more hours than I initially anticipated however I can also report I met so many wonderful students that the work was well worth the effort. I'm surprised to hear that software development has lagged in this area since there is such an obvious need for advancement. Profit motive is usually the moving force and in this case it would seem that with such a need a solution should be coming down the line. I is always a pleasure to read your blogs, thank you!
  • Hello Dr.! My Dad is 91 and said he has a form of lymphoma..15 days of radiation..no chemo...he is a vet who was a medic...had 9 children, was a police officer..Ford plant employee..eventually independent truck driver..big and old fashioned kinda guy. We were always at odds because I was a real different kind of woman from day one...how about this?
  • D Someya Reed
    To say that "no one has taken the potential or complexity of EMRs seriously" is the equivalent of telling the medical industry that no one has taken curing cancer seriously. What you suggest could not be done even with a functioning "Star Trek style" medical tricorder (which they are trying to build; mostly by offering contest awards). To use another Star Trek reference, your comparison of a GPS unit to the EMRs you envision is like constructing "a mnemonic memory circuit with stone knives and bear skins." What you fail to understand is that before anyone can make anything even in rudimentary version of what you envision, there would have to be agreement among all doctors, all hospitals, all insurers, all pharmaceutical companies, all research labs, all governments, etc, etc, etc, etc to use a single computing platform, the same software, the same telecommunications methods and methodology, allowing full, live 24 hour access to all research and results and on and on and on. Yet no one can seem to even agree on diagnostics or treatment regimens. To take it a step further...how will they make this cost effective for all involved because it's not really all about the patient in the "real world." You've written about it yourself. It's about cost vs. benefit. No system capable of this will ever be considered cheap enough to implement (at least any time soon). The world is going to have to make some huge technological advancements before such a thing could happen. And really, are you going to compare your GPS unit with its limited, definable variables to the hundreds (some have even said thousands) of variables you hear every doctor state exists for every form of cancer at every forum when the question is asked, "Why haven't you cured cancer with all the money that's being spent on research?" It's fine to dream. It's not fine to bad mouth others for their lack of (perceived) accomplishment just as doctors have been bad mouthed for their lack of a cure. Throwing money at a problem doesn't solve everything (or anything). If it did, we should have that cure.
    • Malcolm P
      I think the attitude in the post by D Someya Reed reflects how successful many EMR vendors have been in convincing people that all this is too hard. I agree that it's unlikely that a single system will do what the article seeks, however the lack of deep interoperability in Health IT means that progress is slow because EMR systems are usually closed boxes of data rather than platforms for innovation upon which incremental progress can be made to the broader vision. This doesn't require huge technological advancements as much as enforcements of how key clinical concepts are represented, and enforcing 2-way interoperability so rather than buying an EMR you are buying EMR services which can be used (safely) by multiple applications or other systems. Unfortunately Health IT in 2014 is like general business systems in the 1960's where lack of interoperability and semantic standards are used as a lock-in strategy. Until purchasers start forcing companies to interoperate at a deep level (not just sending out HL7 messages) we won't be able to progress.
  • http://doctordalai.blogspot.com/2014/10/epic-fails-and-deadly-it-cultures.html One of the main problems with EMR's and other medical software such as PACS is that those who BUY the software are not those who USE the software. That little disconnect takes away all incentive for the programs to actually work.

Leave a Reply