10 February 2009


Yesterday afternoon in my ER:

Me: Well, sir, I have some good news for you. The CT scan of your abdomen looked perfectly fine.

Patient: That's a relief.

Me: So, were you perhaps wondering why you got a CT abdomen, given that you are here for a fainting spell?

Patient: Yeah, that had me kind of worried -- I figured there must have been something really wrong.

Me: Well, there's a funny story to be told, and I'd like to offer you an apology. I was looking at your lab results and I was very surprised to see a CT scan result in there. See, the CT scan was ordered on the patient in the bed next to you, who has a kidney stone. And it appears that nobody noticed until you had already undergone the study.

Patient: Wow. How did that happen?

Me: It looks like it was my fault. I entered the order in our computerized system, and apparently I just clicked on the wrong patient name. I have no idea how I could have done that, but there it is.

Patient: But the scan was OK?

Me: Yes.

Patient: And no harm done?

Me: No, other than the fact that you got a dose of radiation you didn't need.

Patient: Okay, then. Good thing you were just ordering a scan, and not an enema!

Me: True dat.

The same thing happened again later the same shift. This time, I caught the error immediately, since I was double-checking. I'm still unsure if it was a computer glitch or user error.

So when President Obama said in yesterday evening's presser, "Why wouldn't we want ... an electronic medical record that will reduce error rates, reduce our long-term cost of health care, and create jobs right now?"

Well, I began to wonder.

It's taken as axiomatic that an EMR and E-prescribing reduce error rates. That has not been my experience. In the four years we have used one, I have found that we have simply substituted one type of error for another. In the old days, I would get calls from pharmacists who were unable to decipher a hand-written prescription from one of my partners. Now I get calls from pharmacists about prescriptions written incorrectly -- the wrong formulation selected from the drop-down menu, or the dispense and frequency fields filled out nonsensically. I got in a real pissing match with a pharmacist the other day about a prescription for Tylox (oxycodone with 500 mg tylenol) when they only carry Percoset (oxycodone with 325 mg tylenol). Apparently the computer had substituted the less-common Tylox in automatically, and these cannot be altered over the phone, so the patient was unable to fill his prescription and had to come back to the ER.

Don't get me wrong. I think we need an EMR, for a variety of reasons. I can manage my physicians with the data culled from it, we can view operational "dashboard" parameters of the ED, it does simplify and expedite order entry. On the whole, it is a strong positive for the ED. It may reduce costs by allowing us to do more with less resources.

But what it does not appear to do is reduce errors, not in and of itself. When politicians and policy-makers hold an EMR out there as a panacea for medical errors, it is a false hope and a false promise.


  1. And that is why *I* always ask if I think an order is incorrect. Sure a few of you ER dr types get pissy if you feel you are being questioned, but most are perfectly fine to take 15 seconds to say "yes I do want a PE study and an abd/pel on that patient."

  2. I'll be honest, I have no idea how those things work, but to me, that sounds like programming errors, maybe not the selecting the wrong patient, but definitely the medication errors. There should be safeguards in place that make sure you're not creating some new formulation (i.e. dispensing capsules at 220mg when the only formulation that exists is a 200mg tablet). It seems to me (again, no clue) that it should be possible to at least request verification from someone if the test doesn't match up with some other characteristics (like if you accidentally ordered a pregnancy test on a male patient).

    The hospital/chain (even independent) pharmacies should be linked to the point where inventories can be queried to make sure that what you're prescribing is available.

    Heck, even insurance could be linked to make sure that what you're prescribing is a preferred drug, rather than having the patient go all the way to the pharmacy to find out the copay's going to be $200 vs. $25 (although that probably gets into the whole insurance vs. doctor choosing the therapy debate).

  3. You can always write an order in the wrong chart - paper or electronic. There will always be errors - thank God this was minor one. Also, cudos to you to admitting you screwed up and did not blame it on someone else.

  4. Those who create such programs sell them to physicians by promising us they will increase our collections, but they sell the concept to the government by telling them that they will decrease their expenditures.

    EMRs are magical like that.

  5. It is refreshing to read of your honesty in admitting a mistake. That is one of the important parts of error reduction.

    Another is the removal of penalties for admitting mistakes. A third is prevention of discovery of internal quality improvement records by plaintiffs' lawyers. Many people will not be as honest, if they feel that it may cost them a lot of money later on.

    Politicians rarely admit mistakes, so they are not the ones to make intelligent decisions about error reduction.

  6. Please don't take offense, but why are we proud of you for being honest?

    (and moving rapidly along...)

    Why are we blaming the electronic record system for your error? As ERP points out, errors also happen with paper charts -- but are probably not so easily picked up as errors.

    When I got a copy of my hospital chart a few years back, I found someone else's heart rhythm strips carefully affixed and whole pages of orders for an elderly man in another room. Hours were wasted one day as the staff tried over and over to call the wrong doctor (same last name) to consult for my kidney failure. This was in the ICU. (I'm theorizing that an electronic system might show the choices for nephrologists, or at least point up the existence of identical last names in the list of physicians. The guy they ended up calling in was an ID specialist. Sadly, we ended up needing him!)

    Despite failing to give me stress-dose steroids before, during, or after surgery (hence, an Addisonian crisis ensued)-- the steroids were forgotten two more times in the following days (once in ICU, once on the floor). I tend to think that the electronic record might have flagged that. I mean, if getting a zero cortisol back doesn't cause someone to react... you know?

    Some system of redundancy has to be in place to catch all this shit.

    I blurted most of the story out over at Happy's place back in December, having finally decided that people need to know as much as possible about what can happen in hospitals -- where everyone is busy and not everyone is top notch.


    The hospitalization resulted in the declaration of a sentinel event and a big part of the hospitals plans for correction that had to be submitted to the state involved switching numerous protocols to an electronic record system.

  7. Bianca Castafiore,

    Please don't take offense, but why are we proud of you for being honest?

    Because so many are not honest, when honesty may result in misfortune. If we are to change things, we must encourage the good behavior as much as we discourage the bad behavior.

    It might be nice, if this were not the case, but evolution has provided us with a species that is not completely, nor even mostly, altruistic.


Note: Only a member of this blog may post a comment.