11 August 2010

In which I shall depart from the party line

The conventional wisdom has been that as the crisis in American healthcare has deepened, as the number of uninsured Americans grew and the access to primary care dwindled, the nation's ERs have been choked with patients seeking primary care, with non-urgent complaints, with trivial stuff that chokes up the departments and distracts staff from the truly ill patients with real emergencies.

ACEP and the house of Emergency Medicine have pushed back hard on this perception. They contend that most ER patients are in fact in the ER appropriately and that the real problem is a lack of inpatient beds which force admitted patients to be boarded in the ER, thus exacerbating the overcrowding crisis. ACEP is right to focus on patient boarding -- it's a real problem, and it's a more tractable problem. But I've always thought it strains credulity to claim that there are few non-urgent patients in the ER.  It actually seems a self-serving bit of fiction, designed to protect the turf, protect the business, to justify further investment in the ER.

I saw this triumphant tweet from ACEP's public affairs office yesterday:

Twitter / ACEP: Full CDC report, only 7% n ...

And my initial response was: hogwash.  7% of patients non-urgent?  No way.

So I started by actually reading the linked publication: the National Health Statistic Reports 2007 Emergency Department Summary (warning: PDF). The methodology is simple: they looked at a statistically valid sample of ER visits and complied the descriptive data, including triage level.  Apparently everyone uses the 5-level triage scale now, or enough that they could extrapolate to national figures, and I am sure that their statistical prowess far exceeds my own, so I'm not going to quibble with the results.  As stated, only 7.9% of ER patients were triaged at the lowest level of urgency.  The triage scale, for those not familiar, is this:
Triage Levels
Sensible enough.  So are the PR guys at ACEP right?  Are 92.1% of patients in the ER in fact, emergent?  Maayyybe. But I think not. Note that green category, "semi-urgent."  That includes another 21% of patients. Now this is where things really get subjective. What's the difference between a level 4 and 5 triage?  What's the difference between Semi- and Non-urgent?  I have no idea. Sure, there's a definition (green means the patient should been seen in 61-120 minutes), but in my experience the triage nurse simply picks a level kind of arbitrarily when the patient is on the low end of the spectrum.  Quite frankly, the nurses tend to use green as "non-urgent" and blue as "so fricking non-urgent that I am mad at you for coming to the ER."  So my contention would be that the more accurate interpretation of the NHSR report reflects the reality that somewhere around 29% of ER patients are not true emergencies.

This is also consistent with the larger trend we have been seeing in medicine - the rate of ER visits over the last decade has increased at twice the rate of population growth. To some degree this is due to the aging of the population and the increase in the number of Americans living with chronic disease. It is also, I suspect, due to the slow death of primary care and the rise in the number of uninsured and medicaid patient who effectively have no access to primary care. (It's worth noting that the uninsured patients in the NHSR report skew to the lower acuities.) 

This is also consistent with my experience. It's reasonable to take triage acuity as a sense of whether a patient had, prospectively, a potential emergency. But when you do look retrospectively at diagnoses, it paints a very different picture. Of the 22 million pediatric visits, about 9% were due to otitis media and URI's alone. When I look at the most common diagnoses in our ER (a relatively high-acuity ER) I see a lot of not-exactly-emergent diagnoses on the list:
  • Lumbar sprain
  • Bronchitis
  • Headache
  • Migraine
  • UTI
  • Unspecified Viral Infection
  • Lumbago (really?)
  • Otitis Media
I admit that it's hard to sort of from a list of diagnoses whether a patient really "belonged" in the ER. A 18 year-old female with cystitis and an 88 year-old with urosepsis might both show up under the same primary diagnosis. But still, lots of "urgent care" stuff there.

Now none of this is to say that I blame patients for coming to the ER with these complaints. Generally they are doing the best they can in the system we have. If there were primary care docs and urgent care clinics who could care for their urgent problems, many would choose to go there. Even more would choose to go to urgent cares if they had an economic incentive to do so -- like a higher co-pay.  However, the ER is "free" for many and we are open 24/7 with no appointment needed, so we become the convenience clinic in addition to the safety net.

In an ideal world, we would have a front desk with two doors behind it: one to the main ER and one to an urgent care clinic.  Or even if we had the ability to screen patients and refer them to clinic appointments.  But EMTALA makes such endeavors dicey: you make mistakes with that sort of policy and you can be in a world of trouble.  Moreover, the economics of office-based practice are so marginal and the crummy payer mix of the average ER makes it worse, so few hospitals or medical groups are eager to open up such urgent care centers. So we are stuck seeing all comers, the emergent as well as the urgent and the semi/non-urgent.  It's not an ideal situation, but it is not going to change any time soon.

It does not behoove ACEP, however, to make and persistently repeat claims which are false or misleading about the acuity of the patients we serve. It reduces our credibility in the policy debates and wastes energy on an issue which is a distraction from the other (more tractable) issues challenging the emergency care system.


  1. As a nurse who often triages, I just want to support your assertion that there are an awful lot of non-urgent visits under category two. Your description of cat. 1 is spot on. For example, I usually reserve level 1 for pt.'s in whom I cannot find a medical complaint: homeless pt. who wants socks and a sandwich, pt who came to get a dental check-up, parent who wants sports forms completed etc. Any sort of non-urgent medical complaint gets a 2: common colds, simple rashes, chronic pain management, etc. Some appropriate ER visits also get level 2's: simple suturable lacs, bad ankle sprains, abscesses. Level 3 is when you get to pt.'s who really needed to come to the ER: appys, CP, pneumonia etc. I think about a third of our patients don't need to come to the ER.
    -whitecap nurse

  2. Was thinking about your last part, re "In an ideal world, we would have a front desk with two doors behind it: one to the main ER and one to an urgent care clinic."

    Maybe 15 years ago I had driven 400 miles to Boston with what later turned out to be an abcessed tonsil.

    This generated two visits to Beth Israel's Walk-in clinic; the first had relatively relaxed timing; the second (after two days of what turned out to be the wrong abx) had them ushering me into the back before I'd even walked across the room to sign in and sit down*.

    As far as I could tell, the walk-in clinic was a separate triage for those sev1, sev2, and maybe sev3, but the treatment area was still the same ER as for those who came in the ER entrance.

    I was still on parental insurance at the time, so I don't really know how billing worked - I know I didn't have any bill at the time. And since the staff seemed to be from the same pool it doesn't help with general overloading.

    But perhaps an approach like this works to prevent what you linked to in Chicago?

    * Which, given that swallowing wasn't working so well as I walked in from the parking lot, is in retrospect unsurprising.

  3. our triage nurses only assign a "5" to suture removals and prescription refills. we send 4's and 5's to fast track. i would say about 40-50% of our patients qualify for fast track.

  4. Four and a half years ago, I was in the ER for what turned out to be a non-urgent complaint. (First trimester miscarriage; I needed a rhogam shot, I was totally distraught, and it was new years' day.) Because of the possibility of a ruptured ectopic, I got triaged back fairly quickly, but once they confirmed a non-viable intrauterine pregnancy, we got to spend a lonnnnng time waiting for that rhogam shot. Which is fine.

    About an hour into our wait, we heard Quite a Commotion out in the hall. It turned out that a floridly psychotic patient who had been brought in after eating mouse poison had produced a knife from somewhere, cut through his restraints, pulled out his IV and his foley, and gone haring off down the hall bare-assed naked. . . and out into the waiting room. He was subdued and returned to his room, hollering and shrieking.

    Ten minutes later, the doc came in with my rhogam shot and my discharge paperwork. "Sorry about the wait!" he said. "Fortunately, after that little episode, a whole lot of people in the waiting room decided their problem wasn't that emergent after all! I wonder if I can hire that guy to come back next Saturday!"

    That was a horrible day, one of the worst of my life. But I smile whenever I think of that doctor, and the position he must have been in to be GLAD that a crazy naked man slipped his restraints and scared all the non-urgents out of the waiting room.

  5. I guess the problem I see with looking at the stats from the final diagnosis -- looking at it from a patient perspective more than a medical professional perspective -- is that there are non-urgent conditions that you know are non-urgent, and then there are conditions that you only know are non-urgent after a little patient education. There really needs to be some way to differentiate between those two categories when looking at this sort of thing. Although admittedly individual education level, etc. makes it a bit tough to standardize.

  6. Sigh ... there is actually more logic to the ESI system than that. Or, at least there is supposed to be.

    OK, at the risk of being a total geek, here goes.

    The first "decision point" is based on severity: basically the "is this person going to die without imemdiate help"? These are the RED patients, at the highest triage level. Cardiac arrest, severe resp distress, STEMIs, anaphylaxis. These patients cannot tolerate any time in the waiting room.

    Patients that don't quite fit that category go on down the line to the next severe category. This would be your level 4's. ESI designates these patients as "concerning," to to either pain, mechanism of injury, or concerning VS -- things like that. This would be chest pain without the STEMI, shoulder dislocations, high fever, and the like. ESI puts it this way: these are the patients that can tolerate some time in the WR, but you would give your last bed to them.

    After that point, you are triaging based on predicted resource use. Category 3's will need more than one resource. Category 2's will need exactly one resource. Category 1's don't need any resources.

    A resource is defined as: lab, Xray, parenteral injection (im, subq, iv), consult, ekg, resp therapy. Each of these equals one. So, if the patient needs 5 Xrays, that is one resource. If they need 1 Xray and a u/a, that equals two resources. (Point of care testing does not count -- bedside UPTs and blood sugars aren't considered a resource).

    So, a wheezing patient that will need a CXR and a breathing tx is a level 3. So is the abd pain that will need lab, ct, xray and fluid.

    A paient with an injury that needs an Xray is a level 2. So is the dysuria that just needs a UA.

    Where the patient that just needs and Rx, no matter how "real" their c/o is, is a level 1.

    So basically Level 5 and Level 4 = emergency.

    Anything below that is suspect. It might or might not be an emergency. I would say that a laceration that could be considered a level 2 could still be an emergency; where probably 80-90% of the abd pain we see gets a level 3, but isn't any more of an emergency this time than the last 4/6/8 times the patient came in for it.

  7. Right. And regardless of triage level, I find that most patients I see on a daily basis did not need to be seen by me. They could have waited a day or more to see their PMD or longer. Many I can tell within 30 seconds of talking to them that it is not urgent.

  8. In addition, I think that it is safe to say that a decent porting of your level 4s aren't really emergencies, either.

    Just because there is something "concerning" in triage (and I say this as someone who primarily works in triage), doesn't mean the patient actually had an emergency.

  9. In practice, there are really only 3 triage levels: Get 'em back now, get 'em back next, and get 'em when you can. This has been the case in the six ED's I have worked in anyway. Creating finer distinctions may look nice for admin, but does nothing for those in the trenches.

  10. Greetings. While I have to confess a "bias" here as I work on behalf of the urgent care side of things, we did dig deeper into this study when it came out as well (looking at the "back end" disposition of the visits), since we have long discussed the determination of the "levels" with our colleagues at ACEP - and how challenging it is for real conclusions to be drawn from them.

    We did a press release on what we took away from the full study http://www.businesswire.com/news/home/20100819006009/en/CDC-Study-Shows-Continuing-Urgent-Care-Centers if you'd like to peruse.

    It's our belief that urgent cares CAN help with some of the "likely don't belongs" in ERs - and all of us have the challenge of helping educate the patients about what decisions to make in the heat of the moment.

    That said, I agree that the boarding issue is a much more contributory one from everything I hear - though perhaps more challenging to address as you are talking about systems issues now. Throw EMTALA in there and...

    It's interesting to hear all of your perspectives - thank you for sharing. We do believe we are all in this together, and understand that you folks in the ER suffer quite a bit - hope someday we can be a good part of the relief effort for everyone.

    Lou Ellen Horwitz
    Executive Director
    Urgent Care Association of America


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