11 May 2011

Can we save costs by reducing ER usage? A counter-intuitive take.

When we look at the $2,300 Billion that the US spends every year on health care, and the annual rate of increase of about 5% (inflation-adjusted), it's fairly clear that we need to get costs under control, or health care will bankrupt the country. One area of focus for possible cost savings has been the Emergency Department. The ER is viewed as a very expensive place to provide care, so that's not surprising. It's also viewed as overcrowded, which is true, and that it's used inappropriately by many patients, which is more debatable. 

On that point, there are conflicting perspectives. I have criticized ACEP in the past for promoting misleading statistics, when they have claimed that only 7% of ER patients are "non-urgent." This was based on frank misuse of the triage categorization system and simply does not pass the sniff test for anyone who has ever worked in an ER. A better estimate might be 25-40% of ER cases as non-urgent, or representing conditions which could have been treated elsewhere. However, I'm not sure I'd call this "inappropriate use" since for many patients there is no alternative option. The barrier to this is multifactorial. Unfunded or underfunded patients have limited avenues to access for semi-urgent care because clinics simply turn them away unless they can pay cash. Even funded patients are challenged to find primary care physicians who are able to see them in a timely manner, a problem exacerbated by the shortage of primary care physicians. Furthermore, hospitals in many areas have been complicit in driving up ER utilization by agressively promoting the ER as a convenient place to receive low-wait or no-wait care. 



So, it's a fair question -- are there potential savings to be realized in diverting low-acuity patients away from ERs? Legislators in MA, SC, and WA think there might be, and there have been suggestions that we should take measures to divert the less-urgent patients from the ER, perhaps using payment reforms or financial incentives/penalties to do so. Just yesterday there were hearings in the Senate HELP committee on whether non-urgent use of the ER can be reduced.

The question is whether this would save much money. ACEP says "no." They claim that the Emergency Department care accounts for only 2% of the nation's health care budget, and that reformers won't find much fat to cut here. To me, that 2% number also sounds implausibly low. The nation's ERs see 124 million patients annually. The average charge is about $5,000 for an ER visit (doctor and facility, source: PDF). Assuming a standard discount between charge and cost of about 70% (hard numbers on this are very hard to come by, but it's a good estimate) gives an average cost of an ER visit at $1,500. So that suggests that actual cost of ER care is $187,000,000,000. Given the total health expenditure of $2.3 trillion, we're more like 8% of the total health care budget. ACEP bases their claim on the respected AHRQ survey which found that the 124 million visits cost only $47 billion, or $376 per visit. While I greatly respect the AHRQ, this also, unfortunately, doesn't seem to pass the sniff test. I'd be interested to see a rigorous estimate of the actual cost of emergency medicine's services, and I suspect that it's somewhere between the two estimates. Which means that it's a non-trivial amount of money to be considered, somewhere in the tens to hundreds of billions, and a fair target for cost reductions, if there are any to be had. 

Which brings us to the next question. Since there may be saving to be had, would reducing the number of non-urgent ER visits do much to realize said savings?

This is where I can get back on the ACEP party-line bus, and say, I don't think so. The general concept here is that the marginal cost of an extra ER visit is very low. The ER is already there and the lights are on, the doctor is already there, and the CT scanner is already there (and paid for in advance). Yes, you need to scale up a bit as the patient volume expands, but a study, admittedly a few years old, found that the marginal cost for a low-acuity ER visit was about 7% of the average ER charge, or about 20% of the actual cost of that low-acuity visit. This makes sense, in that the typical low-urgency visit is one that is low resource utilization. It takes fairly little time and almost no ancillary services to diagnose a UTI or a sprained ankle.

Myles Riner, who I know through the ACEP reimbursement committee, put some similarly provocative data over at ACEP's Central Line blog: Much Ado About Very Little – the Deferral of ED Care Boondoggle. He presented some data which he had reviewed from Anthem Blue Cross, which looked at costs attributed to a year's worth of non-admitted ER patients. This includes 637,000 patients, at an average cost of $1,500 per patient, which is much more in line with my real-world experience. His findings mirrored the data-clustering finding I posted last week, in that the most expensive patients are far, far more expensive than the least:

So, assuming (and it's a big assumption) that we could take the bottom 40% of ER patients with regard to their cost -- and I think it's reasonable to assume that cost, acuity and urgency are reasonably correlated -- and successfully redirect them elsewhere, we would save a grand total of 12% of ER costs. Depending which estimate of overall health care costs ERs represent, that would net the economy a savings of $5 billion to $20 billion. (If you could convince only the bottom 20% to go elsewhere, which might be more realistic, then you'd save between $2 and $7 billion.)

That's a drop in the bucket, folks. And it would require major, major restructuring of the health care system, because these people represent nearly 50 million patient encounters. Some of them would give up and not seek medical care if the ER was closed to them, and that would be OK. (I'm thinking of the sunburns and such, the truly trivial complaints). Some would fail to get medical care and get worse; a simple UTI would turn into a case of pyelonephritis, and then present to the ER sicker. Most, however, would need to go to their primary care office or to an urgent care clinic -- resources which at this time either do not exist or do not have capacity to care for this huge volume of patient encounters.

It would also be very difficult to accomplish. Many times it's impossible to predict in advance whether a given ER case will be easy or hard. That blood in the urine could be a simple UTI or it could be renal cell carcinoma. The baby with the rash could be a viral exanthem or it could be meningitis. The heartburn could be heartburn or it could be unstable angina, and you'll never know until you are seen by a provider with the experience and judgement to differentiate the two. So, prospectively, I have no idea how to sort out the 40% who could go somewhere else, even if they had somewhere to go, which they don't, and even if they could be easily deterred, which they can't.

Furthermore, I'd make the argument that, especially for low-resource-utilization patients, the ER provides an efficient, high-value service. In a well-functioning ER (sadly, not the majority of US ERs), a fast-track type patient can be in and out in 45-90 minutes. No appointment needed. Just show up, get seen & treated. For someone who is in pain, or has a challenging job schedule, or child-care responsibilities, there is real value to that sort of service. If you have to take a day off of work to get evaluated for a minor medical problem, that cost is substantial; an ER that can provide the needed care may wind up costing not just the patient but the economy less in the long run. Yes, it costs more than a doctor's office, but since the doctor's office is unavailable for this sort of care for many people, I don't see the validity to that comparison. The fact that a low-to-moderately complex medical problem can be addressed quickly and cheaply in an ER is a measure of the success we have in operating efficiently. 

The best policy, of course, would be to have a network of primary care and urgent care offices which have excess capacity and are open for extended hours, if not 24/7. That would be a better and cheaper way to deliver acute care for a large minority of patients. That doesn't exist, unfortunately, and until it does, the ER remains not only the safety net, but a relatively low-cost alternative for patients who cannot access care elsewhere. Legislators and policymakers are misguided if they think they can achieve significant cost savings by reducing ER utilization. 


  1. So many things to love, & challenge, in your post.

    LOVE your "pareto" treatment of ER economics (few patients generate the lion's share of cost), so resonant with Gawande's recent New Yorker article on 's work in Camden NJ. Inexplicably few analyses of "what's wrong with health care in the US" ever even sniff this vital reality.

    CHALLENGE your notion that there's little to be gained from reconfiguring care so fewer people wind up in ER.

    Reconfiguring care obviously won't drop ER's share of spending from 8% to zero. Less obviously, but far more importantly, it WOULD change behavior around when, where & how care is obtained.
    To invert 's famed dictum: "different is more" - that is, different non-emergency care is qualitatively "more" health, thanks to the "better" health behaviors it potentially produces.

    Said another way, it's not the comparative UNIT COSTS of care that is important, tho it shouldn't be ignored; rather, it's what diverting care to more appropriate venues means for and does to care behavior, systemically, that's important, and potentially economically significant.

    Finally, LOVE this concluding bit: "The best policy, of course, would be to have a network of primary care and urgent care offices which have excess [we could quibble about "excess", but let's let it go for now] capacity and are open for extended hours, if not 24/7. That would be a better and cheaper way to deliver acute care for a large minority of patients. That doesn't exist, unfortunately...".

    The thing is, it MAY exist one day, if, as you have done, more people think differently about the challenges facing our health system.

  2. I agree with you on this one. I used to work in an ER for a number of years and that has always been my take on it. It always strains resources and flusters the staff, but in terms of actual resource utilization simple low acuity patients really don't soak up that much. And more to the point, even if we could somehow eliminate the cost of 60% of the patient load an ER sees (across the board) that would only serve to match cost increases - essentially, treading water for a 60% reduction in one aspect of medical care. Should we strive to be more efficient? Absolutely. Do low acuity patients (especially those who can't or wont pay and illegal immigrants) cost the ED money? Yep. Would that really make a big enough difference to justify yelling and screaming about the illegals and make instituting sweeping changes worthwhile? I think not as well.

  3. I think a bigger problem is doctor abuse of ER.

    We get patients all the time who see their doctors in the office, the doctors realize they need to be admitted and rather than call the hospitalists and arrange an admit just tell the patient to go to the ER and get admitted.

    Part of the problem is our ER docs will accept the patient when and if they call instead of encouraging direct admits. Our hospital has done nothing to streamline admits to the hospitalists.

    Also the urgents cares must be staffed by morons. Example, I got a patient from an urgent care yesterday who was sent to the ER because she had a subconjunctival hemmorrage. No other problem, vision was fine, no pain.

    Another big cause of the log jam is the ER docs who keep doing the same thing expecting different results. Same few patients getting the same tests over and over again with negative results. The chronic abdominal pains who come in and occupy beds for hours and hours. Healthy young patients who say they've been vomiting for days but have no ketones in their urine and heart rates under 100 getting IV hydration and lab. Give them a script for phenergan suppositories or zofran and say see you later. Don't tie up a bed for 3 hours.

    Its not the low acuity patients that are the problem. A big part of it is the low acuity patients being made into high maintenance patients.

  4. Or live in Sweden where we have primary care open until between 10pm to midnight and at the well run primary care centers the possibility to ring up in the morning and get a time that afternoon or the next day. But then you have to have a government run system that is organized to give the best care to the most number of people, as opposed to the US system where some get the very very best/over the top/unnecessary care and many need to come to the ER for their UTIs (and get good care of course).

  5. I am curious about the financial impact on the hospitals themselves. If you could magically remove 40% of the patient population you would likely eliminate a large source of revenue. I have had to go to the ED for things like severe cuts, broken bones and migraines that could have been treated elsewhere except for medical advice or lack of options. As it's an order of magnitude difference in billing between urgent care and the ED with insurance how can a hospital afford not to have me as a patient? The doctors may hate it, but their job depends on having all those unneeded patients. If they lose money now, what would it look like if there was effective utilization?



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