Four months ago I wrote an essay for this blog (Oct. 22, 2011, "Measuring Quality and Value?") about the government telling us how to practice medicine in the guise of improving the quality and value of medical care in the United States. In that essay I mentioned a new rule they want us to follow in deciding whether patients who come to the emergency department with headache should have a CAT scan of the brain to look for a serious cause.
As I have said before, when folks from the government talk about quality and value, they're really focused on dollars spent. Value is defined in terms of the relationship between quality and cost. When you improve quality while holding cost constant, or reduce cost without compromising quality, you increase value. Even better would be improving quality and reducing cost at the same time.
So the feds (meaning bureaucrats at the Centers for Medicare and Medicaid Services, CMS) told us we do too many CAT scans of the brain in our evaluation of patients with headaches (that were not the result of an injury). They said we should use certain criteria, which they delineated for us, in making these decisions. By doing so, they claimed, we could improve the quality of care by sparing patients needless and potentially harmful radiation, while simultaneously reducing cost. This is just the double bonus we should be looking for in the calculus of value. What a deal!
The problem with the new CMS rule was that the National Quality Forum (NQF), the highly respected body that reviews and recommends measures of quality, rejected this one. And the American College of Emergency Physicians (ACEP) said the CMS criteria were fundamentally flawed. And now my esteemed colleague, Dr. Jeremiah (Jay) Schuur, has published a study in the Annals of Emergency Medicine demonstrating the unreliability of the CMS measure. Dr. Schuur is the Director of Quality, Safety & Performance Improvement in Emergency Medicine at one of Harvard's major teaching hospitals. He knows his stuff.
There's someone else in this story who should know his stuff but apparently doesn't. And that is Dr. Michael Rapp. Dr. Rapp is the Director of the Quality Measurement and Health Assessment Group at CMS. Dr. Rapp believes emergency physicians are less than competent when it comes to evaluating headache patients and rely on CAT scans far more than they should. Rapp pushed forward with this "quality measure" despite the NQF's rejection and over the emphatic protestations of ACEP. Now, I must go off on a bit of a personal tangent here. I find this disagreement especially vexing because I have known Dr. Rapp for many years. He was president of ACEP a dozen years ago, and I first met him several years before that, when he was a candidate for election to the ACEP Board of Directors. What a grievous example of going over to the dark side!
I am a staunch proponent of the cost-efficient practice of medicine and as willing as anyone else to embrace a clinical decision instrument that allows us to avoid expensive testing without missing important diagnostic findings. Over the years I have learned a bit about the process of developing such aids to our practice.
Don't worry, I am not going to try to turn you into an expert on evidence-based medicine (EBM) by the end of this essay. But it's not that complicated. Let's start with a large group of headache patients. This is a group that has been thoroughly evaluated, and we know which patients turned out to have something bad and which ones didn't. We'll make a list of characteristics, things about patients that we can see when we interview and examine them, that we think might help distinguish those with something serious from all the rest. We'll look at our population of headache patients and see who has what from our list. Then we'll do some fancy statistical modeling to come up with a list of low-risk features, such that patients who have all of them were extremely unlikely to have anything serious causing their headaches. That is how we derive a clinical decision instrument. Then we have to validate it by using it prospectively, on a new group of headache patients, and see how well our prediction rule performs in identifying the ones with something bad. What we're really looking for is simple: if a patient meets all of the low-risk criteria, that should reliably predict that the patient will turn out to have a non-serious cause of his headache. If it works, we have a sound method of identifying a group of patients for whom we don't need to order CAT scans.
(In case you were wondering, yes, there are other tests we sometimes do for headache patients, like a spinal tap if we think a patient's headache might be from meningitis, but the focus here is on whether headache patients need CAT scans.)
At this point you should be wondering the same thing I was when I learned last year about the new CMS quality measure. Was it based on a well-validated clinical decision instrument for determining which emergency department patients with non-traumatic headache need CAT scans? And by now you can guess the answer: a resounding no.
So Dr. Schuur, my colleague at Harvard, did an excellent study to examine this new CMS rule, and he concluded: "The CMS imaging efficiency measure for brain CTs ... is not reliable, valid, or accurate...." I will not attempt to summarize his findings here, but if you really want to read it and do not have a subscription to the Annals of Emergency Medicine, drop me a line and I will e-mail you a pdf.
Those of you who have come to recognize the statement "I'm from the government, and I'm here to help you" as one that should strike fear in your heart will find none of this the least bit surprising. I'm sure you can now surmise what happens to me every time I think about CMS. Yes, I know. It's just a headache.
So the CMS rule is bad. We need a different, better rule, or no rule at all. Sounds fine.
ReplyDeleteI have to ask the question, though, what's your ideal system? Because if we're going to go forward on the basis of a principled opposition to government interference with physicians' decisions, we need to figure out what will replace Medicare and Medicaid. Because ultimately you wouldn't want the government spending hundreds of billions of dollars a year on something and deferring all questions of quality or necessity to the people selling the good.
I mean, the government spends more on healthcare than it does on the military. I certainly hope that there is someone other than Lockheed Martin & Co checking to see that the bullets fire, the rockets launch and the ships float. While the government is the largest payer for healthcare services, they are going to try and regulate what they are paying for. If you bill them, they will come.
It seems to me we have two options: find a way to get the government completely out of the business of paying for healthcare, or try and replace things like the CMS rule with the best science available.
I agree with you about the two options, so my choice is the second one. President Obama is a proponent of comparative effectiveness research. Let's do it, and let's use the evidence. We already have some evidence, of course, but we need a lot more. If CMS just makes up rules to fit the amount of money they want to spend - which my dear friend (and ACEP past president) Rick Blum has opined is how they do things - that may reduce cost, but it is unlikely to preserve or enhance quality.
DeleteI completely agree. One thing that's hard for outsiders to understand is that we need cutting-edge research into doing nothing, like we have cutting-edge research into new cancer drugs or scan modalities.
DeleteWhether it's not getting a CT or not placing a stent or not leaving old people who fall on Coumadin, doing nothing requires the best science available.
You can't patent doing nothing, and you can't bill for it, so "wait and see" is at an automatic disadvantage as long as the potential money from patenting discoveries drives a lot of our research spending.
The Onion's cogent take: http://www.theonion.com/articles/powerful-rest-and-fluids-industry-influencing-doct,2634/