What’s Wrong With This Study?
Tests such as CT scans and ultrasounds add to hospital bills, but doctors said that such tests given right after patients showed up in emergency rooms only helped with diagnosis in roughly one of three cases, the study — published in the Archives of Internal Medicine — said.
Full article on doctors requiring fewer tests to diagnose patients.
I don’t know that anything is wrong with the study, but there is something very wrong with the headline message. A one in three chance of saving your life is a very reasonable risk reduction to pay for.
One in three sounds worth it to me.
I’ve heard that Emergency Room doctors are so rushed that rather than take the time to talk to patients it becomes natural to just order a battery of tests that will speed the diagnostic process. Diagnostic tests also provide an acceptable excuse to move on to the next patients while the tests are being performed on earlier patients.
Economists and health policy wonks often make the blanket assumption that diagnostic images and tests add a lot to costs. I don’t believe this is necessarily true. In reality, these tests add a lot to charges, but don’t necessarily add to the hospitals cost unless performed outside the hospital. In a cost-plus system or a fee-for-service system, additional tests add to costs. In an integrated system, these would be absorbed the cost for performing them would fall.
It’s not 1 in 3 on saving your life. It’s 1 in 3 on diagnosing what’s wrong. Not everyone who goes to the emergency room is dying. Furthermore, every time anyone that I met ever went to the emergency room, they always knew what was wrong with them. 1 in 3 sounds like a lot of dart throwing, especially considering the additional evidence that people often provide upon admission.
The 1 in 3 number is interesting, but needs further context. To make a judgement on the value of the tests, it would be necessary to look at:
1. The costs (presumably for the insurer, unless the patient is uninsured)
2. The benefit of the diagnoses (ie, were these potentially life-threatening/harmful conditions that were diagnosed, or was making an early diagnosis not particularly important in helpin hte patient avoid harm).
3. The harms patients may suffer as a result of the tests (such as a misdiagnosis or causing the physician to ignore a more serious condition).
I realize that these outcomes (save cost) may be nearly impossible to measure, but these are the conversations we need to be having when we talk about treatments and testing.
I’ll preface my thoughts with the caveat that while I will be a physician one day, I’m not one yet. Take this with a grain of salt.
First, this study looks only at diagnosis. It’s possible that the diagnosis could be made just from the H&P, but that the imaging is of use in guiding management or therapy.
Secondly, in my limited experience shadowing in EDs, it’s not uncommon for imaging to come back before the “routine labs” that together with H&P accounted for 90% of the diagnoses. Stat images can be ready to interpret before stat labs, though this might vary from site to site and test to test. It could be that the emergency department would have made the diagnosis with just H&P/labs, but … well, it was an emergency, and the labs didn’t come back quickly enough.
Moreover, these patients are those who have been admitted, which limits the ability to make inferences about the appropriateness of imaging ordered in the ED. Admitted patients tend to be sicker than those “treated-and-streeted” in the ED. Patients treated in the ED and sent home weren’t included, which makes it hard to draw a conclusion of “overuse.”
Other examples include conditions whose diagnosis becomes more clinically apparent as time progresses. Take for instance (possibly… I’m not a doctor yet) epidural hematoma. If someone comes to the ED 20 minutes post-major head trauma and seems perfectly lucid, they could still have a life-threatening brain bleed. They get sent to the scanner. They don’t have epidural hematoma, but are admitted for [some other reason]. The admitting physician doing the H&P a few hours knows this, for the simple reason that they’re not being admitted to the neurosurgical service. Moreover, the passage of time makes it more clear whether this is or isn’t an epidural hematoma.
That said, I think there is something to be said for a solid history as the basis of diagnosis. I would hesitate to leap from there to “ED doctors are ordering too much imaging.”