In addition to calling it the “ER,” using cell phones in said “ER,” and engaging in baby talk, we can now add “scientific studies” like this to the list of things that drive me friggin batty.
The American College of Radiology published this study that purported to analyze the “appropriateness” of outpatient CT and MRI scans ordered from primary care clinics at an academic medical center.
In the study, researchers at the University of Washington used “appropriateness criteria from a radiology benefit management company” to determine whether CT scans and MRIs ordered by the lowly primary care physicians met “criteria for approval.”
Then researchers compared studies that did meet “criteria for approval” with those that did not meet “criteria for approval” and found that 26% of the studies ordered were considered “inappropriate.” The authors listed several examples of “inappropriate” studies such as obtaining a brain CT for chronic headache, obtaining a lumbar spine MR for acute back pain, ordering knee or shoulder MRI in patients with osteoarthritis, and ordering a CT for hematuria during a urinary tract infection.
Here’s the thing, though. The study states that “only” 24% of the “inappropriate studies” had positive results and affected patient management. In other words, if the researchers had not performed the “inappropriate studies”, they would have missed clinically significant findings in a quarter of patients. The conclusion of the “study” is that because the sensitivity of appropriate studies is higher than that of inappropriate studies, primary care physicians need help to “improve the quality of their imaging decision requests.”
Want some help? Here’s some help for you: Stop the Monday morning quarterbacking and create a policy at your academic institutions so that none of the lowly primary care physicians can obtain a diagnostic radiology test without the esteemed radiologist’s approval. Lowly family practitioners can order the tests and you researchers just veto them when they cross your desk. Think of all the money and wasted testing you’ll save. Oh yeah … then you can be legally liable for the bad patient outcomes when you don’t allow the test.
Why doesn’t one of you suggest that as an official ACR policy at your annual meeting in April?
Those tests don’t look quite so “inappropriate” when you don’t have the benefit of a retrospectoscope, do they?
P.S. Have family practitioners ever done a study to determine how many of the additional radiographic tests recommended in a radiologist report (i.e. “hip fracture present, cannot rule out pathologic fracture, recommend MRI and bone scan”) were retrospectively “appropriate”?
Frustrating to not be able to read the full article but from the abstract it certainly sounds like proof that the “criteria for approval” are overly restrictive and can result in patien harm.
Now if the study could show that specific unmet criteria always resulted in negative findings then they would be on to something.(Probably not possible given the sample size).
The other question not mentioned, in the abstract at least, is if any “100% criteria” exist.
Scanning to (only) confirm what is known from examination and other tests is equally wastefull.
Interestingly the only free article in current journal is:
“Why Radiologists Lose Their Hospital Contracts: Is Your Contract Secure?”
Sorry Whitecoat, I don’t think I will ever break the habit of saying “ER”, but I promise not to use my cell phone when I go to the ER…err..ED…ack, sorry, but that just doesn’t sound right. Too many drug-pushing commercials make me think of something other than emergency department when I hear or say the initials ED. Can’t help you with the radiology study either. .
Speaking of uneeded head CTs, did you read the PECARN study on children and head CTs?
http://www.pecarn.org/publications/documents/Kuppermann_2009_The-Lancet.pdf
They basically looked at 43,000 children and tried to come up with criteria to rule out intracranial processes. Retrospectively they got a 99.5% sensitivity.
My partners and I have been doing fewer scans, but I think we’re in a catch 22 here: scan and you’re liable for giving the kid cancer for ‘unneeded’ radiation exposure. Don’t scan and you’re liable for the 1/10,000 head bleed. How long before we see the lawyers advertising about it on TV?
Does a patient get radiation exposure from MRIs?
NO!
Thanks Doc 99.
Re: Radiology…Crayzee Central has a policy that ER MDs can’t order extremity MRIs. I hope you don’t think this is a bad idea…Before the policy, patients were getting MRIs for knee pain ROUTINELY. Waste of time and money!
Want some help? Here’s some help for you: Stop the Monday morning quarterbacking and create a policy at your academic institutions so that none of the lowly primary care physicians can obtain a diagnostic radiology test without the esteemed radiologist’s approval.
Sounds similar to what its like for doctors (and their staff) seeking prior-approval/prior-authorization for certain medical procedures from insurers/TPAs.
Yowza. A quarter of the “inappropriate” scans changed the course of care? That strikes me as proof the PCPs are getting it right; it’s the “appropriateness” standard that’s wrong.
I’d rather be lucky than good.
I’d rather be lucky than good.
Skill (being good) is reproducible when you need it.
Luck is not something you can count on.
Luck is certainly not something you want any professional to count on, unless it is your opponent’s lawyer, who is counting on luck.
WhiteCoat
The study you cite strikes me as an unrealistic solution in the real world, especially the non-academic one. The big reasons that radiologists, for the most part, do not refuse exams is 1) biting the hand that feeds us 2) not knowing the patient’s entire pertinent history and physical so I don’t want to be liable for refusing an exam on a patient I don’t really know. 3) A matter of practicality– If we had to thoroughly screen ALL requested exams then the backlog for a routine outpatient head CT would be 3 weeks.
On the other hand, clinicians have to be familiar with the appropriate reasons for ordering a certain imaging study. There has to be some sort of check system. I get requests from PAs and nurse practitioners and even PCPs who still don’t know the difference between a CT or MRI and when you should order one over the other for a certain problem. But instead of phoning us for a consult, often times they just order the suspect test. The ACR does have criteria but it of course, offered as a guide, with a big disclaimer on every page. I”m sure what the solution is but I know defensive medicine sure drives the process.
That’s just the point of the post.
It’s easy to criticize someone after reading a normal exam and wondering why the idiot doctor ordered it to begin with, but it’s another thing entirely to say I’ll use my own clinical judgment prospectively to say which tests shouldn’t be ordered.
I know that there are unnecessary tests being ordered – like the knee MRIs from the ED that Nurse K mentioned. How are those likely to change management? Maybe some of them warrant a call to the ordering doc to find out what they’re looking for and why they ordered the test.
I just think it takes a lot of gonads for these researchers to write a paper criticizing the judgment of the primary care physicians when 25% of the “unnecessary” studies had clinically significant findings.
That’s the challenge. If the primary care docs are so piss poor at their jobs, let’s see the researchers do better.
Why would the ACR even publish something like this?
Based on where the decision making process appears to be taking place, I think it’s not the ‘retrospectoscope’ they’re using but the ‘retroproctoscope’. The fact that a quarter of the ‘unneccesary’ studies showed stuff we needed to know should show that our hunches have at least enough value to warrant a test or three. Otherwise we might as well just pin a list of diagnoses up on the wall and throw darts.
WC – Here’s a thought, when the insurance company declines to pay for the “unindicated scan,” let them penalize the doctor who ordered the study, not the ones who in good faith performed and interpreted the study. Problem solved.
That is only appropriate if the scan is not not ordered with the reasonable expectation that it will assist in the diagnosis of the patient.
Does this study address that, or does the study just look at the rate of positive findings?
Fifty-eight percent of the appropriate studies had positive results and affected subsequent management, whereas only 24% of inappropriate studies had positive results and affected management.
While I only have access to the abstract, I would love to read the explanation for the way they determined that a study was ordered, but did not affect management. how do they know the thought process of the ordering doctor?
I do not mean to defend defensive medicine. I do want doctors to use their critical judgment. I would like to know how the reviewers determined that the decision was inappropriate.
Regardless of their reasoning, why is a test inappropriate, if it produces results that affect treatment 24% of the time? Using this number, without relating to a patient condition, is not helpful.
I think you just identified the problem with determining what is and is not “defensive” medicine.
The ACR has appropriateness criteria because it is a major organization for a large specialty but the criteria has little weight in practice. It boils down to how much uncertainty our society will tolerate. If we create and enforce criteria for let’s say ordering a CT pulmonary angiogram to rule out PE in a patient, and the patient does not fulfill this criteria and the clinician does not order the test– is he or she liable if the patient is found to have a clot later on?
Max, please comment on this. Some patients will always fall outside of “criteria” which is why the art of medicine trumps everything else which is why criteria do not help much. On the flipside some of the imaging requests I see are truly astounding. A knee MRI on an 80 year old with endstage DJD on his xray and chronic pain. A routine CT of Abdomen-pelvis for an acute GI bleed. The former will not change management and the latter is a waste of time, resources and radiation.