Sunday, July 11, 2010

Why Total Body Scans are Scams: Maze vs Bayes

Folks across the country are paying hard cash for total body scans, abdominal aortic aneurysm testing, CAT coronary artery scans and carotid artery evaluations to prevent disease or find important lesions early. It’s a seductive argument, and it’s a scam.

Ordinary patients don’t understand about pre-test probability and positive and negative predictive values. Indeed, all physicians were taught to consider Bayesian theory when ordering diagnostic tests. This is very tough concept for patients to grasp. A critical principle of proper diagnostic testing can be summarized in a single sentence.

If an individual is unlikely to have the medical condition under consideration, then a diagnostic test that yields a positive result is likely to be a false reading.

Here is an illustration demonstrating why patients need to understand this issue. While the forthcoming example is hypothetical, I guarantee that every physician has seen very similar patients in their practices. While the patient presented here has a presumed cardiology issue, every medical specialist and primary care physician can land in the same trap. When this occurs, patients suffer.

A 30-year-old non-smoker sees me in the office with chest pain that is readily relieved with antacids. It is very unlikely to be angina, and probably represents simple heartburn. If I arrange for this person to undergo a cardiac stress test, and the result is positive, then it is much more likely that  the test result is wrong than that the individual has true heart disease. This is not simply my opinion, but a conclusion based upon mathematical and statistical principles. However, try explaining this to a patient with a false positive stress test result. Despite the physician’s reassurance that the test result is erroneous, the patient will likely become anxious and remain unconvinced. Such a patient can easily slide, or be pushed, down a medical cascade that may include cardiac catheterization, or even stenting of a coronary artery that was not responsible for the patient’s symptoms, and should have been left alone.

The key is that diagnostic tests need to be ordered when the patient has a reasonable chance of having the condition under consideration. (If the physician is nearly certain of the disease, then the test may not be needed.) This determination is made on the basis of a careful history and physical examination. When stress tests and various scans are ordered casually by physicians, or requested by patients, then this opens a pathway into a medical labyrinth with no easy way out. Would you prefer to agonize over a false positive test result that pushes you toward medical quicksand, or avoid an unnecessary test in the first place?

Of course, there are rare individuals who have benefited from a scan that was ordered for the wrong reasons. These folks understandably are convinced that the scan saved their lives. These anecdotes, however, which make for potent testimonials, should not change established medical diagnostic principles. Every day, folks become millionaires after purchasing winning lottery tickets. Since nearly 100% of lottery tickets become bookmarks or end up in landfills, we know that this is a poor strategy to accumulate wealth. Should every person undergo a CAT scan of the head every year because it is theoretically possible that a few might benefit by accident? Dumb luck should not be our diagnostic touchstone.

Total body scans, and all of their cousins, are examples of medicine at its worst. It is a commercial enterprise that bypasses sound medical principles and judgment. These entrepreneurs proffer a promise that they know they cannot fulfill. It’s a scam clad in a white coat. For the majority of their unsuspecting customers, a positive result will be wrong and a negative result will guarantee nothing.  I realize that an ordinary patient will celebrate when his total body scan is negative, but this is not how medicine works.  You can have a normal EKG performed weekly, but this will not prevent a heart attack or exclude significant coronary artery disease. 

Still thinking about that cardiac scan being advertised in the newspaper or on television? Do yourself a favor. Buy some snake oil instead. The result will be the same, but you won’t waste nearly as much money and you won’t end up with a stent.

Many patients who have endured a ride on the medical cascade may feel that they were rescued from certain disaster.  I'd rather rescue folks from the cascade.

10 comments:

LeisureGuy said...

Don't you wish you could prescribe a course in Bayesian statistics for your patients and make sure they take it?

Of course, were that possible, I'd have an entire college curriculum ready to roll out for certain politicians...

Nicholas Fogelson, MD said...

Here's an interesting concept:

Right now what we provide to patients is greatly controlled by insurers, and many doctors moving to or thinking of moving to a cash only no insurance model. In theory, they are restoring the true doctor patient relationship by cutting out a middle man that is going to mediate the care. At the same time, the patient becomes financially responsible, in some way at least, for the medical care they consume.

In this new model, is it more likely that patients will consume medical care that is not good for them? If every CT scan were paid for in cash, would every CT scanner then start doing whole body scans for patients that requested them, or would there be some that chose to screen more carefully?


Another thought -

What about these screening lung CTs for long term smokers? I'm not sure of the performance of that test in this population, but it may be a similar issue.

Michael Kirsch, M.D. said...

Thanks commenters. I think it is inarguable that medical overutilization is fueled to a great extent by the patient not paying for the care. If the physician or the patient casually thinks that an MRI makes sense, the trigger is pulled since neither of them are paying for it. If the patient bore some financial responsbility for diagnostic tests, then this would serve as a brake on the system. I recognize, however, that this could also pose a barrier for truly needed care. A balance would need to be established.

Anonymous said...

Naked body scanners courtesy by TSA are also scam and almost everybody agrees to be irradiated.

Shawn said...

I disagree with a lot of the above. First, insurance won't pay for the prophlyactic use of body scanners, so the patient does have a financial interest in the procedure. Seoond, I think in some populations they are absolutely helpful in savings lives. Take for example: Mid-fifties, 40 pack year history with some occupational exposure, with two brothers diagnosed with late stage lung ca in their late fifties. Should I just wait for the inevitable bad news when it's too late to treat, or get the lung scans now and very probably catch it in stage 1 or 2. I think I'd choose the latter. Thanks.

Michael Kirsch, M.D. said...

Shawn, I disagree with you. Cost aside, I think you have more to lose than gain from a total body scan. I realize to an ordinary person, the scan sounds like it can't hurt and can only help. I have tried to point out in the post why this is a fallacy. Thanks for your comment.

Josip Skoko said...

I LIke your site and good information
positive and negative predictive values of body scan,Thanks
Full Body CT scan

Anonymous said...

as erroneous as you believe these scans to be, if one could have saved my husband from sudden death at 42 of an undiagnosed abdominal aortic aneurysm, i would have been on my knees thanking God for the ability to buy one myself and not have to wait for a Dr. to order one. Of all the stupid things we people waste our money on, a whole body CT scan is not one of them. If it saves a life once in a blue moon its worth it. Quit scaring people away from something that could be a good thing, doc!

Anonymous said...

The commenter above, perhaps inadvertently, reveals the fallacy of the argument in the article: A sample is not a population. It's a common error to make when trying to employ statistics in everyday problems. If you'd like more background, sign up for my class next semester!

Equity Options Underdog said...

I love how MDs refer to others as "ordinary people". While it is true that wearing your seatbelt every time you drive is unlikely to save your life on any given day, we still advise people to always wear them. Why? because the cost to do this is near zero (one second of our time per click). In our lifetimes we may only have a single accident where the seat belt actually saves our lives but we collectively think that all the thousands of seconds we spend needless clicking our seat belts is worth it. The only thing distinguishing this from MRI scanning for early diagnosis of a once-in-a-lifetime life threatening medical condition is the cost to the health care system. If not for the cost these holier-than-thou MDs would not have such heartburn with routine MRI scanning. Indeed, using population statistics to argue against routine MRI in order to save costs is about as convincing as telling me I should not wear a seatbelt. I care about my "n of 1" event because I am a sample not a distribution. And for that I will get scanned.

Add this