Medicine Is Not An Exact Science (Countering A Defensive Medicine Myth)

As I’ve argued on this blog before, “defensive medicine” — i.e., the claim that doctors routinely order useless medical tests and procedures solely to prevent the appearance of malpractice — is a myth. By and large, if a test or procedure won’t help prevent the patient from developing a serious complication, then it won’t help the doctor avoid a lawsuit if something goes wrong. Conversely, if a test or procedure would have helped the patient avoid a serious injury or complication, then, typically, the standard of medical care required the test or procedure anyway. Tort liability arises only when a doctor fails to do what a reasonable doctor would do and that failure causes the patient harm.

As I detailed again back in August, states that enacted draconian “tort reform” laws — which made it impossible for injured patients to pursue all but the most egregious cases of malpractice involving millions of dollars in damages — have not seen any reductions in their health care costs. Tort law and malpractice lawsuits simply don’t have much of an effect on health care and health care spending in the United States. That’s not surprising, considering that malpractice payments comprise a mere one-tenth of 1 percent (0.11 percent) of national health care costs. Plenty of unnecessary medical procedures are performed every day, but the motive for them is the obvious one: to make money.

However, the myth of defensive medicine just won’t go away, and on Christmas Day, a doctor under the pseudonym “Skeptical Scalpel, MD” posted that “defensive medicine is ubiquitous and not going to go away soon. Health care costs will continue to rise.” He recommends “a massive culture shift” among doctors and patients.

It’s unclear what kind of culture shift Dr. Scalpel has in mind, but the three examples of “defensive medicine” he gives suggest he wants doctors to scrap the differential diagnosis (the process of elimination that forms the backbone of clinical medicine) and stop reaching out to colleagues for advice on complicated cases.

Here’s his first example:

A young man with chest pain arrives in the ED. After taking a history and examining the patient, the ED MD is 99.99% certain that the patient did not have a heart attack or a pulmonary embolism. But he’s a little short of breath. He remembers a case of a fatal PE with only minimal shortness of breath, orders a blood gas and CT angiogram of the chest.

Any physician who is “99.99% certain” that patient did not suffer a heart attack is unfit to practice medicine. A patient history and physical examination will never give any physician enough information to be “99.99%” or 98.78% or 73.64% or 22.13% or 00.01% certain of anything. Clinical medicine doesn’t work that way. A doctor in that circumstance might reasonably be “pretty sure” or “confident” that the patient isn’t having a serious, urgent medical condition, but to ascribe any specific probability to it — much less a probability (“99.99%”) that is an incredible 3.89 standard deviations from the mean, with an equally surprising zero confidence interval — means that the doctor has utterly failed to understand the limits of a clinical medicine, and has wrongly presumed a level of mathematical precision to their analysis. 

I’m sure Dr. Scalpel meant “99.99%” as an informal shorthand for “pretty sure.” But there’s a reason he did it: because replacing his “99.99%” with “pretty sure” makes the example far less persuasive. It sounds a lot better to claim some doctor’s gut-level intuition is actually a medical fact asserted with “99.99% certainty,” but it’s dead wrong, and it reflects a serious problem with the doctor’s mindset.

Truth is, the doctor in that example has committed malpractice because they have not applied the differential diagnosis, which requires ruling out the most serious conditions. Doctors in clinical settings aren’t supposed to vaguely speculate about probabilities, they’re supposed to start with the worst that could happen and then work their way down. If a doctor can, without harming the patient, obtain more information that can help them rule out a serious condition, then they must do so.

In this case, it’s a young man (presumably a fit one, without any significant comorbidities) who had chest pain significant enough to send them to the ER, and with a little shortness of breath. He could indeed be suffering a heart attack. So what to do? Crack open their chest and go for a transplant?

Of course not. I don’t know why Dr. Scalpel says the next step is to “order[ ] a blood gas and CT angiogram of the chest.” From a malpractice lawyer’s perspective, what I’d look for as the next step would most likely be an electrocardiogram (as the Mayo Clinic says, “This is the first test done to diagnose a heart attack.”). Then normal blood labs would likely be drawn and reviewed, potentially including an arterial blood gas (though unlikely, given the minimal shortness of breath; the patient might instead get a pulseox). If those showed problems as well, an echocardiogram would be done, and, lastly, and typically only after the other tests have shown anomalies, would an angiogram be done.

And what would be so wrong about that? Is is such a terrible burden on the health care system to run an electrocardiogram and normal blood labs on a person with substantial enough chest pain and shortness of breath that they brought themselves to an ER? The cost of each is trivial, and the risks are non-existent. Dr. Scalpel is advocating for the equivalent of a mechanic saying, “I don’t need to look” when your brake light comes on because the mechanic is “pretty sure” it’s just the sensor. Odds are, they’re right — but you don’t go to a mechanic or a doctor to play Russian roulette (where, odds are, you’ll be fine).  “Pretty sure” just won’t cut it in matters of life and death.

Dr. Scalpel’s other two examples are even worse: a young girl with lower abdominal pain gets an ultrasound for appendicitis (among the least invasive, least expensive, and most helpful tests in history — remember this funny GE ad for their portable ultrasound?), and a surgeon calls up an infectious disease colleague to confirm he or she can drain an E. Coli-infected post-operative wound by draining it, without antibiotics.

Is that the big boogey-man of defensive medicine? Two patients who might really have life-threatening conditions receive cheap, non-invasive tests, and a third patient with a dangerous complication gets a quick informal consultation from an expert? If we’re really at that level of penny-pinching in our healthcare system, then there are plenty of other places to look for waste — like in the trillion dollars in damage caused every year by malpractice.

[UPDATE: Skeptical Scalpel has posted a reply, with the rather curious argument that all of the testing he believes is the standard of care is actually worthless, and with a couple more assumed tests lumped on the examples.

On the first example, he asserts that a CT angiogram is performed every time a patient has chest pain, which, as I explained, simply isn’t correct: a whole battery of non-invasive, inexpensive tests are performed first, with the CT angiogram following worrisome results on those tests.

On the second example, he now argues the ultrasound is worthless in ruling out appendicitis. If that’s correct, then the ultrasound makes no difference one way or another in terms of the doctor’s liability; however, the very source he cites for that claim concludes “Bedside ultrasound is helpful in patients with suspected appendicitis to confirm the diagnosis.” Recognizing that, he then lumps on the assumption that the patient will also be given a CT, for unexplained reasons.

On the third example, he notes (correctly) that surgeons have far more experience in treating post-operative wounds than infectious disease specialists — one of the reasons why his example didn’t make sense in the first place — and then notes (again, correctly) that the infectious disease specialist would likely refuse to sign off on the surgeon’s routine patient without actually accepting the patient as their own.

It’s important to bear in mind what we’re talking about here. I think doctors by and large try to do right by their patients, ordering the tests they think will rule out serious conditions. Skeptical Scalpel, however, thinks that a significant portion of the medicine practiced in this country is useless, and done solely to avoid liability, and the three best examples he can come up with are debatable (the ultrasound for suspected appendicitis), meaningless (the post-operative ID consult that usually won’t happen and which won’t change the surgeon’s liability anyway) or simply wrong (the young man with chest pain automatically receiving a CT angiogram).]

  • Alex R

    I disagree with you on all counts.

    In the first example, yes, an EKG and a blood gas would be the next logical step. But the doctor orders an unnecessary CT angiogram. In the second one, again, a CT is unnecessary. As for the third example, if the ID doc were to perform a “quick, informal consult”, he wouldn’t be acting too responsibly. His bill and the unnecessary antibiotics add to the bill. The antibiotics also contribute to bacterial resistance.

    Defensive medicine is a reality.

    • It seems you’re putting the cart before the horse, assuming the patient did not have the serious condition, which you then use as a post hoc justification for calling the test unnecessary.
      For the first, why do you assert the doctor must “order an unnecessary CT angiogram?” The vast majority of ER patients with chest pain and shortness of breath do not receive one, because the possibility of a heart attack is ruled out by the less invasive tests. That’s why they’re done at all — if the next step after a clinical exam was a CT angiogram, the rest of the tests wouldn’t be used at all.
      For the second, the issue is an ultrasound, not a CT. Why do you consider that unnecessary? Does it not give the doctor information they need to rule out a possible, serious condition?
      For the third, it seems you’ve missed part of the predicate. Surgeons treat post-operative infections all the time, and calling an ID specialist won’t do anything to help them from a liability standpoint if the surgeon breaches the standard of care. The sole issue will be: was it the standard of care for the surgeon to prescribe antibiotics? If the case was complicated enough to warrant calling an ID specialist, then the call indeed should be made.

      • Really, dude?

        What makes you think you have any qualification whatsoever to comment on appropriate medical treatment? Are you so arrogant as to think that you could possibly know even a fraction of what physicians know after years of training and experience? Or would a little humility get in the way of lining your pockets?

        • Is it your understanding that, in a malpractice lawsuit, a lawyer can “line their pockets” by standing up and proclaiming what they personally think should be appropriate medical treatment?

        • Lily P.

          “Alex” and “Really” don’t seem to realize that it isn’t the lawyer telling the judge
          and jury what the standard is… it’s OTHER DOCTORS who serve as

          In my opinion, a doctor is far
          less likely to get sued by listening to you (and other med mal legal
          bloggers) than by listening to Dr. Scapel’s dubious advice.

        • Indeed. A lawyer’s assertions about the medicine doesn’t mean much in a court of law — it’s all about the experts, who are themselves qualified physicians who based their opinions on the prevailing standards and practices in the profession. “Alex” and “Really” are arguing against the medical profession itself.

  • Lily P.

    Ha! It looks like you’ve annoyed doctors who are advocating substandard care. If they’re among the small percentage of doctors who get sued repeatedly for sucking, then it sounds like they deserve it.

    • If you’re not very good at your job — or you just don’t want people ever looking over your shoulder — then your best bet is to claim that your job is so complex that mere mortals can’t even review it. Thus, the response from medical professionals to critiques of defensive medicine and the like is always the same: when it comes to diagnosing and treating conditions, medicine is a mathematical science free from error, but when it comes to proving malpractice, medicine is a dark art incapable of description or interpretation.

      The sad truth is that many doctors just plain don’t want to do the differential diagnosis, they want to go with their gut feeling. That’s not good medicine, and they are rightly held responsible for the consequences.

  • You do your argument no favor by lumping new assumptions on top of assumptions that were already dubious.

    The numbers in your chest pain example are made up out of thin air. More to the point, your example is different from Scalpel’s: if, indeed, a young man with chest pain has shown abnormal EKG and lab results, then he most certainly should have an angiogram! Or are you saying that young men never, ever, ever have heart attacks?

    Similarly, your ultrasound example has totally missed the boat — the example Scalpel gave doesn’t involve a biopsy, it involves a suspected appendicitis. Your argument seems to be a general assertion that ultrasounds should never be used because they might find something suspicious… an argument that, thankfully, carries no weight in the medical community. Come back with a more specific example.

    Your ID example is even worse — you put another cart in front of the cart in front of the horse by assuming they’ve all been sued for unspecified reasons. Focus on the facts presented: whether a surgeon should prescribe antibiotics for an infected post-operative wound. That’s the question at hand, and if the surgeon doesn’t know the answer, shouldn’t they reach out for help?

    Your analysis of the “malpractice system” as a “joke” is largely correct: the vast majority of negligently injured patients receive no compensation at all. Indeed, most of them have the truth withheld from them, and never know the cause of their injuries. Implementing a much broader compensation system — something I’ve supported in the past — would likely be more fair, but it would also cost far more than our current system, because it would compensate far more people. Are you willing to pay that price? Most people like you arguing against malpractice liability refuse to pay that price, they’re just raising a red herring for the purpose of avoiding responsibility altogether.

    • VKA

      I suppose I should know better than to argue with a lawyer over the internet. This will be my last post, and you can get the last word if you desire. I will seek to explain the errors in your reasoning as best that I can.

      My EKG example was the same as Scalpel. It is you who does not understand medical practice- a non-invasive and ‘free’ test such as an EKG isn’t usually explicitly mentioned in medical conversation. Practically anyone in an ER gets hooked up to a monitor (which checks EKG, O2 sat, and BP) regardless of why they are there- meaning it’s already available to the physician as they are taking a history. Thus, the patient in the example is almost certainly fine (how much is “almost certainly”? The above study in the UK of patients 13-35 does not mention a single pulmonary embolism in 400 patients. Believe me- a pulmonary embolism would have been mentioned if found in a young male with no risk factors) The ordering of the angiogram despite the lack of any non-invasive tests is defensive medicine and a waste. It is perceived to reduce liability even when it does no such thing.

      Ultrasound as well: I was not arguing that no one should get an ultrasound, but that it’s use must be tempered and used in appropriate circumstances. Using it in patients who are not good candidates based on clinical history results in false positives that cause more harms. Are you familiar with the concepts of number needed to treat? Attributable risk reduction? Or Beysian probability analysis? Basically, if an ultrasound would not reduce risk of a missed appendicitis by enough to offset the harms of incidental findings, then it should not be done. How much risk it reduces is dependent upon the clinical scenario. For example, if the patient is complaining of right lower quadrant pain, but jumps off the table when the surgeon is finished- no ultrasound should be done and the patient should be sent home (the jump without pain indicates the child is most likely faking the symptoms- appendicitis would leave the child in agony after any such jump).

      Finally, with the ID scenario you are attacking a straw man, and trying to change the facts presented as well as the question. We are not talking about incompetent surgeons (which I hope you agree are a small minority). We are talking about competent surgeons who know the correct antibiotics (as indeed the vast majority of surgeons do). They do not require any aid from another specialty- but they will request it anyway, solely because in the event that the infection spreads (always a distinct possibility regardless of the surgeon’s actions) they want to be in a legally stronger position. This is defensive medicine.

      I think what you fundamentally misunderstand is that in each of the above cases, even if the doctor does “everything right”, a bad outcome can still occur. That is the randomness and complexity of the human body. Doctors practice defensive medicine because in the event of those random occurrences- bad outcomes despite the best treatment- they want to be be in a position to win any lawsuit.

      Finally, you state that a system like the one I proposed will vastly escalate costs. However, this has not been the experience in the vaccine injury program. There, set reimbursement rates based on provable injuries has cut costs dramatically- where before the implementation of the program the court costs threatened to bring down the vaccine industry. Similarly, New Zealand has also kept costs to a remarkably low level with a medicine-wide compensation system. There is no evidence that a broader compensation system based solely on negative outcomes- not upon blame or “malpractice”- will increase costs.

      • In each, you assume your own conclusions.

        In the first, you’ve now redefined the example to involve a young man with mild chest pain, mild shortness of breath, who has been had totally normal findings on EKG, SpO2, blood labs, and echocardiogram. Such a patient will not get a CT angiogram for liability reasons. Period. It is not the standard of care to do so. If you’ve seen this before in your practice, you’ve seen it for one reason alone: to pad the doctor’s and hospital’s bills. Truth is, you’ve probably never seen this happen either, it’s an invented hypothetical that stretches the bounds of imagination. The far more common scenario in medicine is for a patient with chest pain and shortness of breath to be sent home without an EKG or blood labs — after which they expire as the result of a preventable heart attack.

        In the second, you have again left specific examples to instead generally complain about ultrasounds. Stick with the example Scalpel gave: under a risk reduction analysis, what, exactly, is the potential harm of the ultrasound? We’re not talking about, say, men receiving a PSA analysis for their prostate, where the benefit of knowing is really clear in all circumstances. We’re talking about a child with lower abdominal pain. What unnecessary procedures could result from an ultrasound? Be specific.

        In the third, you’ve changed your example again, and still missed the point: the ID consult (which is the sole thing Scalpel complained about) does nothing to change the doctor’s liability, because the ID answer doesn’t change what the standard of care is for the doctor. The doctor would know that. If the doctor was truly being defensive as you allege, they wouldn’t even bother with the ID, they’d just prescribe antibiotics at reckless abandon. But that wouldn’t do anything either: as Scalpel claims, the standard of care typically does not call for antibiotics.

        I completely understand that bad outcomes occur all the time despite perfect care; I spend a lot of my time explaining that to seriously injured people who call our office asking us to review cases that involved a really bad outcome but no malpractice. You “fundamentally misunderstand” that a bad outcome does not equal a successful malpractice lawsuit. A successful malpractice lawsuit requires competent medical testimony establishing a breach in the standard of care that caused an injury to the patient — I spend a lot of my time explaining that, too, to seriously injured people who call our office asking us to review cases that involved malpractice and a really bad outcome, but an outcome unconnected to the malpractice.

        Again, the problem here is your definition of “defensive” medicine. Malpractice law is not a random win/loss game. The malpractice system is unduly hostile to plaintiffs — that’s what researchers at Harvard School of Medicine found in 2006 — but, by and large, it only awards compensation when the plaintiff has presented compelling medical testimony establishing a breach in the standard of care that caused an injury to the patient. If a “defensive” test wouldn’t benefit the patient, it wouldn’t help avoid liability, either.

        The compensation program is a different discussion, but I noticed that you simply avoided the answer. It seems you’d be quite happy with that system only if it cost less, and you’d oppose it if it turned out to cost more. Your concern isn’t fairness, it’s just lowering your costs.

        • RuggerMD

          You said:
          “Doctors in clinical settings aren’t supposed to vaguely speculate about
          probabilities, they’re supposed to start with the worst that could
          happen and then work their way down. If a doctor can, without harming
          the patient, obtain more information that can help them rule out a
          serious condition, then they must do so.”
          Actually doctors are supposed to take a history and physical exam and then make a decision as to what the diagnosis is. If you need help in doing that, you order labs, tests, and radiological tests to help you make a definitive diagnosis.
          Most of the time we can easily come to a definitive diagnosis. When we can not we make a differential diagnosis and then start the defensive medicine cause of lawyers like yourself who are only out to make a buck.
          Oh and by the way, doctors DO NOT make extra money by ordering extra tests in a hospital, just so you all know.


        • You write “actually,” as if you’re going to disagree with me, and then you agree with me that the differential diagnosis is the standard of care.
          If “most of the time” you shoot from the hip that just means most of the time you don’t bother to practice medicine. Then you have the nerve to complain about it when you actually do have to practice medicine — as if you’re allowed to give up and go with your gut when it would take hard work to figure out what the patient has and needs.
          Tort liability makes doctors like you better at your job.

        • ML

          When you say that “doctors DO NOT make extra money by ordering extra tests,” is that a universal fact? The US Attorney’s Office seems to disagree with you:

          As for less egregious examples, don’t doctors benefit when they make more money for the hospital by ordering unnecessary, expensive tests? Are you saying that doctors don’t benefit from making their employers happy? They know who butters their bread, right?

        • RuggerMD

          “most of the time” means most of the time the differential diagnosis we have made in our head is very clear as to which one is the diagnosis. And no, we don’t have to document that just for you. Unfortunately for you, you think that practicing medicine means ordering a bunch of random tests, but it really means diagnosing your patient and treating them properly.
          And just so you know, I don’t practice defensive medicine. I talk to the patient about what their diagnosis may be and I give them options for the best treatments, then I allow them input as to what treatments they want. When I am uncertain as to the diagnosis, I tell the patient what the differential diagnosis are, and I tell them the likelihood of each, and how we can eliminate or confirm diagnosis. Then I discuss with the patient as to which test we should perform, depending on likelihood of sensitivities/specificity of those tests versus harm or cost to that patient. Then I allow the patient again to make their informed decision as to how best to come to a conclusion.
          That is what most physicians do.

          I hope you have as much disdain for physicians as you demonstrate on you blog when you or your family get ill. You will surely see a lot of differential diagnosing and defensive medicine.

          ML, your first example is the oncologist billing for giving chemo, not ordering a bunch of tests.
          No, most doctors do not benefit when they order expensive tests. There should not be any kickbacks.

        • As I understand it, your argument is: “doctors must do the differential diagnosis, and of course I do that, but it’s okay if they don’t, and if they don’t then too bad for you.” Again, if you’re actually doing a differential diagnosis, then neither documentation nor defensiveness makes any difference: you will do what the standard of care requires, and either prevent the harm or, if a harm happens, will not be liable. You seem to want a world in which it doesn’t matter if the doctor actually does their job or not.
          As for doctors not benefitting from excessive testing, you’re nuts. Most doctors in hospitals are scored on billing and billing efficiency, and it goes right into their pay. Most doctors with their own practices charge a markup for everything. The more care, the more profit. Atul Gawande has written about how, in Texas, they passed severe tort reform, then watched CT rates skyrocket — because a whole bunch of practices bought CT machines and started billing for them directly, at a premium. Go figure.

  • Really, dude


    Do you find it ironic that you are arguing with expert physicians about how to practice medicine while also maintaining that a lawyer’s assertions about how to practice medicine are irrelevant?

    • I find it ironic that, despite all these clues, you have no idea where the standards of care for doctors comes from: it comes from other doctors. The standard of care in malpractice cases is proven in court exclusively by way of other doctors. You don’t seem to get that.
      I happen to know the standards applicable to a patient with chest pain in the ER because it’s sadly a common missed case, a case where people die every year because some lazy, arrogant doctor is “pretty sure” the patient isn’t having a heart attack and so doesn’t bother to do basic labs and an EKG. I’ve seen many, many undiagnosed heart attack and PE (and stroke) cases, and have discussed all of them with many Emergency Medicine physicians, some of them medical school professors, all of them deeply familiar with standards like the ACEP guidelines.
      Unsurprisingly, you, Scalpel, and everyone else has had to run away from the original examples given, because they didn’t make any sense: in no hospital in America would a young man with mild chest pain be given a CT angiogram on the basis of nothing more than the complaint of chest pain. Even Scalpel knows that, despite not being an Emergency Medicine Physician. His example is thus reduced now to an absurdity: a CT angiogram for a young man with a normal EKG, normal blood labs, and a normal echocardiogram. It simply doesn’t happen that way, and neither Scalpel’s pseudonymous assertions nor your anonymous assertions will make it so.

    • ML

      How do I know you’re an “expert physician”? Would you mind displaying your expert qualifications? I’m more inclined to believe the med mal attorney whose credentials are available on this website versus the anonymous trolls who seem to be taking this issue a bit too personally (which makes me wonder about whether they are–or will be–regular defendants in these sorts of cases).