• Nonbeneficial treatments and moral asymmetry

    This is a follow-up to my post earlier today. Consider two vignettes:

    1. A patient requests a treatment or diagnostic test from a physician. The physician reasonably believes that the service is of no medical benefit to the patient (perhaps based evidence so strong the vast majority of physicians would concur). At first the physician is reluctant to perform the service, but the patient so desperately wants the treatment/diagnostic test — even paying full price out-of-pocket for it — that the physician yields and delivers it anyway.
    2. A physician offers a treatment or test to a patient that is widely known to be of no medical benefit. (Perhaps this physician doesn’t know, but let’s presume it is not controversial that it is not beneficial.) All things being equal, the patient would rather not undergo the treatment or test, but the doctor is insistent that it is necessary. The patient yields and the service is delivered. (If it matters to you, and to make this case parallel to vignette 1, let’s suppose the patient is paying full price out-of-pocket.)

    In both vignettes a transaction occurred that one party was somewhat reluctant to participate in and the other party wanted very badly. In both cases the treatment is assumed to be medically nonbeneficial. In both cases the patient pays out-of-pocket; no insurance is involved. The difference is that in vignette 1, the physician yielded to the patient’s values and in vignette 2, the patient yielded to the physician. Yet the outcome is the same.

    Questions: Do we feel differently about these two cases? Is delivery of the service moral in one case and not another? Would it have been OK for the physician in vignette 1 to refuse provision of (nonbeneficial) care? Would it have been sensible for the patient in vignette 2 to decline it? Would your answer change if the possibility of obtaining second opinions in either case was removed (e.g., an emergency situation or one with limitations on access to alternative care)? Are these vignettes morally symmetric or not?

    Which vignette is “patient-centered” care? Which is “consumer-directed”? Which is (or potentially is) “rationing”? What if some IPAB-like board had declared the treatment as nonbeneficial?* What if the doctor had some “skin in the game” (e.g. operating under an ACO or some capitated arrangement)? What if the patient did not have “skin in the game” (fully insured)?

    It’s not hard to draw a line from policy we debate to vignettes like these. Is our reaction to such vignettes consistent with our policy preferences?

    * Note that the IPAB cannot do this under current law.


    • I think that there is a pretty strong consensus around two basic ideas: first, we go to physicians because we believe that the physician is a trained expert who can advise us how best to resolve our problems, and second, in a free society we have the right to throw away our money on any ridiculous thing we want, as long as it doesn’t hurt anyone else.

      In the first case, the physician has fulfilled her duty by explaining to the patient that the treatment is generally regarded as ineffective. Since the patient is paying entirely out-of-pocket, he isn’t harming anyone else by going ahead and purchasing the treatment (I’m assuming that the resources for the treatment aren’t scarce, so providing the treatment to the patient doesn’t deprive a worthy recipient.) So, no problems here, at least in terms of patient / physician interactions.

      The second case is much more difficult. Why is the physician recommending the treatment if it doesn’t work? Is it because the doctor gets some sort of financial reward for the treatment i.e. she is a partner in a group practice that just purchased an MRI machine, and so she benefits every time she can direct patients to the device? If that isn’t the case, then why is the doctor recommending the ineffective treatment? If it truly is widely recognized as ineffective, then the doctor is ignorant or incompetent, violating the idea that the physician is a highly trained expert. So in both scenarios, there seems to be a serious ethical problem on the part of the physician — in both cases the patient’s trust is being abused, in the first case intentionally for financial gain, in the second case by incompetence.

      The difference between the two scenarios that you have posited is that there is an asymmetry in the relationship between physician and patient, and that changes how the purchase of ineffective treatment is viewed. And, as always, we get to spend our own money on whatever foolishness we please.

    • Here’s a way to add a little more complexity to this problem: the test that the patient is requesting is likely to lead to a “therapeutic cascade,” more testing, perhaps even treatment down the line. (Think PSA test.) Much of that downstream testing and treatment has serious side effects. (think prostate biopsy, and prostatectomy) including death.

      Now what is the ethical thing for the physician to do?

    • We need to add in risk to the equation. While there is little risk (ordinarily) from the MRI, there is always the chance of a false positive, a study which is over read. This could lead to repeat studies and perhaps not needed surgery. While I think I mostly agree with Brad’s quote from Berwick down below, I feel no need to make things easy.

      My partner and I provide anesthesia for the children who need MRIs at our facility, along with adults as requested. We see what we think are a fair number of inappropriate referrals. We often redirect the patients to the radiologists and have them speak with the PCP directly, so they get a more appropriate study, or none. When we find out a study has been done elsewhere relatively recently, we confirm that they really want another study.

      Our service has grown a lot. We get way too many adults who really dont need our care. It is too easy for the surgeon or primary to just check off the anesthesia box. We have a very nice nurse working with us. When we think it is pretty clear the patient is not that anxious, and it is not a head first case, we have her sit inside and hold the patient’s hand. Works nicely most of the time, safer and cheaper. Kind of like passing up free money on our part, but seems the right thing to do.


    • Austin
      I like where you are going on the posts today. There are real benefits in having these vignettes. I would even make a suggestion. Crowd sourcing and insertion of a polling applet on questions of relevance would serve the readership well. I would like to see how the community answers a refined, focused question.

      (The above entry asks a lot of questions, too many to answer)

      Just something to think about:

      1) Are there any diagnostic tests that are dead ends, ie, that wont potentially lead to more testing or have a NPV/PPV of 100%. Is there potential harm? If yes, the ethical quandary is heightened.

      2) Is giving in to patient demands detrimental, knowing the possibility of potent placebo effect. Is there therapeutic value in this? This is very real btw, the lit is bldg quite rapidly.

      3) By testing a patient, do you prevent them from showing up in another setting, utilizing more resources than intended (there are citations on this)

      4) Is testing one patient, with concurrent psychosocial stressors/anxiety likely to produce a difft result than in an mentally healthy patient. See #3.

      I could go on.

      This is ethics, economics, and clinical dilemmas at the x-roads. This is fodder for CER type dilemmas, and is illustrative of hard this will be–and see my short list as evidence. Lots of exceptions.


      • Dear Brad —

        I’m puzzled by your first question on diagnostic “dead ends” which don’t lead to further testing, in particular, your example of a test with an NPV or a PPV of 100%. How could this increase the ethical quandary? Suppose a test has an NPV of 100%: that means that a negative test result guarantees that the patient doesn’t have the disease, and so we could rule out further testing with complete confidence. Similarly, a PPV of 100% would imply that a positive result is incontrovertible proof that a patient has the disease i.e. a guarantee that this is not a false positive. How could such perfect information be anything but beneficial to the patient?

        It seems to me that the real ethical quandary is when a test has weak predictive power, for instance an NPV or a PPV of 50%, because in that case the test results are completely non-predictive, so the patient does not in any way benefit from the procedure — we could diagnose just as well by flipping a fair coin.

        • Precisely. I believe its your interpretation of my statement, mainly if operating characteristics of diagnostic test are anything less than perfect, the result is scenario you describe. We agree.

    • I’ve experienced a variant of vignette 2. As a college student, I had a rash on my hands that might have been, in the nurse’s opinion, second stage syphilis. I assured her it was not, but she refused to refer me without testing, which wasted my time (took 3 days to get the test, get the results, and get an appointment with a doctor to look at my rash) and the insurance company’s money.

      Ultimately, there is a tension between the clinical judgement of medical professionals and patient directed services. As long as someone else is paying, patients don’t object strongly to unnecessary interventions.

    • Seems easy to me, in that the second scenario involves either deceit or professional incompetence on the doctor’s part, and the first scenario doesn’t.

      In each case the physician’s only ethical obligation is to provide the patient with a professional opinion which is (1) medically sound and (2) uninfluenced by personal interests that conflict with the patient’s best interest.

      The patient is free to follow the doctor’s advice or reject it. The physician is free to acquiesce to the patients’ demands or refuse to accommodate them. So long as the physician has satisfied conditions 1 and 2 he has fulfilled his professional and moral obligations. End of story.