Withholding information from patients: it’s not OK

Over at his blog Gene Expression, Razib Khan expressed his disgust with the paternalism of the medical profession. His disgust was not misplaced, but he did bypass some of the subtleties.

Razib, a respected science blogger, ran smack into a frightening piece in Time. In it, doctors discuss whether and how much genetic information should be given to parents about their kids. The doctors featured evaluated a child’s genome, found a potential for early-onset dementia, and chose not to tell the parents. Let’s break this down.

Paternalism vs. Autonomy

Paternalism in medicine isn’t prima facie bad. We count on professionals to use their superior knowledge, experience, and relative objectivity to help us. The premise is, “you know more than I do; help me.” The foundation of medical practice rests on centuries of paternalism taken to every imaginable extreme: surgeries performed without consent, diagnoses being withheld, forced sterilization. Over the last fifty years the reliance on paternalism has given way to our value of patient autonomy. This has been an uneasy struggle, one that lives in a grey world of few right answers.

But there are many things we as a society and medicine as a profession do agree upon. The basis of patient autonomy is informed consent. Patients should be given every opportunity to understand what we propose to do for them and be given a chance to assent or decline any test or treatment.

Once a patient has given consent, information must continue to flow. I cannot ethically order a test or perform a procedure and then withhold the outcome from the patient.

Since there are a lot of grey areas, it’s helpful to see if there are any bright lines. A patient who comes to see me signs a general consent to treatment which covers much of what we do. I don’t need to explicitly ask permission to look into an ear or to analyze kidney function. These are presumed to fall under a general consent because they are relatively minor interventions that are a normal part of an exam and commonly without grave implications. Contrast this to a CT scan of the abdomen. This is a more “extraordinary” intervention, one with potential risks (radiation exposure, exposure to potentially toxic intravenous dye) and one whose results may give life-altering information (a tumor, a need for surgery, etc.).

Let’s take these roughly-sketched guidelines—specific consent should be obtained for extraordinary procedures with unusual risks or high probability of life-altering results—and take a look at the Time article.

The piece is not terribly written, but it leaves out a lot of important information. That being said, we can assume that a genomic analysis was conducted of some children, and doctors puzzled over whether to give the results to the parents.

Ethics require that the parents (who are surrogate decision makers for the children) be given informed consent. The genomic analysis should not be done at all without informing the parents what we can or cannot learn from the test. Genetic predisposition to disease can be very mushy: it’s easy to see if someone has sickle cell anemia, but not so easy to see if they will develop hypertension.

For the purposes of this discussion, I’ll assume that whatever this “dementia gene” the doctors found gives the child a very strong chance of dementia at an early age. In this case, they cannot ethically withhold this information from the parents. What if there were a “maybe, sort of” predisposition? Same answer.

In fact, if the parents gave their informed consent, and the information included the uncertainties inherent in genomic analysis, they have the right to all the results, along with interpretations from a professional who understands them. If they were not given proper informed consent, the test shouldn’t have been run. A doctor who orders a test must have it’s consequences mapped out in advance, and shouldn’t be wondering after the fact what to do with the data.

This scenario assumes an ideal collaboration between doctor and patient, where communication flows freely, where a doctor uses their expertise to guide patients’ decisions. The complex nature of medical knowledge and its potential for life-altering impact (and expense) argue against any patient being able to order any test for themselves, including a genomic analysis. The information gained from such an analysis requires professional interpretation so that people can use the information wisely. A test result that sounds horrible to a patient may in fact be insignificant, and vice versa.

There are risks to people having unbounded ability to use medical testing and treatment. For patients to agree to this important premise, we as physicians must respect their autonomy and human dignity.


  1. I have a couple of questions
    How can you examine someone’s kidney by looking into their ear? I’ve got to say, you’ve really aroused my curiosity, there.’
    And also, is there a potential to treat a child’s dementia, if it’s discovered before it happens (like with this genome test)? Can you even treat dementia? That would make such a test a good idea, right?

  2. The ear and the kidney were separate examples : )
    The potential for intervention is not relevant to whether the parents are given the information. This sort of thing is to be discussed before ordering the test.

  3. twitter correspondent reports they messed up consent. no genetic counselor involved. when she puts up a blog post, i will update. but the issue seems more to be that they messed up and it wasn’t highlighted (or not divulged explicitly) in the piece.

    • im curious to see. it’s critical. gotta be careful doing complicated tests without the right person to interpret.

  4. SurgPA

     /  November 1, 2012

    While I agree in principle with your thesis (and don’t withhold results from patients myself), informed consent implies that the patient is informed in a manner he or she can understand (appropriate language, education level, etc.) How can you obtain informed consent from a patient who is not equipped (intellectually, educationally) to understand the complexity of the issue you discuss? Given the increasing complexity of medical decision-making, how can you be sure your GED-educated patient with no understanding of statistics and limited understanding of basic science or human biology really understands that treating his hypertension is done, not with the end-goal of lowering the blood pressure, but ultimately to reduce by X% (absolute, or relative) their life-time risk of renal failure/stroke/MI, and that NNT > 1 implies that you may subject him to chronic meds without preventing a life-altering event? Even with a simple operative risk discussion, I am continually amazed that, when told that the risk during cholecystectomy of bile duct injury/leak is ~ 1:200, patients view that as a near-guarantee that it won’t happen, and fail to realize that it actually occurs in 1-2 of the 300 cases we do each year.

    I am not arguing for paternalism, but I fear that our current bias toward patient autonomy is frequently based on a faulty premise that patients’ superficial understanding equips them to make the best decision.

  5. im totally with you on this. some paternalism is unavoidable in a system that relies on professionalism. we have to communicate with each patient in a way they understand, we have to ask are patients about their values, and we have to help them make decisions that are in line with their medical needs and their values.

  6. theshortearedowl

     /  November 2, 2012

    The idea of informed consent gets pretty complicated with genetic tests. A given study might find a 10% increased relative risk of dementia in a particular population; if the patient doesn’t meet all of the criteria of that population and there are no other studies, what does that mean? Do you tell them? If so, can you be sure they understand the limitations of the science? Besides, if there is little or nothing they can do about it, is there even a point to telling them?

    With genomic-scale data, you are implicitly “testing” for thousands of bits of genetic information, with wildly varying degrees of risk and evidence for that risk. How do you explain all of them?

    Furthermore, with any genetic data, you are not only testing the patient, but also their parents, siblings, children and other close relatives. The implications for patient confidentiality are not trivial.

  7. JustaTech

     /  November 2, 2012

    Informed concent becomes even harder with genetic screening like 23andMe. My in-laws did it (knowing that they both have horrible health histories) and then wanted me (not a geneticist, but a biologist) to explain all of it to them.

    When we went over their results 23andMe had very long pages explaining the potential implications of screening for some specific diseases (BRCA 1 and 2, Parkinsons’, dementia, etc), but my mother-in-law didn’t want to bother with all that, so she just clicked through. Thankfully, nothing came up, but I really wish that there had been some way to insist that she actually read and understand what the screening meant.

  8. A. Marina Fournier

     /  November 5, 2012

    Theshortearedowl wrote:
    Besides, if there is little or nothing they can do about it, is there even a point to telling them?

    I’m left-handed, short, female, obese (working on that!) bipolar, insomniac, diabetic, a migraineur, hypertensive with high cholesterol. Every week, some study comes out saying that one of those conditions above are due to something I have no control over (genetically, usually), and that I will die earlier because of it.

    I just love those reports.

    Now, as to the paternalism of the medical profession, it can’t hold a candle to the sexist paternalism of the current GOP politicians, who are in several places, telling OB/GYNs that it’s okay to lie to their patients about their health, an abortion, birth control, the health of the developing fetus (and how it might affect her).

    It’s scarier and scarier out there. I thank fortune that I live in California.

%d bloggers like this: