Article Text

Download PDFPDF

Autonomy and negatively informed consent
  1. U Kihlbom
  1. U Kihlbom, Philosophy, Örebro University, 80172 Örebro, Sweden; Ulrik.Kihlbom{at}


The requirement of informed consent (IC) to medical treatments is almost invariably justified with appeal to patient autonomy. Indeed, it is common to assume that there is a conceptual link between the principle of respect for autonomy and the requirement of IC, as in the influential work of Beauchamp and Childress.

In this paper I will argue that the possible relation between the norm of respecting (or promoting) patient autonomy and IC is much weaker than conventionally conceived. One consequence of this is that it is possible to exercise your autonomy without having the amount of and the kind of information that are assumed in the standard requirement of IC to medical treatments. In particular, I will argue that with a plausible conception of patient autonomy, the respect for and the promotion of patient autonomy are in certain circumstances better protected by giving patients the right to give their negatively informed consent to medical treatments.

  • informed consent
  • autonomy
  • negative beliefs
  • positive beliefs
  • negatively informed consent

Statistics from

Though the practice of informed consent (IC) to medical treatments is widespread in western societies today and has very solid support within healthcare, it has critics claiming it is too weak. Robert Veatch for instance, claims that the practice of IC does not sufficiently safeguard the interests of the patient, and that it “will have to be replaced with a much more radical, robust notion of active patient participation”.1

IC is very often justified with an appeal to respect for personal autonomy, at least as a major ground. It is argued that patients ought to have a right to give their IC since this is the best way to respect and/or promote the autonomy of the patient.2

Contrary to this view, I will argue in this paper that the concern for personal autonomy is, in some circumstances, a reason for weakening the requirement of IC. By giving patients less information and of a different type than traditionally suggested, it is possible to secure patient autonomy with a more active role of the physician. If we care about patient autonomy, which we often should do, then we should relax the general practice of IC, and in some types of situations let the patient give a negatively IC rather than a positively IC, or so I will argue.

Note that my claim is not the idea that a patient may autonomously decide to bring their self-determination to an end. A patient may prefer that a trusted physician decide what he or she judges to be the best course of action. Though it seems perfectly possible and sometimes reasonable to make such a waiver, it would be a matter of giving up your autonomy, no matter how autonomous your decision is. In contrast, my claim here is that a relaxation of the requirement of IC would in some situations be a way to strengthen or at least to protect the autonomy of the patient.

I will concentrate on the relation between personal autonomy on the one hand and IC on the other. The notion of personal autonomy is notoriously blurry and is used in many different ways. Here I will take Beauchamp and Childress’ notion as a point of departure and make it somewhat more precise. But even if unambiguous, the individual or “atomic” conception of autonomy has been subject to much criticism. However, it will hopefully be clear that the present suggestion opens up for a relational dimension of personal autonomy since it requires substantive patient-doctor relationships of confidence or trust. Following this, I will assume that personal autonomy is a good thing most of the time, that should be both respected and promoted. I will not address the issue when considerations of autonomy may be irrelevant or should be overruled by other types of normative considerations.

Moreover, other considerations that presumably would serve as arguments in favour of my claim, such as the well being of the patient and health promotion, will be left out. There will be no discussion of the problems of how to actually accomplish a genuinely IC in the clinical setting. The empirical evidence of the many difficulties associated with IC is overwhelming.3 If significant, the problems of attainting such consents probably supports my view, but I will not pursue this line of argument here. Lastly, no tangible and specific suggestion of how to shape the practice of negatively IC at the clinical level will be delivered. My discussion will indicate some rough lines, but I leave these, admittedly difficult, details for now.


Why is it so commonly thought in western healthcare that the principle of autonomy supports the practice of IC? An important explanation is probably the interpretation of the autonomy idea that to exercise ones’ autonomy is a matter of being the direct and intentional cause to what happens to oneself, and, in turn, the presupposition that this can only be the case if you understand and are aware of what is happening to you. This idea is codified in Tom Beauchamp and John Childress’ influential book Principles in Biomedical Ethics in which IC is more or less understood as a way of practicing or manifesting personal autonomy.

Beauchamp and Childress “…analyse autonomous action in terms of normal choosers who act (1) intentionally, (2) with understanding, and (3) without controlling influences that determine the action” (p69).2

There are many objections to this notion. The only one I will point out here is that it is too thin. A patient fulfilling these conditions may make a decision that has no anchoring within the person, nor is anything said about the competence of practical reasoning. We might suggest the following notion of autonomy: I act autonomously if I, as the agent, am the power that freely and competently achieve my own ends by choosing what I have good grounds to believe to be, the best means to my ends. More explicitly we should say that the decision of a person (P) is autonomous when the following conditions are fulfilled:

  1. P’s decision is caused by P’s intention to make it

  2. P’s intentions are not controlled by anyone else

  3. P’s beliefs about the alternatives and their possible outcomes are well-founded

  4. P’s intention coheres with P’s other beliefs and desires

  5. P would revise one or several of his beliefs/desires if they would not cohere

The first condition is obviously a causal condition saying that P should cause the decision by requiring that it is P’s intentions to make the decision that causes the decision. The second condition supplements the first by requiring that no one else other than the agent control the intentions that cause the decision. The third is an externalistic condition about the agent’s beliefs of the decision and its effects, which requires that they be more or less justified. This includes having well-founded beliefs about probabilities of the different outcomes. The fourth condition is an authenticity condition of coherence that should capture that P’s decision is his own. The last one is a competence condition put on P’s practical reason. All these conditions are vague and can be interpreted in different ways, but that need not worry us here.

By invoking the two latter conditions, this conception is both stronger and more plausible than the notion of Beauchamp and Childress.

It is the third condition that corresponds to Beauchamp and Childress’ condition of understanding and they are the crucial ones here. Beauchamp and Childress say that understanding “only” required that this condition is satisfied to a substantial degree, “not full or even nearly full” (p69).2 What counts as a substantial understanding is contextually determined according to Beauchamp and Childress. Even though this account is not very clarifying, their illustrations of what the principle of autonomy applied to medical practices requires, indicate quite clearly that they identify understanding with comprehending information about voluntariness, diagnosis, procedures, risks, and prognoses. So IC seems to be the obvious conclusion when it comes to consent (Beauchamp & Childress, p100-5)2 The concept of IC they discuss is along the following lines: a patient has given an IC when the patient

  1. is competent, and

  2. has the capability of understanding the information

  3. has received information of:

    1. purpose of the treatment

    2. period of time

    3. methods and means

    4. all the significant difficulties and risks that are likely to occur

    5. that the treatment is voluntary

    6. that the consent can be withdrawn at any time, and

  4. on the basis of this information gives his/her voluntary and explicit consent to undergo the treatment.

I take it that to have understanding of a medical treatment in Beauchamp and Childress’ sense is a matter of having well-founded beliefs of the treatment and its significant outcomes. If so, Beauchamp and Childress’ claim seems to be that if a patient is to have understanding (relevant for autonomy) about a treatment he or she must have information of what the methods and means are of a certain medical treatment and of all the significant difficulties and risks that are likely to occur.


There might be many reasons why it is important that the patient receives information of means, methods and risks, but it is not obvious that personal autonomy is always among them.

The present suggestion of negatively IC implies that a patient can take an autonomous decision with less information about the actual methods, means and risks. A first step to spell out this suggestion is to make the distinction between positive and negative beliefs.

A positive belief is a belief the content of which says that this or that is the case (or will be, or has been the case). A negative belief, on the other hand, is a belief the content of which says this or that is not the case (or will not be, or has not been the case). For instance, when writing this I have a positive belief that there is snow on the ground outside my window and I have a negative belief that it is not snowing and another one that it is not raining. I also have the positive belief that it was snowing yesterday, and the negative belief that it will not snow within the next couple of hours.

As we have seen, one assumption behind Beauchamp and Childress’ traditional way of justifying IC is that for the patient to exercise her autonomy it is necessary that she has positive belief in the methods, means and risks concerned, and that they concern what the, for example, methods amount to.

This is mistaken. To exercise your autonomy, you do not need to know how your ends are realised, given that you have good grounds to believe that they will be realised. You might, instead, have a number of well-founded negative beliefs, beliefs about what will not happen to you. A combination of positive and negative beliefs where the beliefs about means and methods are negative rather than positive may well suffice for exercising autonomy. As an effect, the demands put on the patient’s ability to process information and deliberate becomes weaker and at least in some kind of cases, much more reasonable.

We might say that a patient has given a negatively informed consent when the patient:

  1. is competent, and

  2. has the capability of understanding the information

  3. has received information of:

    • 1 Purpose of the treatment

    • 2 That it is possible to receive more information if wanted

    • 2 That the treatment is voluntary

    • 3 That the consent can be withdrawn at any time

  4. has well founded beliefs that the physician will choose the treatment that best promote his/her values

  5. has well founded beliefs that the physician will choose the treatment, the risks of which are in accordance with his/her attitudes towards different kinds of risks.

  6. on the basis of this gives his/her voluntary and explicit consent to undergo the treatment and express his/her voluntary and explicit wish not to have more information.

The notion of negative IC differs from the notion of IC in the following respects: in IC, the patient has received information of purpose of the treatment, period of time, methods and means, of all the difficulties and risks that are likely to occur, and this is not the case in negative IC. On the other hand, in negative IC, includes condition (d)–(f), and this is not the case in IC.

To sustain this idea, I will discuss a number of different examples of different complexity, starting with the simplest one:

Suppose that I decide to send an email to a colleague and use my computer in the usual way to do this. It is obvious that I do not need to (positively) know how the means to my ends work in order to have well-founded beliefs about my decision in order for my decision to be autonomous. I know very little of how computers and the internet work, but that does not limit my autonomy.

It might be objected that this example is irrelevant for the discussion of autonomous decisions to undergo medical treatments since the information does not concern what happens to me, at least not directly.

To answer this worry, suppose that I have a severe headache and take a couple of painkillers to get rid of it. To have sufficient understanding for acting autonomously, I surely need to have good grounds for believing the pills will relieve me of my headache. It seems also reasonable that I also should have well-founded negative beliefs about that taking them will not bring with them significant risks for side-effects. However, I need no positive beliefs of how they chemically work in my brain, to have sufficient knowledge for making an autonomous decision. Contrary to the first example, this is a decision to undergo a medical treatment.

Now, it might be objected that this example is irrelevant in a further way to the discussion of IC to medical treatments. In the clinical setting, there are likely to be doctors and others involved. What threatens autonomy here is that they will take the decision concerning your treatment and they will know things about, for example, methods and risks, which you will be ignorant about. Your autonomy is lost or severely infringed by their decisions, or so it may be argued.

Again, to make an autonomous decision, I do not need to know, exactly how the means to my ends work, given that I have well founded beliefs that they will work—that they will be the best means to my ends, even if decisions and actions of others are among them.

To illustrate: suppose that you visit your brother who moved to Hong Kong five years ago. You head out for a meal together and at the restaurant your brother asks you what you would like to eat. For obvious reasons, you are ignorant about the food that the restaurant serves, and in contrast to your brother, you can neither read nor speak Cantonese. Would it be possible for you to make an autonomous decision about what course you would like to have without knowing what the different courses are and how they taste?

The answer seems to me to be positive. You can exercise your autonomy by saying that you want to eat something delicious, new and typically Chinese, and letting him decide where to go and what food you will have. But this only holds given that you have good reason to believe at least three things about your brother: that he knows your preferences about food, that he knows what the different dishes taste like, and that he will decide in accordance with your preferences. You do not need to have well-founded beliefs of what kind of food will be served.

What you need positive belief about is merely the result in terms of the realisation of your ends, not how the means, whatever they are, actually work. In order to make an autonomous decision you need also, obviously, negative beliefs about the meal and about the outcomes. That is, you need to have well-founded beliefs about what the meal will not involve or result in, that you will not be served something you dislike. Now, you have good reasons to assume that your brother will decide in a way that realise your ends better than any alternative you yourself can come up with. So, you are using him as the best means to achieve your ends as you might use any other means. This is not, then, a case where you are waiving your autonomous decision; you are not giving up your autonomy and putting yourself in his care. It is still you who, to use rather grand words, determines your life on that night.

Matters seem to be analogous in the clinical setting. A patient can take an autonomous decision to undergo a medical treatment without having (positive) knowledge of the treatment and risks. If I, as a patient, have a choice between giving an IC or a negative IC, it seems that the latter alternative may well be the one that both respect and promote my autonomy better than IC. Furthermore, if I, as the patient, choose to let you, as the physician, determine my treatment, and I have well founded beliefs that you will choose the treatment that best promote my values, and that the risks of the treatment you will choose, is in accordance with my attitudes towards different kinds of risks, I will exercise my autonomy, not waive my right to exercise it.

This should not come as a surprise. After all, many of the means we use are we not very familiar with. These states of ignorance do not that threaten autonomy.


It seems obvious to me that in some circumstances, negative IC is a better procedure than IC if we want to secure the autonomy of the patient. But, there are some objections against negative IC that are naturally forthcoming but not very worrisome.

One such objection says that negative IC may lead to paternalism, which is bad. True, it may lead to paternalism. But is equally true of IC where the physicians can frame the information to suit certain interests. Another objection concerns liability. It might seem troublesome to have negative IC from a legal point of view. Something might go wrong and if the case is brought to court there will be problems of verifying that the first two conditions that are unique for negative IC, 1 and 2 above, are satisfied. But this can be avoided by having the patient in print authorising a treatment without further information. A related objection concerns non-disclosure: negative IC may be an easy way of not withholding important information to the patient. But this overlooks that negative IC is not a question of concealing information, the information should be there if wanted. Moreover, if the conditions for negative IC are fulfilled, there will be no information that, the disclosure of which would lead the patient to revise the given consent.

There is, at least, one good objection against letting patients give their negative IC to medical treatments. It is the objection that a practice with negative IC is unfeasible. Condition 1 and 2 seem to require, not only that the physician knows more than most physicians possibly can know about their patients, but also that the patient have well-founded beliefs about what the physician knows, and well-founded beliefs that the physician will act accordingly. It should be noted that negative IC does not require that the physician should know that these conditions are satisfied, only that they are satisfied.

However, this rules out negative IC in situations where the physician and patient know little about each other. Negative IC requires a substantive patient-doctor relationship of confidence or trust. So, it can hardly be a reasonable practice in all kinds of medical contexts or all kinds of treatments. However, there are long-term treatments involving close patient-doctor relationship where these conditions can be satisfied, or something close to this. Moreover, my main concern here is really to argue that the requirement of IC can be relaxed without impairing patients autonomy, and for this to hold we may find a practice that lies somewhere between IC and negative IC.


I have argued that the relevant conception of patient autonomy is, with respect to understanding, weaker than the one that has dominated the debate in bioethics. I have also argued that given this conception of autonomy, the norm of IC to medical treatments ought to be relaxed in contexts where substantive patient-doctor relationships of confidence or trust are established.


This paper originated within the research project Monitoring and Improving Ethical and Medical Praxis in Perinatal Medicine, at the Research Program in Biomedical Ethics, Uppsala University and founded by Vårdalstiftelsen. I wish to thank A Dawson, MG Hansson, E Rynning, M Sheehan and A Wrigley for valuable comments on earlier drafts.



  • Competing interests: None.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Other content recommended for you