What is a proper presumption in an ethical assessment?

A presumption in an ethical analysis allows a conscientious agent to select a favoured option in an urgent situation or when a detailed ethical analysis is impossible for different reasons. For instance, in example 1, the presumption is to save the life of a dying patient when it cannot be ascertained whether the patient wishes to be saved. Put another way, if a clinician decides not to give a life-saving treatment, the burden is on the clinician to prove that there is a valid expression of the wish not to be saved probably in the form of an advanced directive.

• Example 1: Life-saving treatment (where the patient's wish is not known)

• Option A₁: To let die

• Possible benefit: wish to die followed

• Possible harm: life sacrificed

• Most probable outcome: life sacrificed (presumably most people want to live on)

• Option B₁: To save life

• Possible benefit: life saved

• Possible harm: wish to die violated

• Most probable outcome: wish to die not violated (presumably most people want to live on).

Effort should be made to ascertain whether the patient has previously expressed a wish not to be resuscitated so that an ethical decision of whether to save life does not need to be made based on a presumption. However, when the patient's wish is not yet known, most would agree that the presumption should favour option B₁ because the damage of letting die is irreversible, whereas the patient may claim the right to die at the next available opportunity. Based on the assumption that most people want to live on, the potential harm associated with option B₁ is also less likely to occur than the potential harm associated with option A₁. Therefore, an ethically sound presumption in example 1 should favour option B₁ whose harm is of less gravity and less likely to occur than that of option A₁.

Two options in scenario 2

A full ethical analysis to assess the possible benefits and harm for the patient and potential third parties may be impossible in scenario 2 if the patient refuses to reveal further information including the sexual history. In this situation, the presumption may favour either one of the following two options:

• Scenario 2 (where the patient refuses to reveal further information)

• Option A₂: Not to tell HIV status

• Definite benefit: the right to privacy respected

• Possible harm: serious harm to third parties

• Option B₂: To tell HIV status

• Possible benefit: third parties saved from harm.

• Definite harm: right to privacy infringed

Should we invariably favour the option associated with harm of less gravity?

An ethical analysis becomes more complicated when it involves the competing interests of more than one parties. Some people may suggest that the gravity of harm in option A₂ for the potential third parties in scenario 2 apparently outweighs that for the patient in option B₂. They may further propose that when there are competing rights between two parties, the option associated with harm of apparently less gravity should always be favoured regardless of the likelihood of the harm occurring. My response is that this viewpoint has to be vigorously refuted. In example 2, X dies and his organ may be harvested to save Y's life. However, X has previously expressed his wish against organ donation. Our options are to follow X's wish or harvest his organ against his wish. Even a consequentialist would unlikely suggest that option B₃ which can save a human life should be favoured.

• Example 2: Organ donation

• Option A₃: respect X's wish

• Outcome: X's right to dignity/privacy/bodily integrity is preserved

• Harm: Y will die

• Option B₃: harvest the organ against X's wish

• Outcome: Y will live on

• Harm: X's right to dignity/privacy/bodily integrity is breached.

In some countries where an opt-out system is in place for organ donation, the presumption may lie in favour of harvesting the organ if the wish of the patient is unknown. However, in case the patient has opted out from donation, most conscientious agents would not suggest that it is proper to harvest his organs against his wish, even though the harm caused by option A₃ is arguably of higher gravity than that caused by option B₃. There are two reasons. First, the right of a dead person to dignity and/or bodily integrity and/or privacy is still worth our utmost respect. Second, X is not ethically obliged to give an organ to Y and Y has the right to accept a donated organ but not the right to take X's organ against his wish. It can be seen from this example that the gravity of the harm of different options should not be the only decisive factor in an ethical analysis. So, even on the premise that the gravity of harm associated with option A₂ is higher than that of option B₂ in scenario 2 (which one may deny), in the determination of a proper ethical presumption, it will be relevant to consider also the rights and ethical obligations of different parties and the likelihood of the two options to cause harm.

The ethical obligations to protect a third party from harm in scenario 2

In scenario 2, a potential benefit of option B₂ is that the potential third parties who are also HIV positive may receive HIV treatment. Assuming that a third party can be located, the next relevant issue is whether the patient has an ethical obligation to notify this third party so that they can be tested and receive treatment if needed. The possibilities are two fold. First, the patient contracted the virus from this third party. In this case, the patient has no ethical obligation to prevent harm for this third party. Second, the patient transmitted the virus to this third party. It is vital to bear in mind that it is the virus that causes the harm where the patient is only a reservoir for transmission of the virus. It cannot be said that the patient is ethically culpable for the harm as the patient did not transmit the virus to the third party with intention (rightly so assumed because the patient does not know the result of the HIV test yet). Given the harm already done (for which the patient has no ethical culpability), it is difficult to argue that the patient has an ethical obligation to save this third party from further harm. There may be conscientious obligation (for instance for the spouse) but very likely there is none because if a conscientious obligation is present then the patient would not likely have refused to be told the HIV status.

Another potential benefit of option B₂ is that further transmission of the virus may be prevented. Most people would agree that the patient has an ethical obligation not to transmit the virus and the potential third parties have the right not to be harmed. However, an ethical obligation not to harm arguably does not arise so far as the patient is not yet aware of his HIV status. In fact, further transmission can be effectively prevented only if the patient modifies the sexual behaviour, notifies the sexual partner(s) and/or takes antiviral treatment. However, as the motivation to notify the partner(s) was shown to be influenced by the patient's sense of ethical responsibility, concern for the partners' health, the severity of the disease and cultural factors,8 a patient in scenario 2 is probably less motivated to do so. Nor would a patient disinterested to know the HIV status be motivated to take antiviral treatment. Furthermore, a recent study by Kalichman et al showed that the association between HIV treatment and prevention of further transmission hinges on sustained treatment adherence and enhanced behavioural intervention.9 Unless there is a system of mandatory partner notification and HIV treatment (with monitoring) which there is not in most countries at the moment, I submit that the aim of preventing further transmission is too remote to justify a presumption to infringe the patient's right of refusal (even if we grant that the ethical obligation not to further transmit the virus exists).

The likelihood that a third party can benefit from option B₂

In case one argues otherwise that a patient in scenario 2 has an ethical obligation to protect the potential third parties from harm, the following empirical data suggest that the likelihood that option B₂ can benefit those third parties would not be high enough to justify a presumption in its favour. In a systematic review of studies about the effectiveness of partner referral service, a mean of 67% of identified partners were informed of their potential exposure to HIV, and a mean of 63% of those informed came forward for a test. Of those tested, a mean of 20% were HIV positive. In short, there is only 1% to 8% chance to identify an HIV-positive partner from each index case through partner notification.10 As a patient in scenario 2 is likely less motivated to participate in partner notification or referral, the chance that option B₂ can save an HIV-positive third party from harm should even be less than 1% to 8% as cited in the literature. In other words, option A₂ has a fairly low chance to cause harm.

Some people may then suggest that we should focus our discussion more on the welfare of the entire population from the perspectives of public health. There is no doubt that further transmission of HIV may be maximally prevented through coercion, e.g. by passing a law that forces certain groups of people to undergo an HIV test, be informed of their HIV status and receive treatment against their wish. But I argue that a presumption in favour of option B₂ will not be a fair public health intervention. A clinical trial conducted in Mexico showed that the behavioural intervention to promote condom use did not result in significant reduction in the cumulative incidence of HIV.11 In a randomised controlled clinical trial to study the effect of treatment on HIV transmission, in which 1763 infected subjects with CD4 counts between 350 and 550 cells per cubic millimeter were recruited,12 it was shown that HIV treatment has to be given to 60 patients to prevent one case of further HIV transmission in serodiscordant couples.13 This is not at all a claim to discount the effect of providing behavioural intervention or antiviral treatment as a part of the public health strategy to reduce the spread of HIV infection. In contrast, clinicians should encourage all those who may benefit from a treatment to receive it as it is good for the patient and is potentially good for a third party. However, it is a different matter when we consider a public health intervention which encourages infringing the right of refusal of a huge number of patients just to benefit very few other third parties. Most conscientious agents would require that a public health measure should adopt the utilitarian approach to benefit the majority (of those affected) and harm as few as possible. While these empirical data may not fully fit the context of scenario 2, they preliminarily suggest that the magnitude of incremental benefits would probably be too small to substantiate a public health intervention to have a presumption in favour of option B₂ in scenario 2.

Presumption should favour option A₂

To make a presumption in scenario 2 in favour of option B₂, the substantiating evidence has to reach a very high threshold because option B₂ invariably infringes a patient's prima facie right not to know. Considering that the patient has no clear ethical obligations for the third parties potentially at risk of harm, that infringing the patient's right is too remote from preventing further transmission, and most importantly, that option B₂ is not likely to benefit any third party, I argue that the threshold is not reached. Because option B₂ will result in definite harm to the patient likely without any benefit for the potential third parties, option A₂ should be favoured in the presumption.

• Scenario 2 (where the patient refuses to reveal further information)

• Option A₂: Not to tell HIV status

• Definite benefit: the right to privacy respected

• Possible harm: serious harm to third parties

• Most probable outcome: right to privacy respected, no third party harmed

• Option B₂: To tell HIV status

• Possible benefit: third parties saved from harm.

• Definite harm: right to privacy infringed

• Most probable outcome: right to privacy infringed, no third party saved from harm.

For similar reasons, the presumption should favour preserving a patient's right to refuse an HIV test in scenario 1. However, it cannot be over-emphasised that although the favoured option may prevail (particularly when the patient refuses to disclose sufficient information), where possible, a detailed ethical analysis should be engaged to assess how likely a third party may benefit from infringement of the patient's refusal. The disease severity, patient's attitudes towards treatment, number of sexual partners, sexual behaviour, identifiability of partners, patient's willingness to notify partners are all relevant factors. At the conclusion of the ethical assessment, a conscientious agent should be free to select either option without regard to the presumption.