Basket | Login | Register


Risk, benefit, burden and paternalism

Ethics of dementia research

Risk is sometimes described as ranging from minimal to greater than minimal with various degrees of risk between those two poles such as a minor increase over minimal risk (Alzheimer’s Association, 1997, American Psychological Association, 1981). Risk is often discussed along with burden and balanced against possible benefit. There are limits to the degree of risk and burden that is considered acceptable for research involving people who are unable to consent, with greater restrictions being placed on research which does not entail personal benefit.

A risk is generally considered as being low if it does not exceed that associated with the kind of routine medical or psychological examinations that are typical for people from the same group as the participant (e.g. a person with dementia). The US Department of Health and Human Services (2011) defines minimal risk as “the probability and magnitude of harm or discomfort anticipated in the research are not greater in and of themselves than those ordinarily encountered in daily life or during the performance of routine physical or psychological examinations or tests”. The International Ethical Guidelines for Biomedical Research of 2002 allow for levels of risk which are slightly higher than those linked to routine medical or psychological examinations if there is an overriding scientific or medical rationale for the increased risk and provided that an ethical review committee has approved the study. An example of a high risk would be that detected in the phase II trials into the AN-1792 Alzheimer vaccine in which 6% of participants developed severe brain inflammation resembling meningoencephalitis which led to the trial being abandoned in 2002.  

Benefit, which has already been discussed in relation to the issue of consent, is also relevant to the assessment of risk. Perceptions of risk are often balanced against perceptions of possible benefit. As mentioned earlier, the definition of benefit is often too narrow to take fully into account the kinds of benefits that are possible and important to people participating in dementia research, such as spiritual, emotional or psychological benefit.

Burden, like risk and benefit, is subjective as it depends on people’s values, what they are willing to tolerate and what they do and do not find acceptable. Examples of possible burden include having to fast before research interventions, travelling to the research centre in bad weather or when one has reduced mobility, accepting changes in one’s daily routine, carrying out activities or tasks which are stressful, boring or repetitive, experiencing performance anxiety when asked to perform cognitive tests, having to spend long periods of time waiting or filling out forms and so on.

Paternalism is relevant to the issue of risk as it is one approach to minimizing risk. Paternalism usually involves the intentional restriction of a person’s freedom (e.g. the freedom to decide for oneself) on the grounds that the restriction will protect or promote that person’s wellbeing (Jansen and Wall, 2009). Usually the restriction applies to the person whose own good is allegedly being protected. However, in the case of indirect paternalistic interference, it may apply to other people (Dworkin, 1972 in Jansen and Wall, 2009). An example would be the refusal to grant authorisation to researchers to carry out a particular study in order to protect the wellbeing of willing participants (Jansen and Wall, 2009).

Balancing personal risk and benefit to society

The two following quotations reflect different positions regarding the possible gain to society and the risk or burden on/to participants.

 “….ethics is a matter of principled sensitivity to the rights of others … Ethics say that while truth is good, respect for human dignity is better, even if, in the extreme case, the respect for human dignity leaves one ignorant of human nature.” (Blumer cited in Dench et al., 2004, p4).

“It is not cruel to inflict on a few criminals sufferings which may benefit multitudes of innocent people through all centuries.” (Celsus cited in Sodeke, 2005)

The first statement emphasizes human rights and dignity, and clearly suggests that the pursuit of knowledge at all costs is unjustified and undesirable. It suggests that we must not lose sight of the safety, interests and wellbeing of the research participant through the relentless pursuit of knowledge. The second statement, which was made by a 1 st century physician, seems to suggest that suffering may be justifiable if a greater number of people benefit from it but also that it may be more acceptable to allow certain groups or categories of people to suffer for the benefit of others who are considered more worthy. Whilst criminals may have had few or no rights in the days of Celsus, the same argument could all too easily be applied to other groups of people who become marginalized or devalued based on a range or combination of factors such as age, mental health status, prognosis, gender, ability to contribute towards society, sexual orientation and ethnicity.

In the International Ethical Guidelines for Biomedical Research of 2002, benefit to society is defined as being “generalizable knowledge”. This reflects a positivist approach to research. Whilst these guidelines are clearly linked to biomedical research, it is important to bear in mind that generalizability is not the main aim of qualitative research. Whilst the results of a small-scale qualitative study may have relevance to other groups of people sharing similar characteristics, this is something that readers can decide for themselves based on the details provided by the researcher. It is not determined by statistical analysis.

The risk of considering one group of people as sufficiently unworthy as to justify their suffering for the good of the rest of society may be higher if members of that group are stigmatized or lack the power or capacity to protect their own interests, hence the necessity to ensure that people with dementia who take part in research are sufficiently protected. On the other hand, care must be taken to respect the autonomy of people with dementia and to avoid stereotyping them as defenseless victims, treating them as children and even limiting their opportunities to contribute towards research.

Making sense of risk

Historically, risk was considered as something that could be either good or bad, resulting in loss or gain, but Lupton (1999) claims that it has been transformed into something that is entirely negative. Within the social and healthcare system, risk has often been viewed as something that is objective and measurable (Kemshall, 2000), whilst others argue that it is socially constructed and value-laden (Lupton, 2005). Risk can also be perceived as opportunity in that it may result in change.

When weighing up the seriousness of possible risk, the possibility of such an outcome occurring should also be considered. In everyday life, for example, certain drugs for relatively trivial health problems have possible side effects which although extremely rare would be devastating if they occurred. Yet large numbers of people take them every day either without reading the information leaflets or in full knowledge of the risk involved. The likelihood of certain risks occurring may be affected by the situation in which one lives (i.e. whether a person lives alone or has the support of family, friends and the local community).

As a unique individual, each person with dementia has his/her own perception of danger and of the level of risk that s/he is willing to take. Having dementia may even be an additional motivation to take part in research (e.g. to benefit relatives and other people who may develop dementia in the future or as a means to contribute to society in some way). However, as the disease progresses, the capacity to give informed consent to research is likely to deteriorate. For many people, another person will eventually be consulted whose assessment of acceptable risk may determine whether or not they participate in research.

In the context of consent to research, none of the interventions are actually necessary for the participants (except perhaps in the case of emergency research) and people with dementia are considered as being potentially vulnerable research participants. There is therefore a risk of overprotecting people with dementia, infantilising them and depriving them of their autonomy. This could in the long run deprive people with dementia of the possibility to benefit from the results of studies into their particular condition. In some cases, such over-protection may reflect concerns about litigation processes in case of harm as well as genuine concern for the wellbeing of research participants.

The perspective of risk and benefit at the end of life

Glannon (2006) suggests that the relative significance of a possible benefit or risk depends on one’s perspective. With regard to a particular study, outsiders and the potential participants may consider the possibility of benefit occurring and the possible risks involved differently.  A person with dementia receiving end-of-life care may place great value on even the fraction of a hope of benefit despite an awareness of a much greater probability of harm occurring, including the emotional consequences of having one’s final hopes shattered. In such cases, focusing on an improbable but possible benefit (even though this is not the purpose of research) would not be a therapeutic misconception as the person would be fully aware of what was likely and what was at stake.

With regard to benefits which are not directly linked to health, Casaret, Jason and Karlawish (2000) point out that when people are approaching death, the things that they find important may change. They may, for example, place a greater value on dignity, meaning, control, strengthening relationships and addressing unfinished business of a personal nature. In the case of people with advanced dementia, the expression of such concerns or priorities may be difficult but this does not mean that they do not exist. Participating in research may, for some people with dementia, correspond to these new priorities (e.g. by providing meaning or enabling them to feel part of something). For others, it might actually interfere with current goals and priorities (e.g. by taking up valuable time they might prefer to spend with relatives or friends).

Similarly, a person who is very close to death might not assess the possibility of serious risk in the same way as a person who has not yet reached that stage. Some risks may be considered as being more significant and others less so (Casarett, Jason and karlawish, 2000). This would depend greatly on the individual as well as on his/her awareness of his/her prognosis.

One-off or continuous assessment of risk and burden

Robley (1995, in Aita and Richer, 2005) points out that many qualitative studies are unpredictable and that this does not permit a proper evaluation of the benefits/risks ratio which in turn raises concerns about the rights of participants. This is because qualitative research tends to be a flexible and iterative process in which the data is analysed continuously rather than at the end of the data collection process. This may lead to insights which are fed back into the process and may lead to changes in the way data are collected or in any other aspect of the research process (Maxwell, 2005). Whilst ethical approval may have been sought and obtained from the appropriate ethics committees prior to the study, such approval often fails to take into account the iterative nature of qualitative research. Committees therefore oblige researchers to provide precise details (such as interview schedules) and describe various aspects of the study and then grant or refuse approval on that basis. This may result in researchers feeling uneasy about their own scientific integrity and could be interpreted as a failure to accept the validity of research based on the interpretivist paradigm.  Obtaining research ethics approval and governance approval (e.g. from government health services) can be an extremely lengthy process (at least in the UK) and it would be impractical to submit every change along the way to additional approval (although there may be procedures in place for “substantial amendments”).


It has been suggested that studies which clearly do not involve more than minimal risk should perhaps be dealt with more expediently than those involving a higher degree of risk by the IRB (an institutional review board which is a committee that performs ethical review of proposed research) (Emanuel and Menikoff, 2011). It is further suggested that as for many studies the main potential risk is linked to the confidential handling of data, it should be possible for research to commence immediately after the study has been registered and after a one-page form has been submitted accompanied by a commitment to observe data-security protections. However, whilst the ethical review of safety issues does seem exaggerated and clearly inappropriate for many studies, there are other ethical issues to be considered by ethics committees. The National Research Ethics Service (2011) defines its objectives as follows: “to facilitate ethical research that is of potential benefit to participants, science and society …./…. to provide robust, ethical review of proposed research via Research Ethics Committees (RECs). These independent committees put the rights, safety, dignity and well-being of research participants at the centre of their decision making.” Consequently, there may be a need to simplify the approval process for projects involving minimal risk but not at the expense of ensuring the pursuit of good science. 


Jansen and Wall (2009) consider the arguments for and against paternalistic interference in the context of research. First, they argue that paternalistic interference (in clinical trials) is only ever justifiable in the case of studies which impose a serious risk of harm to participants or if the therapeutic benefits do not outweigh any risks of harm. This argument could perhaps also apply to other types of research, such as psychosocial research which might involve psychological risks. According to the principle of distributive fairness, research involves risks as well as benefits which should be shared equally amongst members of society in the sense that specific groups of people should not have an unequal share of these (neither harms nor benefits). Jansen and Wall (2009) identify four different types of harm which paternalistic interference may help protect against: physical harm, psychological harm, social harm and economic harm.

The main arguments against paternalistic interference are that individual liberty (the right to decide for oneself) should take precedence over issues related to welfare and distributive justice and that paternalism is insulting to the individual whose liberty is being restricted (Jansen and Wall, 2009). These arguments are mainly relevant to people who have the capacity to make their own decisions about research and the level of risk they are willing to take (which would nevertheless include many people with dementia, at least but not necessarily limited to those in the early stage), even if such decisions seem foolish.

Miller and Wertheimer (2007) suggest that a person who appears to be willing to seriously endanger his/her life or wellbeing for the sake of science must be insufficiently informed or be subject to some form of “autonomy bearing pressure”. However, it must be borne in mind that people may be taking part in research for altruistic reasons which could be considered as a good reason. Taking part in the same study out of self-interest might be considered as a bad decision in view of the risk to oneself. Bearing in mind the fact that foolish decisions are ethically acceptable provided that the person has sufficient capacity, this is no reason to impose restrictions on participation.  Also, there is no reliable method of ascertaining whether a person wants to participate in a study for altruistic reasons or for self-interest. Another category of participant might include people who have no concern for their own personal wellbeing. This might be attributed to depression but could presumably exist in the absence of any disorder.

Jansen and Wall (2009) argue in favour of paternalistic interference on the grounds that it is unfeasible to have different rules and levels of protection for different research participants as it would be necessary to determine for each person which set might be applicable. A level of protection that is too lax would put the people who are at greater risk of making a bad decision at a disadvantage. Stricter regulations are therefore justifiable.   

The second argument outlined by Jansen and Wall is that paternalistic interference is insulting in that even if it does not violate the right to autonomy, it implies that the potential participant is not capable of making a decision. They argue that this is an unfair assumption as regulations are not targeted at individuals and cannot therefore be justifiably perceived as a personal insult. Nevertheless, it may still be argued that it is unfair to restrict the liberty of some people for the sake of protecting the interests of others which could be perceived as an individualist approach which does not recognise the value of distributive fairness.

  • People with dementia (and their carers, provided that the person with dementia agrees), should be involved, to the greatest extent possible, in assessing the level of risk they are willing to take and the level of burden they are willing to accept when designing studies.
  • An inventory of possible risks and burdens should be made for each study.
  • Researchers should be transparent about potential risks and burdens, and communicate these to participants in a way that the latter can understand.
  • Strategies should be devised to address or minimize risk and burden.
  • Researchers should strive to find the right balance between protection and the promotion of autonomy.
  • Paternalistic practices and attitudes should be avoided.
  • People with dementia should not be treated like children or stereotyped as helpless.
  • It should be recognised that zero risk may be a worthy aim in the development of medication but that it is not necessarily realistic or even desirable in all other aspect of people’s lives.
  • Researchers should have an insurance for any harm resulting from the research. 
  • Potential participants should be invited, if possible, to speak about possible expected benefits linked to participation in a particular study and details of their responses should be noted.
  • Participants should be asked after the study about their experience of it (including details of benefits and burdens) and details of their responses noted.
  • Details of the responses from participants (or in some cases their proxies) about the perceived risks, benefits and benefits of participation should be taken into consideration by researchers when designing subsequent studies.



Last Updated: jeudi 29 mars 2012


  • Acknowledgements

    The above information was published in the 2011 Report "The ethics of dementia research" as part of Alzheimer Europe's 2011 Work Plan which received funding from the European Union in the framework of the Health Programme.
  • European Union


  • Send this page to a friend