Basket | Login


Part 2: Ethical Challenges Linked to Recruitment and to Informed Consent

2019: Overcoming ethical challenges affecting the involvement of people with dementia in research

Part 2: Ethical Challenges Linked to Recruitment and to Informed Consent

The importance of involving a diverse set of participants in research

In Part 1, we looked at issues related to the contribution of people with dementia to research through Public Involvement. This involves them sharing their perspectives, opinions and insight with researchers, thereby having a voice and helping improve the research process. We now turn to issues related to the contribution of people with dementia to research through participation in research studies (i.e. as a research participant) which also raises several ethical challenges in relation to the participants, the researchers and the researcher process.

In accordance with the requirements for good science, researchers must recruit participants who are best suited to providing a response to their research question or hypothesis. Depending on the methodology adopted, this means involving people who are representative of the group targeted by the study and/or who can together produce a broad range of perspectives and experience on a particular issue. Populations are made up of very different people and there is a risk of people with certain characteristics being excluded. All too often, the very people who are most in need of research to improve their condition or situation are the ones who are excluded (Sin 2005). In addition, and in keeping with the principle of justice, measures must be taken to ensure ‘fair subject selection’ whereby the most stigmatised and vulnerable individuals in society are not selected for participation in risky research and the rich and socially powerful for research more likely to entail some personal benefit (Emanuel, Wendler and Grady 2000).

Historically, a first obstacle was to ensure that people with dementia were involved as participants in dementia research. It was often assumed that they were incapable of communicating their thoughts and feelings and that data obtained from them would be unreliable (Hubbard, Downs and Tester 2003). This was combined with an emphasis on biomedical research (Lepore et al. 2017). Regarding clinical trials, people with dementia are the ones who will eventually be using dementia drugs and therefore they need to be involved in their development. However, some groups of people with dementia (e.g. those who are no longer living at home, those without a partner and those with comorbidities) are often excluded from clinical research. This means that the study populations are often not representative of the broader population of people with dementia (Jongsma, van Bruchem-Visser, van de Vathorst and Mattace-Raso 2016). The informed consent process and the inclusion and exclusion criteria for studies are often stumbling blocks in terms of involving a diverse set of people with dementia in research. Behavioural and psychological symptoms of dementia (BPSD) may also sometimes result in exclusion from various studies (e.g. if researchers do not feel able to cope or have concerns that a potential participant might not comply with research requirements).  

People with dementia are best placed to provide information about their experience of living with dementia and their reactions and preferences on a range of issues (with all necessary care, attention and support having been provided to make this possible and to obtain informed consent, sometimes using non-written, non-verbal options). Unfortunately, the perspectives and experiences of people with dementia are still sometimes unnecessarily expressed via the intermediary of informal carers or health and social care professionals. At the same time, it is important to bear in mind that for some ethnic minority groups, the consent of the family is also very important. Similarly, country-specific guidance and legislation should be followed when including people with dementia who lack capacity and where opinions are sought on their behalf from those legally permitted to do so. Restricting the informed consent process solely to the person with dementia may be too individualistic an approach within some communities (Berdai-Chaouni, Claeys and De Donder 2018).

Finally, many people with dementia find themselves excluded from research into issues that are not specifically related to dementia or into medical conditions other than dementia. This seems to reflect an assumption that having dementia is the sum total of a person’s identity, overshadowing everything else, including experiences, hopes and perspectives linked to other aspects of people’s lives, including other diseases.

Dementia is experienced in different ways by different people. People with dementia have different levels of support, different social relationships and roles, and different capacities. They also have a range of different cultural backgrounds with many people identifying with a number of sub-groups including ‘communities within communities’. The diversity of people with dementia should be reflected in research samples but often isn’t.

As mentioned earlier, people who have lower levels of education, intellectual disabilities, lower health literacy, more advanced dementia or are from a minority ethnic group are frequently excluded, as are many older people living with dementia who may also have other health conditions (e.g. depression, cancer or diabetes). People with dementia should not be excluded or miss out on opportunities to take part in research solely on the grounds of characteristics they may or may not have. Ignoring or being ignorant of the needs and interests of large proportions of people with dementia, resulting in them not being involved in research, runs counter to the principle of equity and involves epistemic injustice. It means that some people with dementia are not being granted the same opportunities as other people to contribute towards science (with the potential burdens, risks and benefits that this entails) and thus to society. With regard to epistemic injustice, they are not being recognised or considered as knowers, which is harmful for them in that it leads to discrimination and in researchers missing the opportunity to hear their voice, resulting in limited and biased knowledge about the topic.

Whilst the aim of research is not to benefit individual participants, many participants do feel that they gain something personally out of participation such as a feeling of personal satisfaction from pursuing altruistic motives, increased contact, a feeling of hope, of having a purpose and of empowerment, and access to ‘state-of-the-art’ consultation and disease monitoring, information and management. Potential benefit must of course be balanced against potential harm. However, an overemphasis on harm may lead to the possible benefits of participation being overlooked. Hellström et al. (2007) conclude that there are serious consequences to not including people with dementia in research and suggest that the question should not so much be about whether people with dementia should be included in research, but rather how we can best achieve this and how we can afford not to. Excluding some groups of people from the opportunity to experience such perceived benefits could be considered as discriminatory and institutional racism (Hussain-Gambles 2003).

From a scientific perspective, failing to involve a diverse set of people with dementia in research means that the responses to research questions and hypotheses are incomplete, misleading and at worst wrong. Konkel suggests that not only does the failure to involve a more ethnically diverse set of people in research exacerbate health disparities but it also represents “a missed scientific opportunity to fully understand the factors that lead to poor health and disease” (2015, p. 298). Bhopal and Sheikh further suggest that the exclusion of people from minority ethnic groups “promotes inequality, maintains inequity and is unethical and institutionally racist, if not illegal” (2009, p. 2). From a disability and human rights perspective, appropriate measures should be taken to ensure that people with dementia are enabled and empowered to take part in research on an equal basis as people without dementia.  The following general principles from Article 3 of the CRPD were developed in the context of attempts to ensure the equal involvement of people with disabilities in society could be equally helpful when considering the involvement of a diverse set of people with dementia in research: 

  • Respect for inherent dignity, individual autonomy including the freedom to make one’s own choices, and independence of persons,
  • Non-discrimination,
  • Full and effective participation and inclusion in society,
  • Respect for difference and acceptance of persons with disabilities as part of human diversity and humanity,
  • Equality of opportunity,
  • Accessibility,
  • Equality between men and women.


Recommendations for researchers, research ethics committees and funders

Strive to address inequity so as to ensure that a broad range of people with dementia are involved in research as participants for the following reasons:

  • People with dementia should all have an equal opportunity to be involved in research (in keeping with the principles of justice, solidarity and respect for autonomy).
  • Excluding certain groups of people or people who possess or lack certain characteristics would be discriminatory.
  • Having advanced dementia, or lacking capacity to consent, does not necessarily mean that a person may not participate in research.
  • Findings cannot be generalised (in the context of quantitative research) to people with dementia if participants were from a small sub-section of people with dementia.
  • Findings cannot be considered to represent the diversity of people with dementia if participants were from a small sub-section of people with dementia.
  • Failure to do so may result in treatment, services, care and support which only corresponds to the needs of a sub-section of society with the needs of other members of society going unmet.
  • Failure to do so results in societies failing to benefit from the wealth of experience and of potentially different ideas, perspectives and solutions.
  • Failure to do so prevents the creation and maintenance of an inclusive society.


Additional recommendation for research ethics committees

  • Recognise that much research is still required into supporting people with advanced dementia and this population should not be excluded from participating in research provided that appropriate processes have been followed.



Researchers must recruit participants who are best suited to providing a response to their research question or hypothesis. The principle of justice should be respected which means that everyone should have the same opportunity to take risks, enjoy the benefits and generally contribute towards society through participation in research. Some groups of people with dementia continue to be excluded with the result that the diversity of people with dementia, which should be reflected in research samples, often isn’t. The people who are excluded often have lower levels of education and health literacy, intellectual disabilities, more advanced dementia, co-morbidities or are older or from minority ethnic groups. This exclusion is unethical and unacceptable, leading to some groups of people with dementia being disempowered and their needs and wishes ignored, and representing missed scientific opportunities and incomplete or misleading results.

Attracting interest and addressing barriers to research participation

A key determining factor for the involvement of any person in research is that he or she knows about the study and realises that it would, subject to the fulfilment of certain conditions, be possible to be involved. Even though members of the general public are frequently informed about the latest research findings in newspapers or on the Internet, or hear about breakthroughs on the television, they do not necessarily make or understand the connection between research and those breakthroughs, hear about studies which are recruiting participants or know how to find out about them. Awareness of research and willingness to participate does not, however, necessarily translate into people becoming research participants. Other factors may affect participation such as, for example, a person’s current medical condition, cultural and linguistic factors, financial and time constraints, employment constraints, geographical proximity to research sites and mobility/travel issues (Wendler et al. 2005, Konkel 2015). The timing and effects of various medication may also affect the desire to participate (e.g. people feeling groggy in the morning, needing support to take medication).  In this section, we look at some of the key issues linked to the initial task of informing and arousing the interest of a broad range of people with dementia in research, and how this in turn may affect the participation of a diverse set of research participants with dementia.


Language and literacy

In Part 1 of this report, the issue of literacy and language was raised as a barrier to the involvement of some people with dementia in PI. Difficulties with language, literacy, as well as the ability to use computers, may also result in some groups of people not being aware of research for which they may be suitable participants. Increasingly, Internet platforms are being set up to inform the general public about research opportunities and to encourage people to participate in research[1]. However, this caters for people who are literate, able to use computers and who have Internet access, which is not always readily available or affordable. As these approaches become increasingly popular, there is a risk of more traditional methods of contact becoming less common and of knowledge about (and hence access to) research being ever more restricted to people with a similar set of characteristics.

Differences in levels of health literacy may also affect people’s interest in participating in research. Health literacy is commonly defined as “the capacity to obtain, process, and understand basic health information and services needed to make appropriate health decisions” (Brach et al. 2012, p.1). In this sense, it is associated with healthcare decision making but failure to understand information about health may also affect motivation to contribute towards health-related research (e.g. to improve diagnosis and quality of life, develop treatments, better understand how people are affected by certain conditions and develop non-pharmaceutical therapies and approaches to disease management). In other words, if people do not understand the topic of research and why it is important to address it, it is unlikely that they would be motivated to participate.

Linguistic, educational and cultural issues may also be significant barriers to the participation of people with dementia in research because many of the tools used to diagnose dementia or measure disease progression, as well as instruments to measure research outcomes (all of which are essential components of many studies), are not suitable for people with dementia from linguistically and culturally diverse groups. Some have been translated but many have only been validated in populations from the Western culture. Moreover, language is not the only obstacle as the tools and instruments may in addition be culturally biased[2].  The fact that these have not been developed is an example of inequity and a structural barrier to inclusive research which needs urgent attention. Moreover, the tendency to focus predominantly on written informed consent is also a barrier (e.g. for some older migrants and people with language difficulties) and recorded verbal consent should be considered as an alternative (Berdai-Chaouni, Claeys and De Donder 2018).   

Trust and legitimacy

One possible approach to raising awareness about opportunities to participate in research is through the intermediary of health and social care professionals who have access to a broad range of people. However, distrust of such professionals and the fear of racism are common amongst some minority ethnic communities (Shah 2007, APPGD 2013, Kenning et al. 2017, Condon et al. 2019). The fact that scientists have not always been trustworthy in the past and their general lack of public accountability may hamper the willingness of people to participate in current scientific research. Where there have been healthcare research scandals involving scientific misconduct (such as the Tuskegee study in the United States as described in the introduction), trust in researchers may be particularly low, in general and in particular amongst some minority ethnic groups (Daigle 2019). However, it should not be assumed that all members of a particular minority ethnic group share these concerns or that awareness of such historical abuse is the sole reason for distrust in healthcare professionals and researchers (Wendler 2005).

With specific reference to people from gypsy[3], Roma and traveller communities, Condon et al. (2019) point out that frequent mistrust of authorities occurs within the context of a long history of genocide, banishment, discrimination and rejection by mainstream society. There has also been a long history of abuse, discrimination and persecution against people from the LGBTI[4] community, with homosexuality until quite recently having been illegal and considered in the medical profession as a form of sexual deviation (Peisah, Burns, Edmonds and Brodaty 2018). Many older people will have hidden their sexual orientation, gender identity or intersex status for many years and will be hesitant to engage in research which might result in it becoming known.

People with intellectual disabilities have also throughout history been marginalised. According to Rosner (2015), this may have a negative impact on their willingness to participate in research in that they may have low levels of trust in their own capabilities. Such self-selection/restriction may not always be evident. For some people, it may be linked to language use (the way things are communicated). It may be possible for researchers to build up trust by using the same language.  It should not be assumed that everyone who seems to be able and free to participate in research shares that opinion. Low levels of education, lack of confidence in one’s own abilities, difficulties communicating, internalised negative stereotypes and assumptions that participants would not be allowed or welcome (e.g. for people in group living environments or in situations of dependency) may all contribute to some people with intellectual disabilities not following up on opportunities to participate in research.

The issue of distrust may also be related to the perceived legitimacy of research projects in the sense that a lack of trust in the medical profession and researchers may raise questions about the value and purpose of the study. People from groups which are typically excluded from research may be suspicious about why they are being invited to participate or, conversely, may not believe that the goals of the study are in the interests of people from their communities. Indeed, much of the research which they have the opportunity to participate in does not necessarily reflect the interests or priorities of people from their communities (although there is no reason why such research should be their sole interest).

“Many researchers have been using our community people, our culture, and our knowledge to do business. For example, lots of student researchers get their academic degree through using our knowledge……. I see this research as a business making use of our community’s people, culture and practice”(comment from a person from an indigenous community, cited in Datta 2017, p. 10-11).

Key persons or figures from various communities may therefore be crucial in providing access to people from various communities. It is important to develop a trusted and respectful relationship with these key people and not to ‘use’ them. Sometimes, potential participants from such communities do not even want to have full details about the research as they trust the judgement of the key person. If that person says that the research is good and that the researchers are trustworthy, then people will follow.

In contrast, it is important to be aware of some pitfalls related to trust. Research into the motivations of people to participate in research shows that trust is an important factor influencing decision making (Tromp, Zwaan and van de Vathorst 2016). This trust is very diverse, including, for example, personal trust in researchers or care professionals, institutional trust, trust in research in general and trust in the overarching system (Tromp and van de Vathorst 2018). It is crucial not to take advantage of these types of trust when approaching people for research participation. 

Racial/ethnic bias

In most cases, the exclusion of certain groups of people in research is unintentional but unseen racial biases[5] may sometimes affect who gets recruited into studies (Konkel 2015). Researchers may, for example, assume that a person from a particular ethnic group would not want to participate or unconsciously overemphasise negative aspects of the study which may deter the person from deciding to participate. Bias against studies which would have involved more people from ethnic and other minority groups (e.g. on issues of particular relevance to people from those groups) may also occur at the funding stage. According to Konkel (2015), studies in the United States which focus on health disparities and minority populations are less likely to receive federal funding. This may be partly explained by the fact that only approximately 10% of those responsible for reviewing funding applications are from minority groups.    

As with the involvement of people from minority ethnic groups in PI, it is often assumed that people from those groups do not participate in research as participants because they are not interested. There is nevertheless evidence to suggest that people from minority ethnic groups are just as willing to participate as their counterparts from majority ethnic groups when given the opportunity and provided that the research objectives are presented in a way which is culturally meaningful (Wendler 2005). This means that it is clear what the benefits of the research could be for them or their community and that this is presented in a way that is familiar to them. Wendler concludes that measures need to be taken to facilitate equal access to health research for all minority groups, rather than focusing on trying to change attitudes of people from minority ethnic groups.


The perception and portrayal of dementia and of people from various minority groups

Dementia has, over the centuries, and still is perceived in numerous different ways (e.g. as a divine punishment or test for wrongdoings, an organic brain disorder, a mental disorder, a natural consequence of ageing, something that is contagious and a sign of neglect and poor care of the person concerned) (Alzheimer Europe 2018). The different ways that dementia is perceived and portrayed has been long associated with barriers to help seeking, diagnosis and the uptake of services, and may also affect readiness to contribute towards research. People with dementia are not always treated as equal, respected and valued members of society. Society in general still has a long way to go in overcoming stigma and prejudice associated with dementia. Many people report feeling marginalised as a result of having dementia (Alzheimer’s Disease International 2012). In some minority ethnic groups, an additional factor to consider is the ‘newness’ of dementia as a condition. Members of those groups are less familiar with dementia. This may, for example, be due to recent migration, the population being younger or lower levels of education.

Diagnostic labels can be useful in the clinical context and help people understand difficulties they may have been experiencing for some time. However, Garand, Lingler, Conner and Dew (2009) suggest that some people with dementia may be reluctant to take part in research because they (or in some cases their families) wish to avoid the stigma frequently associated with the diagnostic label, which could take the form of public stigma or self-stigma[6]. Public stigma involves the identification and labelling of people with a socially salient attribute (such as having dementia), attaching negative stereotypes, considering people with that attribute as being ‘other’ (i.e. not like ‘us’), devaluing them and discriminating against them (including denial of equal opportunities as well as social distancing) (Link and Phelan 2001). The attribute is not inherently stigmatising but comes to be considered as such because of the meanings attached to it which are socially salient and generally involve a perception of some kind of threat (e.g. to individuals, society or morality) (Stangor and Crandall 2000). Self-stigma involves the internalisation of negative stereotypes, resulting in low self-esteem and the anticipation of discrimination. Having a lower level of understanding of a particular condition has been associated with greater stigma (Corrigan and Watson 2007). Lower levels of education and difficulties with language and literacy are common amongst (but not restricted to) people from some minority ethnic groups. This may also partly explain higher levels of stigma within certain communities.

Close relatives and friends of people with dementia sometimes experience courtesy stigma (or ‘stigma by association’) (Blum 1991, MacRae 1999, Werner and Heinik 2008). Courtesy stigma involves perceived loss of social status and devaluation of a person because of his or her association with someone who is stigmatised (Goffman 1963). This experience is not limited to specific ethnic groups but may be particularly relevant for groups which are less individualistic. In some communities, the way that dementia is perceived or understood may have implications for the family as well as for the individual, and this in turn may have a negative or positive effect on participation in research. For example, the belief that dementia is a mental disorder that runs in the family may have implications for the status and honour of the family and hence be relevant to marriage negotiations. Consequently, there may be a certain reluctance to take part in dementia research. However, in some minority ethnic communities and in certain religions, people with dementia maintain their status of ‘wise old person’ even in an advanced stage of dementia which engenders more openness about dementia. This differs from the dominant Western discourse which tends to idealise youth and results in older people losing their respected status because of ageing. Islam, for example, emphasises the need to respect all older people (Mohammed 2017).

It is important to bear in mind that many people with dementia already experience stigma in relation to other attributes. Some of these people may be reluctant to participate in research which focuses on that particular attribute (e.g. a comorbidity or personal characteristic) or in which they will be categorised on that basis. The means by which potential participants are identified and categorised, even if used with good intentions, can sometimes be misguided and stigmatising. On the other hand, it should not be assumed that reluctance to take part in PI or research is because of stigma or that people, even in communities where dementia is commonly perceived as a stigma, all consider it as such or would avoid participation for that reason.

Another issue related to diagnostic labels is that if people don’t identify with a particular label or relate to it, they might not feel that they are being targeted to participate in PI or research. Terminology is constantly evolving and there are cultural variations in the use of various terms. Some people are confused about the meanings of the terms ‘Alzheimer’s disease’ and ‘dementia’ as they are used inconsistently by some researchers and healthcare professionals who mix old and new conceptualisations (for more on this issue, see also page 82). In some counties, the term ‘dementia’ (as directly translated from English) is considered inappropriate if not offensive and therefore avoided (e.g. in Finland where the term memory disorders is mainly used). Consequently, people with dementia who might otherwise be interested in contributing to research may not recognise themselves as being eligible or not wish to participate in a study where they are referred to such a way. Moreover, terminology which reflects a strictly biomedical perspective of dementia does not correspond to some people’s own perspective and experience (Jongsma, Spaeth and Schicktanz 2017). This may have a negative impact on their participation in research.

Some terms used to describe other conditions or people from minority groups are also considered derogatory or demeaning. A term which was acceptable and the norm at one point in time (e.g. mental retardation) is sometimes replaced or largely falls into disuse but is still heard from time to time in some geographical areas and can be found in various publications (Nash, Hawkins, Kawchuk and Shea 2012). Lack of familiarity with a particular group of people may result in overlooking distinctions made by people in those groups that are important to their self-identity and how they wish to be considered. In the literature, this is also framed as hermeneutic injustice (Jongsma, Spaeth and Schicktanz 2017). For example, Young, Ferguson-Coleman and Keady (2018, p.1002) point out that Deaf people (Deaf with a capital D) “constitute a cultural–linguistic minority for whom to be Deaf is not a marker of deficit but rather a feature of individual and collective identity”. This is different from people who lose their hearing later in life. Uncertainty or ignorance about the correct terminology to use in order to be respectful towards potential participants may sometimes result in people feeling devalued and not wishing to contribute towards research. This is therefore one area in which the involvement of people from the targeted groups, perhaps with the involvement of gatekeepers, could be particularly helpful.


Recommendations for researchers

  • Be sensitive to the different possible ways that dementia is perceived when trying to recruit people with dementia for research.
  • Be ready to question your own assumptions about dementia and people with dementia.
  • Help people to avoid the stigma they may feel as a result of being seen in a memory disorders clinic by making it possible for data collection to occur in their own home (may also help with compliance and retention).
  • Pay attention to language and respect the linguistic preferences of potential participants.
  • Check how people would prefer to receive information (e.g. not in headed envelopes of an Alzheimer’s disease centre or on documents with letterheads which refer to mental health, dementia or memory clinics).
  • Check when it would be convenient for potential participants to be contacted or visited (e.g. at a particular time of the day or when a certain relative is not likely to be home).
  • Build up trust with marginalised communities by speaking or involving a co-researcher who speaks their language or shares other characteristics of the group.


Recommendations for research ethics committees

  • Try to ensure diversity on research ethics committee panels.
  • Assess whether methods and materials used to advertise research projects are such as to be accessible and respectful towards people from different socio-cultural backgrounds.
  • Ensure county-specific guidance and legislation has been followed to ensure capacity, but also recognise the valuable contribution to research by participants with advanced dementia.
  • Assess whether the spaces where the study is carried out and the research procedures are such as to avoid any stigmatisation of participants.
  • Seek clarification of additional support that may be required for data collection (e.g. the key worker of a person with intellectual disability). 


Recommendations for funders

  • Try to ensure diversity on funding committees.
  • Be sensitive to how symptoms or conditions are framed in calls for proposals.



For some people with dementia, a range of factors stand in the way of them hearing about research, weighing up the pros and cons of participating and potentially volunteering to participate. Examples include difficulties with language and literacy, lack of trust of the people proposing the research or of researchers, not feeling that the research is relevant to people from their communities, not wanting to be associated with studies about dementia because of the way that it is perceived and portrayed and the negative labelling of sub-groups of the population of people with dementia. These barriers are often a result of the way that researchers raise awareness about their studies, the means that they use to attract potential participants, personal racial/ethnic assumptions or bias and lack of sensitivity to perceptions of dementia as a stigma. As was mentioned earlier in relation to PI, informational materials may sometimes be culturally inappropriate and disseminated in places that do not correspond to the cultural diversity of people with dementia. This calls for greater sensitivity and awareness from researchers and emphasises the importance of ensuring diversity on research ethics committees and funding committees. 

The financial costs of involvement

Costs linked to being a research participant may put some people off participating, especially those from a lower socio-economic background. Such costs could, for example, include public transport, taxis or petrol and parking fees for trips to study centres, food, accommodation, minding fees for dependent relatives, loss of earnings, loss of holidays, and refreshments if the study visit takes place in the person’s home. Researchers seeking to involve people with difficulties with language, literacy and learning would need to budget for extra costs and allow for extra time for outreach efforts, interpretation, translation, bilingual experts, researchers with expertise in intellectual disability, adapted information materials, transport to people’s homes for home visits, additional transcription and translation costs for data analysis and costs linked to the development and validation of culturally sensitive data collection tools.

Some of these costs would be incurred at the recruitment stage and others would arise later but all such costs and the process for reimbursement should be made clear from the outset in order to avoid the risk of influencing decisions about the recruitment of a more diverse population in research. Researchers and funders need to take such financial costs into consideration when drawing up and approving budgets for research that is truly inclusive. Cost should not be used as an excuse for failure to involve a diverse population or exclude certain groups of people. Writing in the context of the involvement of people from minority ethnic groups in clinical trials, Hussain-Gambles concludes,  

“Economic considerations should not, however, be used as an excuse for the exclusion of minority people from clinical trials. Good science is expensive, and since the distribution of resources depends on priorities, it is a question of how high up the priority list is the problem of unequal representation in clinical trials”(Hussain-Gambles 2003, p.141).



Gatekeepers are people who formally or informally control researchers’ access to certain groups of participants. They may, for example, be respected members of minority ethnic groups, managers of care homes, legal guardians or informal carers. Gatekeeping can be empowering as well as restrictive with regard to the opportunities that people from various sub-groups of society have to participate in research and the possibilities that researchers have to reach them. It is therefore an important phenomenon to consider in relation to the recruitment of a diverse group of people with dementia in research. Taking the example of research involving the widely neglected group consisting of older homeless people with dementia, Manthorpe et al. (2019) found that staff in hostels for older homeless people were quite good at identifying hostel residents with memory difficulties, some of which may be linked to dementia. They were frequently in contact with them and were able to recognise some of the difficulties they were experiencing in carrying out everyday activities.

Gatekeeping may serve the following purposes, which could be considered in a positive or negative light, depending on one’s perspective:

  • To enable people from different sub-groups to hear about research opportunities, and understand the importance of a particular study for society and for themselves,
  • To enable researchers to access groups of people whom they would otherwise have difficulty reaching (e.g. because of language difficulties, trust issues, cultural and religious traditions governing interaction etc.),
  • To protect people perceived as vulnerable from exploitation or from research activities which may be detrimental to their wellbeing or that of their entourage,
  • To restrict access to potential participants in order to avoid extra workload or responsibilities (e.g. for health and social care professionals).
  • To facilitate or promote research that is likely to be beneficial to people from the communities or sub-groups of society to which the gatekeepers have access,

Gatekeepers have some degree of power to benefit people from the communities they represent by promoting particular research and facilitating the participating of people from those communities in such research. There are, nevertheless, a few challenges that need to be considered when seeking the support of gatekeepers.

When resulting in the restriction of researchers’ access to broader groups of people with dementia who might otherwise have been interested in participating in research, gatekeeping could sometimes be considered as paternalistic and even an abuse of power. It may, for example, deny people with dementia the right to decide for themselves whether or not to take part in research (Lepore et al. 2017), thus failing to respect their autonomy. Nordentoft and Kappel (2011), for example, describe organisational restrictions which make it difficult to carry out research involving vulnerable groups in and outside of hospitals in Denmark. For gatekeepers who do not necessarily know the potential participants personally, there is a risk of them making assumptions based on stereotypes and inadequate knowledge (e.g. about the needs, interests or perceived vulnerability of individuals from various sub-groups such as people in residential care, with advanced dementia, in the oldest old group or with intellectual disabilities). In some contexts (e.g. in many German nursing homes), it is mainly legal guardians who decide whether or not residents will take part in research and they may be hesitant to allow this because of concerns about the research being burdensome to residents (Palm et al. 2016). It may also be the case that they do not know the resident sufficiently well to make the decision about participation.

Doody (2018) describes how gatekeepers may, when trying to promote the inclusion of people with intellectual disabilities in research, actually block their inclusion with their ‘protective power’, which can be positive and nurturing when used appropriately (Witham, Beddow and Hagh 2015), but may also serve to silence and oppress potential participants. This represents an ethical dilemma in which gatekeepers may be caught between the desire to protect people from harm and the desire to respect and promote their autonomy. Doody (2018) suggests that it may be helpful for gatekeepers to have access to feedback given to researchers by the research ethics committees who review study proposals. 

It should also be borne in mind that there are different levels of gatekeeping. Researchers may sometimes need to get past several layers of gatekeepers to recruit older people with dementia for research (Hellström et al. 2007). There may, for example, be directors of care homes who act as initial gatekeepers and different members of staff who must respect the directors’ decisions but may have the power to block or hinder access to potential participants. There may also be legal guardians and informal carers. In the residential care setting, care staff are likely to be most affected by the involvement of people with dementia in research (Doody 2018) and may sometimes have personal reasons for refusing access.  

Witham, Beddow and Haigh (2015) describe the need to understand attitudes and perceptions surrounding dementia and ageing, which are common in some settings in which researchers seek to access various marginal groups to participate in research. This includes social constructions of ageing as dependency and vulnerability, the ‘positioning’ (Harré and van Langenhove 1999) of people with dementia as incompetent and examples of what Kitwood described as ‘malignant social psychology’, including infantilisation and objectification of people with dementia (Kitwood 1990, 1997) and of a general hesitancy and uncertainty in dealing with people seen as ‘other’ or having ‘special needs’. The infantilisation of people with dementia is a cultural cliché that is persistant and problematic. It silences people with dementia, thereby diminishing their autonomy and can be harmful in terms of not providing them with appropriate care and opportunities (e.g. to participate in research). It should therefore be avoided/challenged (Jongsma & Schweda 2018). Witham et al. (2015, p. 35) ask how some groups of people come to be positioned in such a way that they are not even asked in the first place and thus denied choice and autonomy. They suggest the “need to remain reflexive in situating the patient as ‘vulnerable’ and positioning them in such a way that choice and decision making become compromised”.

Researchers who are not members of the communities or groups from which they are seeking to recruit participants may not always understand the extent to which a particular key person represents the interests of the whole community or just a part of that community. Fisher warns against “the unwarranted assumption that opinions of minority scholars and community leaders reflect or override those of the less educated and more vulnerable community members who may be the target of investigation” (2009).

Some communities are very heterogeneous, hierarchical and/or with significant cultural, social or religious sub-divisions. According to Bhattacharyya and Benbow (2013), there are hierarchies and caste systems in some minority ethnic groups. The South Asian community in the Bradford area of the UK, for example, has been described as culturally and socially fragmented with divisions based on caste, status, gender and generational hierarchies. People with intellectual disabilities also form a very heterogeneous group made up of several sub-groups (Doody 2018). People may sometimes object to being placed in a broader or seemingly inappropriate group by researchers. Condon et al. (2019) report a study (not related to dementia) in which Show People[7] were offended at the suggestion that their views would be similar to those of Gypsies, Roma and Travellers and not to those of the general population. This suggests the need for sensitivity to inter group differences that are prominent in many groups, including that of people with dementia.

Once identified and involved in a particular study, gatekeepers may find that researchers are keen to approach them about other research. This could be interpreted positively as evidence of researchers’ interest in involving a more diverse set of people in research. Some gatekeepers may, however, find themselves inundated with requests, have difficulty managing the increased workload and have to reject some of them (Doody 2018). Having successfully negotiated access to a minority group, there may be a tendency for researchers to stick to that group because it has become ‘easy to reach’. This does not represent fairness or progress in reaching out to a wider group of people as other perspectives may still be excluded.

Finally, gatekeepers often play a vital role in recruiting people with dementia from minority groups and invest a lot of time and effort in this task. Researchers may wish to offer them a token of their gratitude. Alternatively, based on the concept of reciprocity, some gatekeepers may expect or request something from researchers (e.g. payment, a service or a contribution to a programme that would be beneficial to people from that group).  There are differences of opinion amongst researchers as to whether recognition of the reciprocal relationship with gatekeepers should be based on a personal gesture or token of gratitude or on a more formal arrangement. Irrespective of the approach adopted, it is important to ensure that anything offered in return for the support of gatekeepers does not have an adverse effect on the choice of research projects to which they lend their support. 

In addition to external gatekeepers, researchers need to bear in mind that they themselves might also be gatekeeping (Sharkey, Savulescu, Aranda and Schofield 2010). They themselves might have implicit inclusion and exclusion criteria that lead to them not approaching all eligible research participants (Tromp and van de Vathorst (2015). For example, they might have a prejudiced belief that someone would not want to participate, think that participation would be too burdensome or think that someone would be non-compliant in the research. Although this practice of individual gatekeeping by researchers is understandable, it is not desirable (Tromp and van de Vathorst 2015), because:  

-          It denies personal choice,

-          It may deny a person a possible benefit resulting from participation in the research,

-          It may reduce inclusion rates,

-          It can cause selective inclusion (i.e. bias) and therefore endangers generalisability (in the case of quantitative research),

-          It can result in an unfair distribution of burden, risk and benefit among people with dementia.

Different agendas/priorities

Research into issues relating to various minority groups often reflects topics chosen by or identified as being problematic by people who are not members of those groups and often for economic or political reasons (Tyack 1995). The actual topics selected may also be influenced by researchers’ lack of awareness of the priorities of minority groups and partly due to hidden agendas of researchers and research institutes (e.g. linked to career advancement, the desire for recognition from peers and society, competition for funds and scientific curiosity). Increasing the involvement of a diverse set of people in research not only requires equal opportunities to participate in research but also equal opportunities to influence the research agenda. This point was made earlier with the regard to the importance of PI at the earliest possible stage, namely when exploring different possible research topics, prior to seeking specific funding. Involving people from minority groups in research should also extend to the choice of research topic.

There are different forms of dementia, the most common one being Alzheimer’s dementia, followed by vascular dementia and then dementia with Lewy bodies, and mixed dementia is also common (Livingston et al. 2017). However, research into dementia is not closely aligned with the prevalence of these different forms. This means that some forms of dementia, such as Lewy body dementia, familial dementia and vascular dementia, are under-researched (Jongsma et al. 2016). Consequently, people with dementia do not all have the opportunity to participate in research related to their specific condition.  

“As I have frontotemporal dementia I am not eligible to participate in much of the research that is going on. I realise there is not as much being done on FTD but I am often even excluded from studies on the topic of living with dementia. I have dementia so why is a lot of this research just about people with Alzheimer’s?” (Petri, person with dementia, Finland).

Relevant questions and hypotheses must be addressed by researchers. It is therefore important to consult members of different minority groups, such as minority ethnic communities, people with intellectual disabilities or members of the LGBTI communities, so as to identify and respond to their interests and concerns (Konkel 2015). Promoting and respecting the values and interests of different communities and protecting communities from harm is what Weijer and Emanuel (2000) described as a matter of ‘respect for communities’.


Recommendations for researchers

  • Create a realistic study population with adequate and appropriate inclusion and exclusion criteria.
  • Budget for the cost of involving a diverse set of people with dementia as participants in research when applying for research funds.
  • Ensure that the time required for informed consent is reflected in project timelines.
  • Discuss with people from the relevant communities about possible participation, the kind of support they might need and the financial cost of this (e.g. travel, interpretation, dietary requirements, additional cost for one or more support people etc.).
  • Invest in a diverse research team. The diversity within a research team enhances sensitivity to possible invisible issues related to the gatekeeper/research team relationship and can help avoid possible conflicts and mis-use of this relationship either by the gate keeper or by the team, and thus enhance access to communities.


Recommendations for research ethics committees

  • Be aware of too restrictive inclusion and exclusion criteria, which cause an irrelevant and impossible research population and diminished generalisability.
  • Ask researchers to describe the measures that they envisage to involve a diverse set of people with dementia in research, including ways to approach possible gatekeepers.
  • Be willing to consider necessary adaptations to existing procedures and open to novel approaches which may be necessary to achieve inclusive research.
  • Assess (in relation to particular payment for particular research) whether reimbursement for participation in research would encourage the participation of people with dementia with limited financial resources and how ethically justifiable this would be.
  • Ensure that anything offered to gatekeepers in return for their support is not such as to influence their choice to support research projects.


 Recommendations for funders

  • Involve relevant individuals and communities in decisions related to funding.



It costs money to attract and involve a diverse set of people with dementia in research. Research funders need to bear this in mind and researchers need to ensure that they consider all the potential costs involved and draw up a comprehensive budget. People with dementia may be less enthusiastic about participating in research if they don’t feel that it is relevant to them (e.g. to people with their type of dementia, to people from their communities or to people who have things in common with them). Gatekeepers often play an important role in enabling researchers to communicate the importance of their study to specific groups of people with dementia which the researchers might otherwise find it difficult to reach or engage with. They may also provide valuable insight into what is important to certain groups and even negotiate something for members of a group in return for the personal investment of some of its members. However, gatekeepers may at times hinder access to certain groups of people as a result of being over-protective and may not always reflect the interests of everyone in what is believed to be the group they represent. Researchers themselves may also gate keep in a way by ruling some people out of participation on the basis of assumptions about what they would or would not want or be likely to do.  

Sampling issues

In the previous section, we explored some of the key issues related to the involvement of a wide range of participants in research, with a focus on practical issues, cultural and societal challenges and ethical considerations. In this section, we reflect on issues which are closely related to the methodological requirements of research sampling, although these also have practical, cultural and ethical implications.

Sampling bias and failure to capture diverse perspectives and experience

As it is rarely possible, in terms of practicalities, cost and time, to include all possible cases in a study, researchers typically select a sample of people to participate in their research. There are several different approaches to sampling in research, both for quantitative and qualitative approaches. Both approaches, including the many variations of each, emphasise the need to involve a broad cross-section of people. This is essential for truly inclusive research but despite concerted efforts to involve people with dementia from particular sub-groups of the population, for example, it often does not happen.

One of the difficulties when seeking to identify people from specific minority groups within an overall population for research (e.g. to try to ensure that the sample is more inclusive) is that details of the characteristics needed are not always readily available. Indeed, in many cases, it would be considered unethical or even illegal to record various details about people. For some characteristics, such as ethnicity or having impairments, it might be possible to ask if participants would be willing to provide such information (i.e. on a voluntary basis). This then poses the problem of providing pre-determined options, including ‘other’ or ‘non-stated’ or classifying open responses afterwards. Either way, this imposes a classification by researchers which does not necessarily reflect real-world characteristics as experienced by or meaningful to the people concerned. For example, people who appear (to others) to belong to the same ethnic group may describe their ethnicity in a number of different ways. In the United Kingdom, for example, some people might describe themselves as BAME or BME or alternatively as Black, Afro-Caribbean, South Asian, Pakistani or Sikh, with many adding ‘British’ (e.g. Black British).

In quantitative research, good quality research (which is a pre-requisite for ethical research) relies on appropriate sampling strategies to avoid selection bias. According to LaMorfe (2016),  

“Selection bias can result when the selection of subjects into a study or their likelihood of being retained in the study leads to a result that is different from what you would have gotten if you had enrolled the entire target population.”

Selection bias would result in findings obtained from an over-represented selection of people from a particular group being generalised not only to other similar people from that group but also to the underrepresented group of people (Simundic 2013). A classic example of this would be the generalisation of the findings of a study only involving White ethnic groups to people from Asian or Black ethnic groups, possibly leading to the development of culturally inappropriate services. Biased sampling results in inaccurate, misleading findings, which are not truly generalisable and the abuse of participants’ time and effort, as well as a waste of funding.

Many quantitative researchers would argue that samples should ideally be random to help ensure that they are statistically representative, thus permitting generalisation from the sample to the larger population. Contrary to popular usage of the term, ‘random’ does not imply that the selection of participants was haphazard but that strict measures were applied to try to ensure that the sample was not systematically biased (see below). Non-probability (i.e. non-randomised) sampling techniques are often considered inferior to probability (i.e. randomised) sampling techniques even though random sampling cannot guarantee the representativeness of a particular sample. This issue is discussed further in the sub-section on sampling hierarchies.

Qualitative research typically involves smaller numbers of participants, often purposefully selected to obtain rich, in-depth and meaningful information which illuminates the research question whilst capturing the diversity of experience. The terminology may differ in that the findings of qualitative research are not derived from statistical analysis and not intended to be generalisable but the goal of ensuring that the sample is sufficient and suited to answer the hypothesis or research question is similar.

Self-selection may occur in some quantitative and qualitative studies depending on the approaches adopted. For example, an Internet-based survey to measure or explore mobility issues of people with dementia would be likely to attract people who are able to use computers, perhaps with support, and have access to one, and exclude those who aren’t and have not. Self-selection may also be influenced by the topic and by cultural differences with regard to readiness to discuss the issue with researchers or people who are not from the community. The opposite may also be the case in that people may sometimes be reluctant to discuss or provide information about a particular issue to someone from the same community (e.g. for fear of gossip or rumours, or to protect the honour and dignity of the family or individual). For quantitative studies, this would risk participants who choose to participate in a study not being equivalent (in terms of the research criteria) to the people who opted out. In qualitative studies, it would risk the people participating having very similar perspectives or experiences and the full diversity of perspectives and experience.

Sample size

Obtaining a sufficient number of people with dementia to participate in research can often be difficult. For research involving quantitative analysis, research ethics committees and funders may be justified in asking for clarification of the number involved (e.g. based on the population size, margin of error, confidence level and standard of deviation) in order to check that this would offer sufficient statistical power, and avoid wrongly failing to reject the null hypothesis. In quantitative research, if a sample size were too small, it could be assumed that the efforts of the research participants, and in some cases their discomfort or burden, were in vain because the goals of the study, namely to obtain a valid response to the research question which can be generalised to the broader population, were not achieved. It could also be argued that financial and time resources invested in the study

“were squandered since ultimately it will contribute absolutely nothing to improve clinical practice or quality of life. The situation becomes even worse if the research involves public funding: A total waste of taxpayer money” (Faber and Fonseca 2014, p.28).

Faber and Fonseca (2014) point out that sample sizes which are larger than necessary are also unethical because this would indicate that some people were needlessly exposed to risk or burden, or that their time, efforts and public funds were wasted. Larger-than needed sample sizes are rarely the problem though. When striving to include people from minority groups in research, attracting a sufficient number of participants can be a real problem and in quantitative research, simple random sampling is often not the best approach. Moreover, actively seeking to include a sufficient number of people from various sub-groups in representational proportions necessitates the involvement of a much larger number of people, with the financial and practical implications that this may incur (Hussain-Gambles 2003). The need to be able to identify people from a more diverse background can also be problematic for qualitative research.

In qualitative research, relatively small numbers of participants are quite common. The principle of qualitative research is that researchers seek to obtain data, often regarding quite complex social phenomena, with a focus on meaningfulness, the diversity of experience and rich descriptions. Researchers are nevertheless often expected by research ethics committees and funders to give some indication of the possible number of participants to be involved in a study even though in many cases the number is determined by the saturation of the data obtained. Once the point of saturation has been reached, further collection of data would be unethical. As with excessive numbers involved in quantitative research, continuing to recruit participants and to collect data beyond the point of saturation would amount to a misuse of human and financial resources, with people sometimes undergoing burdensome tests or questioning for no benefit at all. Asking researchers to justify the number of participants in advance may be based on an inadequate understanding of certain qualitative research approaches, may put pressure on researchers to involve more people than necessary and may be interpreted as questioning the validity of qualitative research by trying to apply standards and criteria for quantitative research.

As a means to involve people with more advanced dementia in qualitative research, it has been suggested that larger samples than usual may be required in order to obtain sufficiently rich data. Reporting on the experience of a few studies, Beuscher and Grando (2009) state that it may also be helpful to conduct multiple interviews, to have multiple readings of the transcripts and to triangulate the findings of the interviews with observational and field notes.

Recommendations for researchers

  • Describe in the research protocol the rationale for the sampling strategy and how this is conducive to the promotion of inclusive research involving people with dementia.


Recommendations for research ethics committees

  • Consider the need for inclusive research when evaluating the appropriateness of the sampling strategy in the context of judging the scientific validity of research projects.
  • Ensure that the assessment of sampling strategies is appropriate to the research design and proposed methodology.

Recommendations for funders

  • Recognise that small sample sizes do not necessarily mean a less effective study provided that there is methodological rigour and reliability in approach.


Because of the time, effort and money it would cost to include all possible cases in a study, researchers typically select a sample of people to participate in their research. The risk of selection bias is particularly important in quantitative research and may result in the results of an over-represented selection of people wrongly being generalised. Random sampling is probably not the best approach to gain a representative sample of people with dementia living in multicultural and diverse societies. In qualitative research, smaller numbers of participants are often purposefully selected to obtain rich, in-depth and meaningful information which captures the diversity of experience. Self-selection may occur in some quantitative and qualitative studies and may contribute towards an unrepresentative and characteristically restricted sample. Obtaining the right number of participants can often be difficult. Too small a sample in quantitative research may result in inconclusive results that lack sufficient power to be generalisable, and too many people in qualitative research (sometimes based on a lack of understanding of the principles of qualitative sampling) may lead to an abuse of people’s time and effort. In both cases, inappropriate sampling leads to a waste of financial and human resources which could have been put to better use.


Inclusion and exclusion criteria

Many of the factors which represent challenges for inclusive research with people with dementia are quite subtle and not intended to prevent people from participating in research. Inclusion and exclusion criteria for research are part of standard research practice for the conduct of good quality scientific research. They are also not designed to be discriminatory. The aim of inclusion criteria is to determine the key features of the target population that will enable researchers to answer their research question or hypothesis. The aim of exclusion criteria is to define features of the target population (who meet the inclusion criteria) which might interfere with the success of the study (e.g. increase the likelihood of them being lost in follow up, missing study visits, providing inaccurate data or resulting in an unfavourable outcome) or put the participants at risk (Patino and Ferreira 2018). A key question for researchers and research ethics committees to answer is whether the rationale for inclusion and exclusion for particular studies is valid (i.e. is there a valid justification for them).

Care nevertheless needs to be taken when developing inclusion and exclusion criteria to ensure that people are not unnecessarily deprived of the opportunity to take part and that they do not jeopardise the external validity of the study. For example, classing every possible characteristic which might lead to an unfavourable outcome (e.g. in a clinical trial) as an exclusion criterion would mean excluding a huge number of people who might like to be involved and the findings having little relevance or generalisability to the real-world population. Inclusion and exclusion criteria therefore have a key role to play in the ethical conduct of research but must be carefully scrutinised to ensure that they are applied in an ethical manner. In this section, we look at several characteristics commonly used as inclusion or exclusion criteria.


Age is usually classed as an inclusion criterion in that it describes the age range of the people the researchers would like to include in their study. There are some good examples of research involving older people in research such as the Alzheimer’s Disease Neuroimaging Initiative (ADNI) (Weiner 2010) and the “nun study” into healthy ageing and dementia involving 678 Catholic sisters aged between 75 and 107 (Snowdon 2003). However, upper age limits exclude many people with dementia from research who could equally contribute towards answering a particular research question or hypothesis. Younger people with dementia may also find themselves wrongly excluded from various types of dementia research, including clinical trials (e.g. because they do not correspond to stereotypes of people with dementia). If the reasons for their exclusion cannot be justified, it should be considered as discriminatory.

F“There is usually a minimum and maximum age limit for participating in dementia research. I don’t really understand why this is necessary. I have dementia, the study is about people with dementia and yet I am excluded. I find this frustrating and unfair.” (Miha, person with dementia, Czech Republic)

It is a fundamental principle and a legal and ethical requirement of drug development and treatment in Europe that drugs be tested on people with the same characteristics as those who will eventually use them (if proven effective and safe). In 1994, the European Medicines Agency (1994) stated that patients entering clinical trials should be reasonably representative of the population that will be later treated by the drug. Yet earlier, in 1989, the American Food and Drugs Administration commented that there is no good basis for the exclusion of patients on the basis of advanced age alone.

According to Beswick et al. (2008), the literature on the age of older people in clinical trials for Alzheimer’s disease is limited but for clinical trials to be truly representative, they should include a large percentage of people between the age of 75 and 90. In a review of 109 clinical trials of therapeutic interventions in adults assessing morbidity or mortality, it was revealed that 20% of studies automatically excluded patients above a certain age (Zulman et al. 2011). Clinical trials therefore often involve participants who are not representative of those for whom the medication will most likely to be used if the trial is successful (Heait et al. 2002, Jongsma, van Bruchem-Visser, van de Vathorst and Mattace-Raso 2016). Many clinical drug trials still have inclusion criteria which prevent older people with dementia from participating even though the largest proportion of people with dementia are aged 80 and over (Brayne and Davis 2012). This means that many older people with dementia are taking medication, which has not been proven safe or effective for them because similar-aged people were excluded from the trials, and that their views and experience are frequently not sufficiently taken into consideration in dementia research. There may be differences between younger and older people with regard to:

  • adverse reactions,
  • pharmacokinetics and pharmacodynamics (biological effects in this case linked to age e.g. affecting metabolism and sensitivity to medication),
  • comorbidities (having additional acute illnesses and chronic conditions – such as dementia),
  • polypharmacy (increased use of medication/older people taking several different drugs each day) (Orwig et al. 2011).

These age-related differences, sometimes expressed as exclusion criteria, may contribute towards the exclusion of older people from clinical trials. Another possible reason for excluding older people from clinical trials is that their inclusion might dilute the active treatment effect, thereby resulting in results which are not statistically significant (Cherubini et al. 2010). This has obvious implications for the development of marketable drugs but also with regard to fairness to the hundreds or thousands of participants who dedicate their time and energy to a study which may be inconclusive. On the other hand, the characteristics of the majority of older people who will eventually take the drugs are different to those of the people who took part in the trial. 

Sox and Greenfield (2009) criticise randomised clinical efficacy trials for attempting to create near ideal circumstances in order to determine whether the intervention could possibly work which do not reflect the conditions in real clinical practice. The EFGCP GMWP[8] (Diener et al. 2013) advised in their  guidance synthesis on medical research for and with older people in Europe that participants in clinical trials, should be reasonably representative of the population that will be later treated by the drug (and hence to include more participants in the older age range) and to avoid the direct extrapolation to older patients of efficacy/safety data and of the drug benefit/risk profile observed in younger adults, as this may lead to safety issues and iatrogenic disorders.

Possible barriers to the involvement of older people with dementia in research in general include, for example, physical and cognitive impairments, difficulties with mobility, lack of transport, a lower threshold for burden, distrust, higher costs linked to the recruitment and retention of participants, polypharmacy and various co-morbidities. However, such factors do not apply to all older people and are not limited to older people. In addition, age is often associated with a more advanced stage of dementia. Hubbard et al. (2003) caution against categorising people into different stages, suggesting rather the need to focus on specific impairments and to use methods suited to the individual, irrespective of the duration or severity of dementia.

It is important to be aware of possible factors which may affect the ability or willingness of older people to participate in research and of various challenges that researchers face when seeking to involve older people in research. However, these factors and challenges should not be unquestioningly seen as legitimate grounds for exclusion. With the necessary time, effort and financial investment, many of the challenges to including older people with dementia in research can often be overcome by proportionate measures. Adherence to recruitment methods and study protocols, which do not take into account the specific needs of older people, with or without dementia, as well as failure to build in additional costs into the design of studies for the recruitment of older people, could be considered as examples of ageism and structural discrimination. This is also not in keeping with the principles of dignity, respect and social inclusion.  With regard to the issue of age discrimination in clinical trials, Cherubini et al. (2010) concludes:

“Older people who are more characteristic of those seen in clinical practice must be included in clinical trials to achieve the goal of safe and effective drug therapy for this growing patient population. (…/…) The design of clinical trials that include older adults more typical of those seen in clinical practice should be reconsidered for inclusion of complex older adults with multiple comorbidities and polypharmacy, addressing concerns about the feasibility of multiple measurements; the risk:benefit ratio of the proposed interventions; the determination of adequate sample sizes, taking into account higher attrition and heterogeneity; the complexity of multicenter trials; and finally, the higher costs of such clinical trials.”

Place of residence

Trial delivery sites for clinical trials tend to be located in and around large towns and cities. This is practical when trying to attract a sufficient number of participants but results in the exclusion of many people who live in rural areas or small islands, especially people with dementia who may have difficulty planning trips, finding their way around unfamiliar places and who need to be accompanied by another person (who in turn may have other obligations and commitments and therefore not have enough free time to assist).

F"I am Idalina and have dementia. I live in Madeira and although this is Portuguese, there is not much research being done here. Living on an island means that people are excluded from a lot of research because studies are all on the mainland. I would be more than willing to travel to mainland Portugal to participate in a study. But I don’t really get to hear about possible research opportunities over there and I’m not sure I would be accepted anyway, and that is unethical in the sense of it being discriminatory.” (Idalina, person with dementia, Madeira)

People with dementia living in nursing homes are sometimes explicitly excluded from studies. In many studies, recruitment simply targets people living independently or no mention is made about living situation as an eligibility criterion (Jongsma et al. 2016). In such cases, people in nursing homes may be excluded because of structural discrimination resulting in unequal access to dementia research and inadequate or unproven care for those living in nursing homes.

People serving prison sentences and homeless people are typically excluded from research (unless a study is specifically about people in those groups). Manthorpe et al. (2019) suggest that older homeless people with memory problems, including those associated with dementia, are often reluctant to engage with services and are difficult to assess and help because of persistent heavy drinking, behaviour that other people find challenging (perhaps in some cases linked to substance abuse), self-neglect and physical health problems. Little is known about the problems and needs of older homeless people with memory problems and it is likely that some of the challenges affecting their help seeking would also be challenging for researchers considering their involvement in studies. There may be issues which many researchers find too challenging and it may be that these are often simply forgotten groups.

In some group living situations, dementia is quite common but seldom mentioned (e.g. in some assisted living or nursing homes for older people). Dementia is harder to diagnose in people with an intellectual disability due to lack of appropriate diagnostic tools. Staff and families may therefore be reluctant for residents to take part in dementia research for fear of raising concerns about dementia amongst other residents and act as gatekeepers, thus denying them opportunities that other people have. People living in various types of residential homes may also be excluded from research because they don’t get to hear about studies, they may have mobility or transport problems or participation might be considered by others as interfering too much with the routines of the home. If formal diagnosis, including imaging, is required for recruitment, older people living in nursing homes may sometimes be excluded because they are considered too frail to transport or because nursing homes are not affiliated with a clinic (Palm et al. 2016).

Having a study partner

In most clinical trials and sometimes also in non-pharmaceutical research involving people with dementia, it is a necessary requirement to have a study partner[9]. Study partners are typically chosen by research participants. Usually, they attend study visits with the person with dementia and may be asked to complete questionnaires and provide information. They may also be asked to give independent reports about the participant’s mental status. They are often a spouse/partner, friend or sibling as they need to know the person quite well and be aware of his or her behaviour or condition. Study partners are generally considered as research participants and thus subject to research ethics provisions. This differs from people who simply have a supportive role (e.g. a key worker for a person with an intellectual disability or a spouse who may remind the participant about meetings or accompany him/her travelling) who are not participants in the study themselves. Study partners should also not be considered as proxy-consenters for people who are unable to consent to participation. These two roles are sometimes combined in one person but are fundamentally different. A study partner is a research participant who acts as an informant, whereas a proxy-consenter has a formal role in which he or she consents to research participation on behalf of a person who is unable to consent. 

The reasons for having study partners include concerns about the condition of the person with dementia deteriorating over the course of the study, about the person with dementia needing assistance to attend study visits and comply with the research requirements and, as stated above, to provide additional information (an informed observer’s perspective) about his or her behaviour, progression or mental status during the course of the research.

Requiring a study partner, irrespective of a person’s actual needs, involves making assumptions about people with dementia (e.g. that all participants need assistance, that their perspective or the information they provide is not reliable and that their condition will deteriorate in the course of the study) and could therefore serve to reinforce negative stereotypes about people with dementia and contribute towards stigma. On the other hand, some studies even require a study partner for healthy controls but this is less common.  

Some people with dementia may have difficulty finding a trusted person to act as their study partner, particularly for research stretching over several years (e.g. they may live alone[10], have no relatives who live close by or do not know someone with the necessary skills, time or interest).

“I would normally agree to take part in lots of different studies but some state you have to have a study partner and even if they don’t, I know I would need support. My wife would gladly help me but she has other commitments. It’s a shame but that’s the way it is.” (Geert, person with dementia, Belgium)  

People who live alone and people whose potential study partners are either opposed to their participation in research or have commitments which make it difficult for them to take on this role may, for example, be excluded.

“Being a study partner necessitates the commitment of time, effort, and insight into the research participant’s cognition and function. In short, being a study partner is work. If an older adult doesn’t have someone who’s able or willing to do this work with her, then the study partner requirement is a barrier to her research participation” (Largent, Karlawish and Grill 2018, p-1-2)

Largent et al. (2018) suggest the need to look for ways, within the limits of methodological and scientific rigour, to make the participation of research partners more palatable (e.g. having study visits outside office hours, offering home visits or transport to the study centre, covering parking costs and allowing remote participation by Skype to complete some study instruments).

There may also be reasons not to involve a partner for fear of drawing attention to a relationship which either the person with dementia or the potential study partner would prefer to remain private. Many people with dementia are to some extent dependent on their relatives and close friends for support or care. According to Davis (2017), the relationship between research participants with dementia and their study partners may sometimes be harmed as a result of their joint participation in research. For example, participants may feel dependent on their study partners for their continued participation in the study, have the unpleasant feeling of being observed or reported on or start to question whether certain interactions are authentic or scrutiny prior to reporting to researchers.

The requirement to have a study partner may reduce the number of eligible research participants and skew the research population. It therefore needs to be considered whether a study partner is really necessary, whether the function they serve could be managed in another way and if not, to look for ways to remove obstacles to their participation. 

Literacy levels and language difficulties

Difficulties with language and literacy, which have already been discussed with regard to PI, may also prevent some people (e.g. from ethnic minority groups or with intellectual disabilities) from contributing towards research as research participants. Participant information sheets, informed consent forms and various research tools such as questionnaires, rating scales and computerised texts, are all key documents which must be accessible and suitable for all research participants. Some documents, such as informed consent forms, are extremely long and complicated. If the level of language is too high or complex and if various documents and research instruments are not culturally appropriate (i.e. have perhaps been translated but not validated on people with similar characteristics as the intended research participants), this will exclude some groups of people, provide ambiguous data (because responses might not correspond to the intended question without the researcher necessarily realising this) and/or involve people as research participants without them having given fully informed consent. Some people (with and without dementia) may even sign without having read the document at all (e.g. due to its complexity and length). Plous and Holm (2015) also highlight the problem of ‘routinisation’ (albeit in relation to online consent) whereby few people read the information provided and often give consent as an act of routine (a kind of formality).  They suggest that ‘routinisation’ threatens the protection of personal autonomy as it involves people not taking the opportunity to reflect on the implications of providing or refusing consent. 

Necessary measures are likely to incur extra time and costs and it would be unacceptable, as mentioned earlier, to use this as grounds to exclude some people with dementia from participating in research unless the measures needed would be clearly disproportionate. In multicultural societies, minority ethnic groups in a particular geographical area may speak over 100 different languages and dialects, and some people may not be literate in their own mother tongue. Including a diverse group of people with dementia in research necessitates the availability of appropriate research instruments to use once recruited. It would be impossible to translate research materials into every possible language. Moreover, materials would ideally have to be translated, back-translated and properly validated. This would sometimes be a study in itself. Certain well-known tests may already exist in other languages and already be validated but may be difficult or costly to obtain. Sometimes, certain minority ethnic groups have a common language, which can be used to address these groups and involve them in research.

Level of education

In some studies, a certain level of education is required (e.g. expressed as a specified number of years’ schooling or as having completed high school or college). In cases where this requirement is not directly linked to the research question or hypothesis, this could be construed as discrimination as it serves to exclude people who do not have that level of education. Whereas some people from majority ethnic groups, especially of the older generations, may have had little or no formal education, this is more common amongst some minority ethnic groups. Many people with intellectual disabilities would also find themselves excluded from research on the grounds of lower levels of educational attainment. 

It might be argued that a specific level of education is needed to be able to comply with the requirements of the research, such as understanding materials and completing tests. However, a lower number of years’ schooling or the absence of educational diplomas does not necessarily mean that a person lacks the ability to participate. Moreover, if research is to be rendered more inclusive, the onus should be on researchers to make their materials and instruments suitable/appropriate for the   population they aim to study (bearing in mind that a target population of ‘people with dementia’ should ideally include a wide range of people, some of whom will have low levels of education, literacy and language abilities etc.).


Provisions can often be made to ensure that people with dementia who have difficulties with mobility or transport can attend study visits. This could, for example, involve having some or all of the study visits in the person’s home or organising and paying for a taxi. Possible solutions to promote the inclusion of people with mobility problems would be likely to involve additional costs and time. These should be budgeted into the study. In addition, as people may be inclined to exclude themselves based on the assumption that they would be unable to get to study sites, measures to promote their inclusion should be stated upfront in any information about the study.


Stage/severity of dementia

Unless there are strict scientific reasons, participation in research should not be limited to people with very mild dementia, but this is often the case. Sometimes, there are clear eligibility criteria which rule out the participation of people with advanced dementia but sometimes exclusion criteria are indirect (based on conditions such as living independently or having a certain level of capacity) or ambiguous (based on “investigator’s uncertainty about willingness, ability, or medical status of the patient to comply with the protocol requirements” - Jongsma et al. 2016, p. 5).

The experiences and contributions of people with more advanced dementia to research are important. Their unjustifiable exclusion from research (i.e. based on a label and associated assumptions rather than on whether they comply with justifiable inclusion and exclusion criteria) contributes towards the lack of an evidence base for the treatment of people with more advanced dementia. Exclusion also results in a gap in understanding of the aetiology of the underlying illnesses. Furthermore, it deprives a sub-section of people with dementia of the right to contribute towards research and to reap any possible benefits linked to participation. There may be methodological and ethical challenges linked to involving people with more advanced dementia in research, especially if people also have some behavioural and psychological symptoms of dementia. However, such challenges can often be addressed successfully, thereby promoting more inclusive research. They should not simply be considered as a justification for exclusion.


As pointed out by Tinetti and Studenski (2011), there are an infinite number of combinations of diseases and treatments which makes the identification of a truly representative population very difficult. Comparative effectiveness research involves the comparison of different available treatments or examines their impact on particular sub-populations. This can be meaningful to doctors when trying to make a decision about or with a specific patient, particularly when treatments which are effective for one condition might exacerbate another condition in the same patient (Institute of Medicine of the National Academies 2009, Tinetti and Studenski 2011). 

In a review of 103 dementia research protocols in the Netherlands covering drug trials, intervention studies and observational studies, 54% were found to exclude people with a psychiatric disorder and 56% people with a neurological disorder. 22% mentioned visual or hearing impairments, 38% medication use and 54% somatic comorbidities as exclusion criteria (Jongsma, van Bruchem-Visser, van de Vathorst and Raso 2016). In 9% of the studies, all five exclusion criteria were mentioned. Jongsma et al. (2016) point out that the average person with dementia has several comorbidities and that excluding people with comorbidities limits the external validity of studies and does not represent the full spectrum of people with dementia typically seen in clinical practice. They argue in favour of conducting pragmatic (i.e. real life) intervention studies involving a large number of people and a limited number of eligibility criteria, with the aim of obtaining more applicable and generalisable findings.

It is important not only to involve people with dementia in research but to involve them in research which results in knowledge that is meaningful to them. People with dementia who have comorbidities may have different priorities than researchers. In a review of over 100 studies into conditions which are common in older people, only 27% mentioned outcomes which were of relevance to the older people such as quality of life or functional status (Covinsky 2011). Whereas researchers may place their emphasis on disease specific outcomes, older people with multiple conditions may prioritise outcomes of a more individual or personal nature. Jongsma et al. (2016) also mention the importance of including patient-centred outcomes in pragmatic intervention studies.

Specific sub-groups of the population

Sampling strategies must ensure that the people who take part in research are the ones whose data will enable researchers to answer their hypotheses/research questions. Difficulties arise when definitions and conceptualisations of some groups of people are inconsistent, ambiguous or too broad. This may also result in some people being excluded from research, and conclusions being drawn or assumptions made about people from certain minority groups that are incorrect (i.e. over generalising and overlooking important contributions from excluded members of those groups).

Different definitions of the term ‘migrant’, for example, may lead to some people having a reduced chance of being involved in a study. It is generally agreed that a migrant is a person who has moved away from his/her country to take up residence in another country, temporarily or permanently, for one of more of a range of reasons. However, there is no universally accepted definition of ‘migrant’ (International Organization for Migration/IOM 2019, United Nations 2019). Precise definitions may vary considerably with some requiring a person to have crossed an international border, others not; some requiring a minimum or maximum period of time in the host country, others not. One could also ask at what point does a person cease to be classified as a migrant (after 10 or 20 years, never?). Sometimes a distinction is made between migrants and expatriates (expats). For many people, the latter[11] are not considered as migrants. This distinction most probably reflects negative stereotypes and value judgements about different groups of people and may also affect who volunteers to participate in studies.

Some terms which have different meanings are used interchangeably. This is the case for the terms ‘migrant’ and ‘person from a minority ethnic group’. Many people from minority ethnic groups are not migrants. Some were born and grew up in the same country where they currently live and have no history of migration. Depending on the research question or hypothesis, if the two terms are used interchangeably, some people would be wrongly included or excluded and this would have an impact on the findings.

Terms used to define various groups of people who may at some point have dementia (i.e. people with intellectual disabilities, from BAME communities, with sensory impairments or from LGBTI communities) may change over time and differ with regard to whether and if so, how they are used across Europe. The term ‘homosexual’ for example, has largely fallen out of use since the 1970s, having been replaced by the term ‘gay’. It is increasingly considered to be an outdated medicalised term with negative connotations (Shaw et al. 2019).  However, it is still used in some countries and in some circles and can be found in books, magazine articles, sermons and everyday conversations. The use of the term would nevertheless be off-putting to many gay people who might find it disempowering.  The terms used may also result in some people being eligible for recruitment, others being excluded and the results of different studies being difficult to compare. The following extract describes issues related to the use of terms for Gypsies, Roma and Travellers:

“Amongst the range of people referred to as ‘Gypsy/Travellers’ are English, Welsh and Scottish Gypsies, Irish Travellers, New Age Travellers, Boat People and Show People, as well as Roma from a variety of central and eastern European countries. (…./…). In continental Europe, Roma people are no longer officially described as Gypsies as the word has come to be associated with racial abuse. Many different groups are described as Roma (e.g. Manouches, Ashkali, Sinti and Boyash people)” (Condon et al. 2019, p.2).

Condon et al. (2019) point out that membership of this broad category is fluid with people marrying into it, leaving it and not wishing to be associated with it if asked for official purposes. Moreover, the official UK census of 2011 only included Gypsies and Irish Travellers (combined as a single ethnic group).

For some researchers, there may also be uncertainty and a blurring of the distinction between terminology linked to intellectual disability (the most common form being Down’s Syndrome), developmental disorders (such as autism which may or may not include an intellectual disability) and terms such as learning difficulty and ‘mental retardation’. There are some cultural variations in the use of some of the terminology surrounding intellectual and developmental disabilities and differences in use between healthcare professionals and lay people (Nash, Hawkins, Kawchuk and Shea 2012) and this could lead to confusion and ambiguity in relation to recruitment.     

Some studies recruit participants from a fairly broad population, such as people with ‘cognitive impairment’, ‘memory problems’ or ‘cognitive difficulties’. This typically occurs in settings in which the specific characteristics of the population are not known (e.g. in residential care homes where many people with dementia do not have a formal diagnosis) or where a clear distinction is not necessary as not relevant to the research question or hypothesis (Palm et al. 2016). Such broad categories, when appropriate, facilitate recruitment, avoid researchers embarking on diagnoses which would otherwise not have been made and avoid people being informed that they have dementia solely for the purpose of the study (see next sub-section). However, terms should not be used as if they were interchangeable if this is not the case.

The inconsistent or incorrect use of terminology surrounding Alzheimer’s disease may result in some people who have been recruited to studies wrongly assuming that they have dementia. Originally, Alzheimer’s disease was more or less synonymous with Alzheimer’s dementia. Recent conceptualisations of Alzheimer’s disease, developed by the IWG and Dubois et al., propose a definition based on Alzheimer’s disease pathology, extending along a continuum from a pre-clinical state, through a prodromal state and on to a state of advanced dementia (Alzheimer Europe 2016). The issue regarding the new conceptualisation of Alzheimer’s disease is not that some people will be excluded from research but rather that some of those who are included may experience harm as a result of a lack of clarity surrounding terminology.

People may interpret information about the new definitions of Alzheimer’s disease within the framework of their current knowledge about Alzheimer’s disease and dementia. Consequently, there is a risk, in the context of research, that they might interpret categories such as preclinical Alzheimer’s disease, prodromal Alzheimer’s disease (IWG) or MCI due to Alzheimer’s disease (NIA-AA) as forms of dementia (Alzheimer Europe 2016). This may have ethical implications linked to informed consent (i.e. it could not be considered as informed if based on a misunderstanding). This is an issue linked to the involvement of people who do not actually have dementia in research which is beyond the scope of this paper but it is nevertheless closely related.  In a recent paper on this topic, Frisoni et al. (2019) concluded that the disease narratives of researchers and society need to be better aligned to ensure that the new conceptual framework is in tune with the social representation of Alzheimer’s disease. Even if there is no mention of dementia, current lay understandings of Alzheimer’s disease, which are still used by many healthcare professionals, combined with insufficient explanation by researchers, may increase the risk of such assumptions being made. An apparent discrepancy between information provided by different healthcare professionals and researchers may also result in a loss of trust in the healthcare and research professions.


Progress in diagnosing dementia is being made but this is not uniform across Europe. In a European five-country study into carers’ perceptions of the timeliness of dementia diagnoses, 30.7% to 45.5% of carers estimated that diagnosis had occurred at the middle to late stages of dementia (Woods et al. 2018). People in the middle to late stages of dementia have limited opportunities to be involved in research and late diagnosis deprives them of opportunities to be involved earlier.

Not being diagnosed at all (i.e. ever) is even more problematic  with regard to the pursuit of inclusive research. Some studies suggest that more than half of community-dwelling people with dementia have not been diagnosed and this corresponds with the findings of a recent systematic literature review and meta-analysis of the prevalence and determinants of undetected dementia in the community (Lang et al. 2017). In this study, the rate of undetected dementia in Europe was 53.7%. A recent study of residents with dementia in German nursing homes found that 30% to 40% had not been accurately diagnosed and that inaccurate diagnoses had also been reported in studies in Norway, Israel and Ireland (Palm et al. 2016). Findings from the European systematic literature review were mixed regarding the number of cases of undetected dementia in relation to place of residence (community or residential setting) or ethnicity (Lang et al. 2017). However, the authors highlighted certain factors which were associated with a higher risk of not being diagnosed such as low socio-economic status, not speaking the local language, living alone, gender (male) and younger age.

Other factors which have been identified as rendering timely and accurate diagnosis of some groups of people more problematic include difficulties with language and communication, cultural perceptions of dementia and health, stigma and taboo, prejudice, structural discrimination and lack of referral by GPs (Beattie et al. 2005, Nielsen at al. 2011a and 2011b). These factors are common amongst people from minority ethnic groups but also apply to many other people with dementia. Nielsen et al. (2015) point out that in Denmark, only 11% of the expected number of older people from minority ethnic groups with dementia receive a formal diagnosis of dementia. It has been suggested that many GPs lack the cultural and linguistic skills to diagnose people from minority ethnic groups as well as the appropriate diagnostic instruments. Also, many people with dementia do not receive a formal diagnosis, either from their GP or from a specialist (Beattie et al. 2005, Diaz, Kumar and Engedal 2015). In some studies, it may be possible to include people who, for a range of reasons, do not have a clear diagnosis of dementia on the basis of reported symptoms. This would enable people who have dementia but do not have a clear clinical diagnosis to be involved in research. However, in such cases, it would be necessary to be consistent (e.g. to describe the study using the same terms and on the basis of the same inclusion criteria.  Otherwise, the findings could be misleading and people would not be motivated to participate (i.e. not seeing the link to their own condition or lived experience).

Relying on recorded medical diagnoses of dementia can be problematic if they have not all been made on the basis of the same criteria. Diagnoses made in some settings and using some criteria may be more thorough than others (e.g. including imaging and complex cognitive testing). There is evidence in some countries that the use of imaging techniques and neuropsychological testing is lower in the primary care sector (Palm et al. 2016). Researchers who need to determine sub-types of dementia on the basis of existing diagnostic findings would therefore be unable to include some people with dementia in their studies. Moreover, Palm et al. (2016) point out that in the late stage of dementia, a differentiation between dementia subtypes is difficult. This means that people in the later stage of dementia who did not initially receive a differential diagnosis would not be able to participate in some studies. On the other hand, whereas the vast majority of research into dementia is about Alzheimer’s dementia, non-drug interventional studies often do not specify the type/cause of dementia (Jongsma et al. 2016).

Despite a well-established and documented relationship between Down’s Syndrome and Alzheimer’s disease, with the first clinical signs of dementia usually occurring around the age of 50, the diagnosis of dementia in people with Down’s Syndrome remains highly problematic (Cipriani et al. 2018). This is partly because of pre-existing levels of intellectual disability, lack of a consistently used diagnostic tool and a huge variation in levels of performance on cognitive tasks, combined with communication difficulties. It may also be because early signs of dementia differ from those experienced by people who do not have Down’s Syndrome (Cipriani et al. 2018). Such difficulties at times result in ‘probable’ rather than confirmed diagnosis. A probable diagnosis of dementia, in people who do, and do not, have an intellectual disability should not be viewed as a reason to exclude them from research participation but should be noted.

Recruiting people with dementia using proxy measures of diagnosis (i.e. carers’ or healthcare professionals’ opinions rather than a formal diagnosis) is one way of including people who might otherwise be excluded but raises both scientific and ethical challenges. First, it touches on the validity and generalisability of the findings in that they cannot be generalised to people with dementia as it is not known whether all the participants actually had dementia. Second, it raises questions about trustworthiness and respect in that the researchers may strongly suspect that the participants have dementia (based on the information they have been given), but have not informed the people concerned. Indeed, doing so could be considered unethical due to the risk of causing harm but it also could also be considered a form of deception and an invasion of people’s privacy. Finally, with regard to respect for autonomy, people should be informed amongst other things about the purpose of the study and why they have been invited to participate.  If this does not occur, the volitional component of consent (i.e. the exercise of free will) would not have been fulfilled (Faden and Beauchamp 1986).  


Recommendations for researchers

  • Reflect on the justification and possible discriminatory assumptions behind the selection of inclusion and exclusion criteria.
  • Take measures to ensure that very old and frail people with multi-morbidities (which includes many people with dementia) are not unfairly excluded from research.
  • Only require a study partner if essential for the research or the participant’s well-being.
  • Consider research approaches for which a study partner would not be required.
  • Explain to research ethics committees the rationale for including population groups that may at times have ‘suspected’ rather than confirmed diagnoses (e.g. people with Down’s syndrome).
  • Explain to research ethics committees the rationale for exclusion criteria which may exclude whole sub-groups of people with dementia from participating in research.  
  • Consider whether certain exclusion criteria could not be avoided if appropriate support was provided (e.g. ensuring that materials are accessible rather than excluding people with less than a certain number of years of education).
  • Avoid inconsistent, ambiguous or too broad definitions and conceptualisations of groups of people with dementia.

Recommendations for research ethics committees

  • Challenge ambiguous inclusion and exclusion criteria in research protocols, which limit the internal and external validity of research.

Recommendations for funders

  • Cover expenses such as PI, logistics support and recruitment of diverse groups to allow for more inclusive research.
  • Foster studies that have relevant outcome measures for people with dementia.
  • Foster studies that contribute towards quality of care, including care for those living alone or in nursing homes and non-pharmacological studies whilst recognising that these studies may have small sample sizes.



Having inclusion and exclusion criteria enables researchers to involve people who will, collectively, be able to provide an answer to their research question or hypothesis and not jeopardise the success of their study or result in harm. A key question is, nevertheless, whether the rationale for inclusion and exclusion is scientifically justifiable and fair. Criteria typically include age, place of residence, literacy and language ability, level of education, mobility, co-morbidities and having a specific diagnosis. Sometimes they are unjustifiable (e.g. based on common practice, assumptions, reluctance to challenge research ethics committees and fear of anything that might risk a particular result not being found). Care needs to be taken when developing inclusion and exclusion criteria to ensure that people are not unnecessarily deprived of the opportunity to take part and that the criteria do not jeopardise the external validity of the study (i.e. resulting in findings from quantitative research studies being generalised to people who have little in common with those who participated in the study.

Issues related to the informed consent procedure

The principle of informed consent is one of the most fundamental foundations for the conduct of ethical research and, in addition, a legal prerequisite for the conduct of biomedical and some forms of clinical research. The aim of informed consent is to protect participants from harm (in keeping with the principle of non-maleficence) whilst respecting their autonomy and avoiding deception or coercion (Smith 2008, Beauchamp and Childress 2012). It is also linked to the principle of justice (or rather to potential injustice) in that people who seem unable to consent, but were not provided with accessible information and appropriate support (e.g. suited to impairments they may have), may be wrongly deprived of the same opportunities to participate in research that other people have. People with dementia may also be able to rely on proxy decision makers and/or advance directives as a means to express their right to self-determination at a time when this would otherwise be difficult or impossible. The use of such options may be considered empowering or restrictive and their ethical, practical and legal implications need to be considered. In this section, we examine possible challenges surrounding informed consent from people with dementia (including those for whom informed consent may be particularly difficult), the role of proxy decision makers and the use of advance directives for research involving people with dementia. But first, we will briefly examine the concepts of capacity and competence.

Capacity and competence 

Informed consent offers participants the opportunity to weigh the possible benefits against the burden and risk of participating in research. It has a volitional component. This means that research participants must be sufficiently informed and that consent must be given voluntarily and without any coercive measures being used. There is also a cognitive component which is that research participants must have the capacity to make the decision at hand and must be competent (Faden and Beauchamp 1986, Beauchamp and Childress 2012). Decisional capacity and competence are sometimes used interchangeably, but their meaning differs slightly. Decisional capacities are the capacities needed to make a decision. Competence is the dichotomous legal judgement whether the capacities are enough to make the decision at hand. This must be in relation to a particular task, at a particular moment in time and under specified conditions (Buchanan and Brock 1990). Competent persons have sufficient decision-making capacities, while incompetent persons have insufficient decision-making capacities.

The area of competence of relevance to the involvement of people with dementia in research is that of decision-making capacity. People may have varying degrees of decision-making capacity. In the case of dementia, it may fluctuate not only because of dementia but also due to a range of psychosocial, situational, medical, psychiatric and neurological factors (Holzer et al. 1997). However, whereas capacity is not an all or nothing matter in that people may possess varying degrees of capacities (with regard to a particular task), competence (i.e. the legal concept) is. A person either reaches the established threshold for competence with regard to a particular task or does not (Buchanan and Brock 1990). So a person is judged either competent or incompetent to make a decision about participating in a particular study. The importance of making such a distinction is to determine whether a potential participant should consent to participation in a particular study or whether somebody else should make that decision on his/her behalf. Lack or loss of the capacity to consent to research should not systematically result in a person’s exclusion from research (Hérault, Bravo and Trottier 2018).

Part of the process of obtaining informed consent involves determining whether a person has the required capacities for the decision at hand. Competence does not equate with perfect rationality in that people have the right to make decisions which others may consider unwise or irrational provided that they fulfil the criteria governing valid informed consent (DuBois 2008). This is particularly important in relation to people with dementia from various sub-groups in society who may have different values, expectations and beliefs to those of the researchers. Other factors such as level of education, socio-economic status, language and communication difficulties and distrust may also affect assessments of capacity and ultimately influence whether a person with dementia is deemed competent to participate in a particular study.

Assessment of the capacity to consent

Informed consent should not be understood as a signed piece of paper or a one-off requirement at the beginning of a study. Rather, it is a process based on verbal, non-verbal and behavioural cues, which needs to be revisited regularly throughout the research (Dewing 2007, Hellström, Nolan, Nordenfelt and Lundh 2007) and accompanied by ongoing reflection about the ethical involvement of participants.

“Consent-seeking should not be thought of merely as an event. The securing of approval from research ethics committees or the signing of a consent form by respondents do not, in themselves, absolve researchers of further need to engage in ethical concerns and reflexive practice. Researchers should refrain from treating such ritualistic enactment of consent-seeking as the be-all and end-all of ethical considerations but should instead engage in a more reflexive approach to the types and levels of consent required before, during and after the act of research” (Sin 2005, p. 290).

This is applicable to all research designs but a degree of flexibility and the possibility for adaptation, which may result in new ethical issues arising were not previously anticipated, is particularly relevant for longitudinal studies and some qualitative research designs (Hellström et al. 2007). A range of approaches can be used to obtain informed consent for participation in research such as clinical interviews, discussions and measures of capacity, including neurological test batteries (Beattie 2009).  

The MacArthur Competency Assessment Tool for Clinical Research (MacCAT-CR) (Appelbaum 2007, Appelbaum and Grisso 2001) is sometimes described as the ‘gold standard’ for the assessment of capacity to consent to clinical research, although not necessarily on its own (Howe 2012). It covers understanding, appreciation of the nature of the situation, reasoning and expressing a choice. Although the tool provides a score, the end decision about capacity to consent is made by the physician administering the tool and this reflects a judgement about the appropriateness of the decision rather than the person’s capacity and hence is a normative judgement (Jongsma 2016). This means that people who make decisions about whether to participate in research which reflect different values, atypical reasoning, eccentric views or perspectives which in some way are interpreted as falling outside the norm, may be deemed as lacking the capacity to consent to research. As mentioned earlier, this would be unethical but may well be the case for many people from sub-groups in society with which researchers are not familiar. On the other hand, it could also be argued that by not providing a cut-off score, the MacCAT-CR provides greater flexibility and, if used sensitively, could be a more inclusive tool for use with people whose scores do not fully reflect their actual capacity to consent.

Examples of other instruments include the University of California San Diego Brief Assessment of Capacity to Consent (UBACC) and the Evaluation to Sign Consent (ESC) (Beattie 2009). The Mini Mental Status Exam (MMSE) is not generally considered as a suitable or reliable measure of the capacity to consent to research (Howe 2012). Moreover, as with diagnosis, care should be taken when drawing conclusions about capacity to consent using instruments which have not been validated for people from various minority groups.

The ability of researchers from different domains (i.e. including non-medical research) to assess capacity to consent to research varies dramatically and some researchers may need to seek assistance from their peers, or at least to reflect on the possibility of having an external assessor of the capacity to consent. Non-medical research does not necessarily entail the same risks to participants as clinical drugs trials and the level of understanding required by participants may sometimes be lower. Nevertheless, it may result in some form of harm and still touches on a range of ethical issues such as respect for autonomy, voluntariness, justice, confidentiality and dignity.

Some organisations and institutions have developed checklists for researchers of things to consider when assessing the capacity to consent to participation in research. The British Psychological Society (Dobson 2008)[12] in the United Kingdom, for example, has a form to guide researchers when participants do not have capacity to consent. This covers whether efforts have been made to maximise the capacity to consent, whether impairments exist which might affect decision-making capacity, whether the person understands the voluntary nature of the research, what it is about and is able to understand and weighed up the benefits and risks, and finally whether he or she has communicated his/her decision regarding participation. Depending on the responses given, researchers are advised to inform themselves about how to enable decision making, discuss the issue with the Principal Researcher or conclude that the person does not have the capacity to grant or withhold consent to take part in the research project.

The ongoing assessment of consent can be challenging in relation to people who already have dementia and whose capacity to consent may deteriorate during a particular study. Hellström et al. (2007) argue that most approaches to consent rely too heavily on cognitive ability, including the recollection and manipulation of facts, do not take feelings and experiences sufficiently into account and therefore place people with dementia in a threatening and unpleasant situation. The performance of people with dementia on tests of cognitive performance may be influenced by a range of factors such as the time of the day, level of concentration and motivation. Cognitive processing may also be affected by a person’s state of general well-being (Damasio 2000). Consequently, tests which are carried out with minimal regard for the individual may affect the results obtained and lead to the unnecessary exclusion of some people with dementia from research.

Flexible and adapted informed consent procedures need to be developed or elaborated in collaboration with members of the groups concerned. This might, for example, include ensuring that information for informed consent is presented in a way that makes it easier for people from specific groups to understand and retain it sufficiently long to make a decision. Dewing (2008) points out that traditional competency-based approaches to informed consent fail to take adequate account of the situational aspect of capacity and the importance of interdependence and relationships. Some researchers have proposed alternative methods such as verbal and behavioural consent (Bamford and Bruce 2000, Bartlett and Martin 2002, Cowdell 2006). Possibilities for alternative forms of consent must also be considered in the case of people with dementia with additional impairments (e.g. visual impairments and intellectual disabilities).

Many people with dementia are able to give informed consent if appropriate adjustments are made and necessary support provided. Process consent is an approach and method to obtain ongoing consent from people “who have an extremely limited capacity, who would generally be thought to be incapable of legally informed consent by others, but on observation can communicate and express their wishes in other ways” (Dewing 2008, p.63). It is a person-centred approach which is also influenced by the concept of social justice and the importance of relationships. Process consent has been implemented, following ethics approval, in the United Kingdom, Ireland and Australia for several qualitative research projects involving people with dementia in the field of gerontology. Dewing (2008) describes the approach as a way to work towards inclusion rather than exclusion of people with dementia in research but recognises that there will still be a cut-off point at which some people will not be able to make or communicate even small choices.

Process consent offers an alternative to the current over-emphasis on cognitive capacity, is flexible and focuses on individuals and their residual capacities and other abilities. It is not yet known whether this approach could be further expanded beyond the specific area of gerontological research and there are a few challenges surrounding the observation and interpretation of states of wellbeing (see Part 3 for more on this topic). However, Dewing (2008) draws attention to the risk of researchers too readily accepting the gold standard of informed consent and emphasises the need to seek alternative methodologies for consent.

Being informed vs understanding 

Information provided to potential research participants with dementia should cover various issues related to a particular study (e.g. what the study is about, why the person is being asked to participate, what participation would involve, potential risks and possible benefits if any, who is responsible for the study, the timeframe of the study and whom to contact for more information, further explanation or in case of complaints) and general issues related to participation in research (e.g. the voluntary nature of participation, the right to withdraw at any time, confidentiality and privacy issues). Participants must be able to comprehend the information provided, demonstrate that they have understood, assess the relevance of information to them and make a reasoned decision.

Some research can be particularly complex, requiring researchers to provide additional information to help people understand various concepts and the purpose of the study. In order to promote inclusive research, it is important to ensure that every effort is made, even for relatively straightforward studies about everyday issues, to ensure that participant information sheets and informed consent forms are accessible to people with varying degrees of cognitive impairment, intellectual disabilities and language and communication difficulties, as well as being appropriate for a culturally diverse population. It may be necessary to develop more than one version of relevant documentation, to provide alternative materials and tools, and to provide support. Watchman (2014) also emphasises the importance of adopting a flexible and individualistic approach to facilitate the active engagement and participation of people with intellectual disabilities and dementia in research.

Assessment of burden, risk and potential benefit

In the context of research, risk is often discussed alongside burden and balanced against possible benefit. Ethical research aims to minimise risk and enhance potential benefit so that the risks to the participant are proportionate to the potential benefits to that person and society (Emanuel, Wendler and Grady 2000). However, the definition of benefit is often too narrow to take fully into account the kinds of benefits that are possible and important to people with dementia participating in research, such as spiritual, emotional or psychological benefit. Moreover, there are ethical and legal limits to the degree of risk and burden that is considered acceptable for research involving people who are unable to consent. For example the Council of Europe’s Additional Protocol on Human Rights and Biomedicine (2005) only allows research without a direct benefit with people who are unable to consent if the research entails only minimal risk and minimal burden.

Additionally, article 15 of this document  contains a series of conditions in the case of people without the capacity to consent to research. These include conditions such as that the results have the potential to produce real and direct benefit to their health, that the research could not be carried out on people capable of giving consent, that the people have been properly informed of rights and safeguards, authorisations from legal representatives have been obtained, previously expressed wishes taken into account and they have not objected to participation.


The issue of direct benefit may be side-stepped subject to fulfilment of the above-mentioned conditions provided that, amongst other conditions:

  • “the research has the aim of contributing, through significant improvement in the scientific understanding of the individual's condition, disease or disorder, to the ultimate attainment of results capable of conferring benefit to the person concerned or to other persons in the same age category or afflicted with the same disease or disorder or having the same condition;
  • the research entails only minimal risk and minimal burden for the individual concerned; and any consideration of additional potential benefits of the research shall not be used to justify an increased level of risk or burden.”

Assessments of risk, benefit and burden cannot easily be made for whole groups of people and there is a risk of paternalism, leading to a lot of people being overprotected and excluded from research without having any say in the matter. Drawing on a relational ethics perspective, Fisher (2009) suggests that failure to include the perspectives of those directly concerned in such discussions may lead to the rejection of scientific research procedures which potential participants and their relatives might perceive as benign and/or worthwhile considering. She states:

“Formulating regulations and ethical judgments solely on the bases of opinions expressed by experts in the scholarly community and IRB members risks treating subjects as ‘research material’ rather than as moral agents with the right to judge the ethicality of investigative procedures in which they participate”(2009, p.5).

As unique individuals, people with dementia have their own perceptions of danger and of the level of risk that they are 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). There is therefore a risk of overprotecting people with dementia, infantilising them and depriving them of their autonomy, which is problematic for the reasons discussed earlier in the sub-section on gatekeeping (Jongsma and Schweda 2018).

With regard to benefits which are not directly linked to health, Casaret, Jason and Karlawish (2000) point out that when people are approaching death, for example, 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, participating in research might 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.

People with dementia, and those whose opinions they consider valuable, 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. Details of the responses from participants (or in some cases their proxies) about the perceived risks, burdens and benefits of participation should be taken into consideration by researchers when designing subsequent studies. It is important to realise that in practice, researchers focus a lot on lowering risk when designing their studies and often forget about the burden that participation may entail. While burden is also a very important factor in potential participants’ decision making, burden for them does not only entail specific procedures. It also includes logistical burden (e.g. travelling to hospitals, time spent participating) and burden for the family and carers. These types of burden are very important in relation to decision making and should already be dealt with by researchers in the research protocol.

Voluntariness and the therapeutic misconception

Participation in research of any kind must be voluntary. There are different degrees of lack of voluntariness ranging from being subjected to research without having had any choice in the matter, being forced to participate on the basis of some kind of threat or deciding to participate on the basis of undue influence (DuBois 2008). Pressure to participate can be quite subtle and may be unintentional. An example would be a person being asked by a healthcare professional or staff in a residential care setting and fearing upsetting that person due to a feeling of dependency on him or her, respect for his/her position or not wanting to jeopardise  an existing positive relationship with that person.

People’s voluntariness could also be influenced by their interpretation of what they are being asked to do and why. The decision about whether or not to participate in research may depend on how research is understood and most importantly, whether a person understands the difference between research and treatment or care. This might not always be as obvious as it first seems, especially in situations where a person is informed about a study by a healthcare professional in a healthcare setting (Lewens 2006). There is some evidence to suggest that people with dementia considering participation in research frequently mistake proposed research for treatment options (Dresser 2001). The risk of this happening may be even greater in the case of people with mild dementia who write combined advance directives for treatment and research (Pierce 2010). Medical researchers and healthcare professionals involved in recruitment for research should not take it for granted that people with dementia understand this difference and also bear in mind other factors which may make it more difficult for some people to understand (e.g. people who are not familiar with the way the healthcare system works and the connection to medical research, who have difficulties understanding the language used, who are stressed about their condition or who feel intimidated by the perceived power status of the person providing them with the explanation).

The provision of treatment and care is generally consented to on the basis of an assumption that it is necessary and will hopefully in some way be personally beneficial. If this distinction is not clear, it cannot be said that a person has given informed consent to research. Medical researchers should try to ensure that potential participants realise that the overall aim of research, including biomedical research, is not to improve their health or contribute to their personal wellbeing.

It is also important to ensure that healthcare professionals responsible for recruiting research participants, who are sometimes extremely enthusiastic about the study, do not ‘oversell’ the virtues of the study or communicate their assumption that it will be beneficial in some way to potential participants. Research participants will not necessarily benefit from better monitoring, for example, and even if this were the case, this should not be a selling point but rather an indication of the need to improve standard diagnosis and care.

Financial incentives to contribute to research in the context of PI were discussed in Part 1 of this discussion paper and linked to arguments surrounding recognition of the value of a person’s contribution to research. With regard to potential research participants being offered financial incentives, the principle of justice needs to be considered because payment could contribute towards the exploitation of some groups of people and skew samples in much the same way as their systematic exclusion would. In addition, for some people sufficient payment might result in them accepting risks or burdens that they would otherwise not have been willing to take.

Vulnerability and disability

Researchers may be faced with difficult decisions about how to comply with rigidly defined informed consent procedures, including additional measures to protect vulnerable participants, whilst not excluding certain groups of people (such as some people with intellectual disabilities, with more advanced dementia or from various minority groups) or including them without their consent (Doody 2018). Raudonis (1992) describes a fine line between “protecting vulnerable research participants and making paternalistic decisions supposedly in their best interests” (Nordentoft and Kappel 2011, p. 368). Alongside the clear- cut perception of competence as something a person has or does not have, Sin (2005) emphasises the fluidity of consent and the dynamic nature of research, which calls for a more reflexive approach and acknowledgement of the socially constructed nature of assumptions surrounding the concepts of ‘normal’, ‘competent’ and ‘informed consent’.

A lot of the protective measures governing the participation of people with dementia are linked to perceptions that they form a vulnerable group of people. Such vulnerability is an acquired status compared to other groups in society who may have been considered as vulnerable all their lives, which is the case for many people with physical or intellectual disabilities. The perceived vulnerability and in many cases over-estimation of vulnerability may partly have its origins in the long history of institutional care of some groups of people and the focus on looking after or caring for them (Doody 2018). The normalisation movement which began in the 1970s, based on the work of Wolfensberger, Nirje and Bank-Mikkelson amongst others, was linked to intellectual disability and promoted ‘the dignity of risk’ rather than protection. Its focus on integration in community life is perhaps also relevant to the goal of enabling people with dementia to participate in research and for researchers, funders and research ethics committees alike to avoid overly focusing on protection and denying freedom of choice to people with severe, profound or significant disabilities (NHS Health Scotland 2016).

The first three paragraphs of article 12 of the United Nations Convention on the Rights of Persons with Disabilities (CRPD 2006) may also be helpful when considering the issue of consent to research by people with dementia, namely,  

  1. States Parties reaffirm that persons with disabilities have the right to recognition everywhere as persons before the law,
  2. States Parties shall recognize that persons with disabilities enjoy legal capacity on an equal basis with others in all aspects of life,
  3. States Parties shall take appropriate measures to provide access by persons with disabilities to the support they may require in exercising their legal capacity.

Consequently, it should not be concluded that a person lacks competence (i.e. legal capacity) to consent to research unless necessary adaptations and efforts have been made to minimise the possible impact of various impairments (e.g. linked to language, communication, memory and abstract reasoning) on decision-making capacity. Similar adaptations should also be considered in relation to other factors, not necessarily physical, mental, intellectual or sensory impairments, which may impact on the capacity of a person with dementia to give informed consent.

Assent and dissent

A key aspect of the informed consent procedure, also linked to the requirement of voluntariness, is that people understand that they have the right not only to refuse to participate in a study but also to withdraw at any time without having to provide any justification for their decision. When a person with dementia continues to participate though unable to confirm his/her informed consent (e.g. based on a proxy decision maker having taken over this responsibility), special attention should be paid to signs of possible assent and dissent. These should equally be respected in the case of people who still have the capacity to consent (Jongsma, Bos and van de Vathorst (2015).

Black, Rabins, Sugarman and Karlawish (2010) suggest that both assent and dissent can be expressed or indicated verbally (e.g. saying ‘no’), behaviourally (e.g. acting agreeably or trying to move away) or emotionally (e.g. through facial expression or posture) and advise researchers to consult with informal carers to help them understand individual participants’ typical means of expression in relation to assent and dissent. This means that even if unable to understand the ins and outs of the study or give reasons for wanting to take part or withdraw, a person with fairly advanced dementia may be able to communicate his/her wish to withdraw or refuse a particular research activity and this may contribute towards decisions to involve more people with advanced dementia in research. However, verbal and non-verbal signs of dissent may be subject to individual and cultural variations. Moreover, Young, Ferguson-Coleman and Keady (2014) point out with reference to Deaf people, much of the focus on non-verbal communication of people with dementia equates non-verbal with non-linguistic, whereas with Deaf people the non-verbal is actually language. Facial expression, raising or lowering the eyebrows, puffing out the cheeks and movement of the eyes all contribute towards the meaning of an utterance. Similarly, averting one’s gaze is not simply a non-verbal marker of indifference but a means of disengagement as no further communication can take place (Young et al. 2014).

Black et al. (2010) describe the concept of assent as a means to respect a person’s remaining autonomy and to enable cognitively impaired people to be involved, to the extent that this is possible, in the decision making process. Careful attention to signs of assent or dissent throughout the research process may also help keep the person with dementia and his/her needs and wishes at the centre of research and avoid him/her being treated as a mere means to an end.

The term ‘dissent’ is used somewhat ambiguously in the literature. One definition refers to objection during a trial (‘the wish to discontinue participation’), also sometimes referred to as objection, distress or resistance. This notion of dissent draws on respect for autonomy whereby dissent should be closely monitored, given that some people who lack decisional capacity are not able (cannot formulate the wish) to withdraw during the trial (Jongsma, Bos and van de Vathorst 2015). Another definition describes dissent as the opposite of consent or assent, consisting of a measure of protection from risks and burden that a person might not want to accept, a means to respect his/her dignity and a shield from distress or from unwanted research activities. Black et al. suggest that the need to respect dissent is, ethically-speaking, hierarchically more important than obtaining assent due to its protective function but nevertheless argue in favour of including people with dementia who are unable to consent and also unable to dissent. They recommend that research ethics committees determine whether any protection beyond proxy consent (e.g. such as an advance directive for research and independent monitor) is needed for people who lack the ability to assent and/or the ability to dissent.


Recommendations for researchers

  • Ensure that informed consent does not consist of a one-off event if the research is ongoing over a period of time.
  • Clarify the process for determining capacity on an ongoing basis.
  • Bear in mind that although a diagnosis of dementia may justify an evaluation of a person’s capacity to consent to research, it does not automatically mean that a person lacks such capacity.
  • Reflect on appropriate processes and legislation where judgement of capacity is difficult.
  • Develop and propose to research ethics committees alternative methods of obtaining informed consent which are neither exclusionary nor paternalistic.
  • If people are unable to write, check whether your research ethics committee would accept oral consent (e.g. in the presence of at least one witness and repeated at regular intervals).
  • Provide details about how and by whom capacity will be assessed.
  • Where possible, arrange for ongoing assessments of capacity to be carried out by those with the necessary expertise who are independent from the research so as to avoid bias.
  • Be transparent about the role of medical researchers so as to prevent therapeutic misconception (e.g. do not sign letters about possible research trials as ‘your treating doctor’).
  • Where possible, try to ensure that healthcare professionals (where relevant) proposing research participation are not also responsible for a person’s treatment or care, so as to avoid putting the person with dementia under pressure to consent and also to avoid the therapeutic misconception (i.e. mistaking a proposal to participate in research for an offer of treatment).
  • Take measures to include people with dementia in the informed consent process, including when they are deemed to lack capacity (e.g. provide information and ask for assent).
  • Closely monitor participants during research and check whether they are still willing to participate and understand what participation involves.
  • Use validated tools, wherever possible and when available, to obtain consent from people from minority groups or with specific characteristics which necessitate special measures.
  • Consider developing or using different methods and tools, where required, to provide people with dementia with information about the proposed research.


Recommendations for research ethics committees

  • Require researchers to describe in research protocols the method used to identify a person who could support the patient in the informed consent process, if needed.
  • Recommend that the healthcare professional proposing participation in research is not also responsible for a potential participant’s treatment or care.
  • Require researchers to provide details of their plans to seek ongoing informed consent and to provide justification if this is not planned.  


Recommendations for funders

  • Ensure that any reimbursement or payment proposed to research participants is in line with available national or local guidance on rates, or in keeping with the principle of fair market value.
  • Encourage researchers to incorporate costs related to obtaining ongoing consent in their requests for funding.
  • Ensure clarity about the process to monitor ongoing capacity if research continues over an extended period of time.


The informed consent procedure is fundamental to the conduct of ethically sound research and closely linked to the principles of respect for autonomy, non-maleficence and justice. Approaches are needed which provide appropriate protection from harm whilst not overriding the right to self-determination (e.g. through blanket judgement of vulnerability). Informed consent should be seen as an ongoing process, continuing throughout the whole research process. It is important to be attentive to signs of assent and dissent which may be expressed differently by people with dementia from different sub-groups of the population. The capacity to provide ongoing consent and the way that assent and dissent is communicated is likely to vary for each person as dementia progresses.

The issues of burden, risk and benefit, and the importance that each person attributes to them, need to be considered, as well as the potential confusion between research and treatment, especially with regard to clinical research. Creative and innovative methods are needed to make it possible for a diverse population of people with dementia to be involved in research (e.g. people with dementia from minority ethnic groups, with lower levels of education, language or literacy difficulties, with intellectual disability or with more advanced dementia). Procedures, attitudes and the environment should promote rather than hinder the abilities of people with dementia to provide consent.

Consent with and on behalf of people with dementia: shared, supported and proxy decision making

Researchers often suggest to potential participants that they discuss their possible participation in a study with their close family and friends, and state that they are more than welcome to contact them for further clarification or with any questions. However, discussions about the involvement of other people in the decision-making process related to informed consent to research are often linked to concerns about a person lacking the capacity to provide informed consent. There has also been a move away from the concept of proxy (i.e. substituted) decision making towards shared decision making and supported decision making. This is sometimes seen as a way to promote autonomy and enable people with dementia who have limited capacities to continue to play a role in decision making. The concepts of shared decision making and supported decision are similar in some respects and may be complementary to some extent. They developed separately but in parallel and the origin of each leads to a slight different in how they are perceived. 

Shared decision making, for example, focuses on a range of skills and practices that can be used to promote a collaborative decision making process. This has been typically used in the context of healthcare decision making (Elwyn et al. (2012). The emphasis is on the decision-making process with the actual decision, especially the issue of who made it, being considered as far less important. The process, which typically involves healthcare professionals as well as patients, involves the use of decision aids, the provision of information and ongoing discussions, all of which are aimed at increasing knowledge, helping people to feel less confused about a range of options and resulting in more informed and values-based choices, better communication and people being more involved in making decisions (Simmons and Gooding 2017). Shared decision making has, however, been criticised for failing to take sufficiently into account the power imbalance which often exists between patients and healthcare providers or researchers and for decision aids often failing to accommodate for the needs of people with low health literacy (Simmons and Gooding 2017).

Shared decision making can also be understood as an individual or cultural choice or philosophy irrespective of cognitive ability. For many people with dementia, participating or sharing in the decision-making process is as important as, if not more important than, making the actual decision (Daly, Bunn and Goodman 2018). In some cultures, there is a deeply rooted set of ethical and cultural traditions which reflect familism rather than the individualist approach common to informed consent procedures in Western culture. Writing in the context of recent attempts to promote a more individualistic approach to informed consent in China, Bian (2015) states:

“Familism is the Confucian family-based and family-oriented way of life embodied in the Chinese mode of decision-making in which all close family members play an important role. Close family members in contemporary Chinese society typically include the patient’s spouse and parents (especially when one’s children are not yet adults), as well as adult children. When facing major decisions for any family member, all close family members come together to make a collective decision” (Bian 2015, p.375).

According to this philosophy, the family is prioritised over the individual, which runs counter to the taken-for-granted practices of individualism, which are dominant in many modern Western societies (Bian 2015). Familism and the practice of shared decision making is common amongst several minority ethnic communities in Europe and is not related to lack of decision-making capacity. Enabling a less individualistic approach to decision making in the context of research may be an important means to promote the ethical involvement of a more diverse set of people with dementia in research, irrespective of the reason for the choice of this approach. 

Supported decision making is about providing the necessary support to enable a person to make a decision, including one that is legally recognised. Whilst the term does not appear in the UN Convention on the Rights of Persons with Disabilities (2006), supported decision making very much reflects the ethos of that convention and is often associated with it. Simmons and Gooding (2017, p.276) describe the rationale for supported decision making as follows:

“In crude terms, supported decision-making promotes the idea that just as people who use wheelchairs are entitled to ramps in order to access buildings, so too people with mental health-related disability - ‘psychosocial disability’ – are entitled to support to exercise choices and their lives. (The same idea holds for people with intellectual disability, or any other disability that may affect decision-making, including the way other people perceive and/or denigrate a person’s decision-making ability).”  

In their recommendations on research (and data sharing) involving people with dementia, Thorogood et al. (2018) point out that research ethics guidelines generally promote supported or shared decision-making.  They suggest that this might include “simplifying consent forms, providing visual or memory aids, taking interactive or educational approaches (where persons with dementia are asked to explain their understanding of consent elements), re-explaining misunderstood information, or involving familiar carers to facilitate explanation and communication of a decision” (Thorogood et al. 2018, p.1338). Such measures could perhaps be further extended to address any barrier experienced by people with dementia, regardless of whether it is linked to a health condition (e.g. such as having a lower level of literacy or being a member of group that is stigmatised etc.). 

In situations where a person has been appointed who is authorised to decide on behalf of a person with dementia about participation in research (i.e. a ‘proxy’ or ‘substitute’ decision maker), it is increasingly excepted that that person should base his/her decisions on the known will and preferences of the person with dementia and not on the basis of ‘best interests’ or ‘welfare’ (Thorogood et al. 2018). Where possible, people with dementia should, as far as possible, be included in the decision-making process. In addition, even people evaluated as unable to give consent for research may preserve the capacity to appoint a research proxy (Kim et al. 2011).

It is increasingly recommended, if not stipulated (European Directive 2001), that proxy consent should represent the person’s presumed will. In reality, this is often not the case. Proxies often do not know a person’s will and preference and make decisions that the person appointing them would not have made (Kim et al. 2013, Thorogood et al. 2018). Proxies may sometimes be too restrictive, thus depriving people with dementia of the opportunity to be involved in research or, on the contrary, involve people with dementia in research when this does not correspond to their wishes, values and interests (Jongsma 2016). 

An important area to consider is that of risk because decisions about participation in research involve considering potential risks, burdens and benefits. According to Landau et al. (2010), a person’s position regarding the relative importance of safety and autonomy may depend on whether he or she is personally responsible for the safety and wellbeing of the person with dementia. People may be more protective of others, than they would be of themselves (Nuffield Council on Bioethics 2009), particularly in cases where the person being ‘protected’ is considered in some way vulnerable. Proxy consent is often described as paternalistic but as Hellström et al. (2007) point out, it may also serve as a means to enable people with dementia, who would otherwise be excluded from research, to engage in an activity which can, in many cases, boost self-esteem, be enjoyable and enriching, and provide a means to validate feelings and experiences.

The recommendation or obligation for proxies to take into account current or previously expressed wishes of a person with dementia who lacks the capacity to consent to research could be considered as a factor contributing towards respect for autonomy, beneficence and non-maleficence, but the phrase ‘taken into account’ does not guarantee that the proxy will respect the wishes of the latter. If the known or supposed wishes of the person with dementia are not acted on (unless there is a morally justifiable reason for not doing so), it cannot be said that his or her autonomy has been respected. This suggestion should not be interpreted as undermining the goodwill or underestimating the difficulties that many proxies experience when making decisions on behalf of people with dementia. Proxies can be very beneficial in combination with advance directives because they may be well-placed to help interpret the kinds of studies a person might have wanted to participate in, drawing on the information provided in an advance directive, their knowledge of the person and possible information about his/her wishes and interests. Some people with dementia grant their proxies the power to override their previously expressed wishes (e.g. in the light of scientific advances which result in unanticipated consequences for participation or on significant changes in the person’s situation). Others may want their advance directives to be followed, without leaving leeway for their proxy to decide.

Some people may be hesitant about taking on the role of proxy. Such hesitancy may be linked to a range of factors such as not understanding the concept of proxy decision making, not understanding the possible legal implications of being a proxy, worrying about the level of commitment needed, being unsure about what knowledge and skills they may need and whether they have that, having language difficulties, finding explanations and documentation difficult to understand, having concerns about having to discuss personal or complex issues with researchers or lacking trust in researchers. In some communities, there may be issues of respect and clearly defined responsibilities which influence who should take on that role. That person might not be the one who understands best the needs and wishes of the person with dementia. In some communities, it may be common to involve several people in such important decisions and not wish to appoint one person as proxy.


Recommendations for researchers

  • Show willingness and be proactive in discussing the issue of shared, supported and substitute decision making with potential participants with dementia and other relevant people. Some people will not be familiar with the different options.
  • Ensure that people with dementia are encouraged, where possible, to make their own decisions relating to participation in research, bearing in mind that some may prefer to appoint a proxy.
  • Do not seek the opinion of a proxy decision maker if the person with dementia has the capacity to decide for him/herself.


Recommendations for research ethics committees

  • Ask researchers to describe how they intend to involve people with dementia who lack the capacity to provide informed consent to research without assistance or support.
    • Be open to potential participants’ possible wishes to involve significant others in the informed consent process.



When promoting inclusive research, it is important to respect priorities and traditions related to decision making which are more family or group orientated (as opposed to the more individualistic approach common in much of Europe). As dementia progresses, people may in any case need more support in order to make and express decisions related to their potential participation in research. Shared decision making helps ensure that when people with dementia lack the capacity to make decisions on their own, every attempt is made to enable them to participate in the decision-making process. In some cases, where they are not able to make a decision themselves, such participation may be just as important as the decision that is eventually made. Supported decision making is about providing sufficient support to enable the person with dementia to make a decision, including one that is legally recognised. Proxy decision making does not simply involve making a decision on someone else’s behalf because proxies are expected to make decisions which represent a person’s will and preferences (and their presumed will if they are currently unable to express preferences). Researchers need to play an active role in whatever form of decision making is legally appropriate and adopted so as to ensure that people with dementia are meaningfully involved to the greatest extent possible and according to their preferences.

Advance directives for research

In the context of inclusive research, it is important to consider the involvement of people with more advanced dementia in research, including those who do not have the necessary capacity to consent and who do not have a representative with the legal right to make such decisions on their behalf. It may be possible to determine what a person’s wishes would have been by consulting close family and friends or an advance directive for research. Advance directives were developed in the 1960s in the United States of America and were originally written to express in writing wishes regarding medical treatment in the event of an accident or illness which made it impossible to exercise self-determination (Vollmann 2001). The moral authority of advance directives resides in the principle of respect for a person’s precedent autonomy (Berghmans 2000, Vollmann 2001), by permitting such  autonomy to be extended into the future, well beyond the point that he or she would otherwise have been able to exercise it. The concept of the advance directive has been further extended to the research situation in some countries. Advance directives may be legally binding or simply advisory depending on their legal status in each country and sometimes on the nature of the decisions to be made (Andorno, Gennet, Jongsma and Elger 2016).

Several ethical concerns have been raised about the use of advance directives in the context of dementia (in general, not necessarily for research). These are mainly linked to arguments about personal identity and changing interests. There is a broad and complex philosophical debate surrounding personhood and dementia. An issue of particular concern and of relevance to the debate about advance directives for research is whether the person who wrote an advance directive is thesameperson as the one for whom it may later be applied, and if not, why the advance directive should be respected.


Parfit (1984), for example, takes the view that personal identity is constituted over time by varying degrees of continuity between former and later selves in terms of a wide range of psychological and physical features. The psychological aspect of personal identity is constituted by the degrees of similarity between two temporally separate selves with regard to a person’s personality, belief structure and desires, which may in certain cases, depending on the degree of similarity and continuity, move from being intra-personal to being inter-personal. In line with this argument, if psychological continuity were to become so deeply disturbed that someone became ‘another person’ (e.g. based on perceived changes in personality due to advanced dementia), would this mean that the advance directive should have no more moral force in relation to that person than it would have, had it been written by a stranger, friend or relative? According to Dresser (1986), a person with dementia may lose the ‘properties’ necessary for retaining personal identity over time, thereby becoming a different moral person. This claim relies on the assumption that people with dementia could lose so much mental capacity that their past preferences and interests no longer seem attributable to them (Dresser 1995, Buchanan and Brock 1989). This claim, that dementia causes a disruption in identity, draws on Parfit’s theory of identity. Based on the assumption that the writer of the advance directive and the person with dementia are different people, it is argued that there is no reason to assume that the directive has any moral authority over the person with dementia.

Dworkin (1994), on the other hand, argues in favour of respecting advance directives. He views autonomy as a reflection of a person’s integrity, whereby people express their character through the lives they lead based on values, commitments, convictions and critical as well as experiential interests. Critical interests include the kinds of things that give meaning to people’s lives and make a person think, had it not been so, their life would have been worse or wasted. Experiential interests cover things that people do simply because they enjoy doing them. Dworkin’s view would seem to imply that having dementia at a particular stage in a person’s life, would just be one stage in a complete life which has already involved different stages. The stage someone is currently in, is affected by interests and concerns which transcend that stage and are important for their life as a whole. The competent and incompetent selves are thus one and the same person. It could therefore be concluded that advance directives should be respected because they are expressions of the critical interests a person has, and that they should therefore take precedence over current experiential interests. However, Jongsma (2016), whilst not rejecting the idea that advance directives should be respected, argues that research participation is not a critical interest for most people and that advance directives for research would be better understood quite simply as ‘declarations of willingness’. Moreover, in keeping with ethically sound research practices, any clear objection to the initiation or continuation of a research procedure should result in the withdrawal of a research participant (Jongsma and van de Vathorst 2015). Consequently, aside from precedent autonomy, researchers also have an obligation to protect people with diminished capacities from undue harm and burden.

Porteri (2018) interprets claims about the disruption of psychological continuity and personal identity as implying that when people with dementia become incompetent, they may become a different and a new person, and that the person they used to be ceases to exist. She points out that potential changes in personality, in beliefs and interests, as a consequence of dementia, are feared precisely because people feel that these changes will affect themselves and not that they will give rise to a different and new person. People with dementia, at whatever stage of the disease, are still recognised as mothers, fathers, partners, sisters, brothers and friends. Relatives and friends have a personal commitment to them and feel that they have responsibilities and duties towards them not just because they are part of the larger human family, but because they are the very same person they used to be (Porteri 2018). Moreover, taking the identity argument to the extreme, if the person is no longer the same person, then the relatives of that person should theoretically have no say in matters relating to him or her. Porteri concludes that body is an important concept in that as long as the body is there, so is the person, adding that bodily continuity as a criterion for considering the person with dementia as the same person better guarantees respect for that person’s rights and wellbeing prior to and after loss of capacity (in this context in relation to participation in research).

The legal status of advance directives for research in Europe (at least for those covering biomedical research) remains unclear. In 2001, the Clinical Trials Directive, to be replaced eventually by the new Regulation on Clinical Trials when it comes into force, makes reference to “incapacitated adults who have not given or not refused informed consent before the onset of their incapacity”. This reference is of relevance to whether additional measures of protection are considered necessary, without there being any further elaboration of the concept of advance directives (Jongsma 2016). The current European legal framework regarding the use of advance directives for research in the context of biomedical research neither explicitly mentions the possibility of their use not forbids it (Andorno et al. 2016). This ambiguity/lack of clarity may fuel unease and uncertainty surrounding the use of advance directives as the sole instrument by which researchers might determine whether or not people with dementia who lack the capacity to consent and have no proxy can participate in research. Finally, the desire to participate in most clinical trials is dependent on participants having a study partner (usually a partner, adult child or close friend). Consequently, without the agreement of a close relative or friend, a person with dementia having expressed the wish to participate in research in an advance directive for research may be prevented from doing so. 

Consenting to future research in an advance directive is also problematic because, as pointed out by Berghmans (1998), it is difficult to give consent for a future experiment which has not yet been devised and which, by the very nature of research, is likely to be innovative. Researchers may find it easier to accept a negative advance directive for research as this does not raise issues about the ambiguity surrounding the risks and burdens people might be exposing themselves to. The risks of research as well as the potential burdens and benefits can only ever be estimated but in the case of decisions made for research that might occur several years later, procedures and methods may have advanced beyond what was initially imaginable.  The ‘informed’ aspect of consent is therefore missing in the case of advance directives for research. However, people can be informed about and helped to understand the implications of this lack of prior knowledge and then make as informed a decision as possible. Those who have a trusted person could be encouraged to appoint and allow the involvement of a proxy (see next sub-section) in future decisions about participation in research.



Recommendations for researchers

  • Involve proxy decision makers, if they have been granted that power, in determining whether the research that is eventually proposed is in line with the wishes expressed in the advance directive.
  • Do not resort to the use of an advance directive for research if the person with dementia currently has the capacity to consent or refuse to participate in research.
    • Consider the current wishes of a person with dementia with regard to continuing with or withdrawing from a study (in terms of indications of assent or dissent). Clear signs of the wish to withdraw from a study should be respected even if the person’s current participation is based on a valid advance directive.  
    • Reflect on personal beliefs about personhood and how this relates to advance decision making.


Recommendations for research ethics committees

  • Recommend that researchers consider previously expressed wishes and preferences when proposing participation in research.
  • Accept preferences expressed in an advance directive as a valid expression of interest in participating in the research.


Advance directives for research provide another way for people to express their wishes about future participation in research, in most cases drawn up before they develop dementia. Ethical concerns around advance directives for research tend to focus on issues related to personal identity, changing interests and that fact that the requirement for people to make ‘informed’ decisions is problematic because relevant and necessary information about particular studies is not available when the directive is drawn up. The notions of a loss of self or personhood and of a complete change of identity, as obstacles to the use of advance directives for research, are rejected by the ethics working group. They emphasise the need to respect any current signs of objection to the initiation or continuation of a research procedure and the importance of involving people who are authorised to do so in determining whether proposed research is in line with the wishes expressed in the advance directive.


[1] Exampes include ResearchMatch, the Alzheimer's Prevention Registry, and the Alzheimer's Association's TrialMatch (Source: Lepore et al. 2017)

[2] For more information on this topic, please see the section on “Challenges surrounding the development of culturally sensitive assessment and diagnostic tools section” in Alzheimer Europe’s 2018 report on Intercultural care and support:

[3] The term ‘gypsy’ is commonly used in the United Kingdom (officially and by members of the gypsy community) but not in continental Europe where it is generally considered offensive.

[4] LGBTI stands for lesbian, gay bisexual, transgender and intersex

[5] Whilst Alzheimer Europe challenges the concept of race and promotes the concept of ethnicity, the term race is sometimes used in this discussion paper to refer to bias and discrimination resulting from persisting beliefs within society about ‘racial’ differences. 

[6] Also in relation to research related to mild cognitive impairment.

[7] Show People are a cultural minority that have owned and operated family-run funfairs and circuses for generations. They often work on rides and attractions at fun fairs in the summer months and settle in the winter to repair the machinery (Horsfield 2017).

[8] European Forum for Good Clinical Practice Geriatric Medicine Working Pary

[9] This is also the case in some clinical trials involving people with preclinical or prodromal Alzheimer’s disease but this is beyond the scope of this report.

[10] The number of older people, including people with dementia, who live alone, is constantly rising (Eurostat 2017)

[11] Examples of expats include people of Scandinavian, German or British origin who have moved to Spain on retirement.

[12] Developed in collaboration with the Royal College of Speech and Language Therapists and the Royal College of Psychiatrists.



Last Updated: Friday 13 November 2020