Informing cognitively healthy research participants of modifiable dementia risk factors: Ethical implications

Introduction

Dementia is a global epidemic that presents profound challenges to individuals, families, health care systems, and societies throughout the world. By 2050, the number of people living with dementia is expected to reach 139 million, and the associated costs to surpass $2.8 trillion by 2030. ¹ Evidence suggests that up to 40% of worldwide dementia can be accounted for by 12 modifiable lifestyle risk factors, many of which begin to impact in mid-life, e.g., alcohol consumption, obesity, and hypertension, and which, therefore, could be prevented or reduced.^2,3 Pathophysiological processes underlying Alzheimer's disease (AD), the most common cause of dementia, may be present decades before the onset of these clinical symptoms. ⁴

The term ‘Alzheimer's disease’ has been variously used to refer to (1) a particular pathophysiological disease process requiring the presence of specific biomarkers, which may or may not precede clinical symptoms, (2) a clinical syndrome characterized by progressive cognitive and behavioral impairment, which is probabilistically linked to the presence of specific biomarkers. Conversely, ‘dementia’ refers to an age-related clinical syndrome characterized by progressive memory and cognitive decline, of which Alzheimer's disease is one potential cause, and for which biomarkers are neither necessary nor sufficient for a diagnosis. There is disagreement over whether (1) or (2) is the most appropriate definition for Alzheimer's disease; for an overview, see Daly et al. ⁵ In this paper, we will refer primarily to (2), though distinguish this from (1) where appropriate.

Efforts to reduce the scale and impact of dementia have led to a growing emphasis on brain health, a holistic state and set of processes that can be actively promoted through lifestyle behaviors.^2,6 It is increasingly recognized that the combination of two broad approaches is required to meet the challenge posed by the dementia epidemic. ⁷ On the one hand, there is the whole-population approach to dementia prevention, which focuses on interventions that aim to reduce everyone's risk across society, regardless of individual risk status for dementia. The rationale for this approach is that many of the risk factors for dementia e.g., obesity and physical inactivity, poor diet, high blood pressure, smoking, low education and lack of cognitive stimulation, social isolation and loneliness, are highly prevalent, and to reduce the absolute burden of disease, scalable interventions that target these risks by efficiently applying available resources to a large number of people are required. ⁷ Moreover, it is well established that dementia follows a social economic gradient, and, therefore, culturally-informed societal interventions that tackle the root of inequities are required to reduce its prevalence for all across societies. ⁷

On the other hand, the at-risk individuals approach to prevention seeks to identify individuals at highest risk for developing dementia, in order to deliver focused risk reduction and preventative interventions to high-risk groups. The later approach is highly targeted and research intensive, and promises to deliver biomarkers of incipient dementia that can be used for future population-level prevention efforts. Examples of this approach are international multi-site and longitudinal programs, such as the PREVENT Dementia Programme. ⁸ These programs collect deep phenotyping data, including brain structure and function via magnetic resonance imaging, amyloid-β load through positron emission tomography, clinical and cognitive data, lifestyle information, etc., from at-risk mid-life individuals, who are cognitively healthy at enrolment, and follow them up longitudinally. Such research programs aim to provide essential knowledge on novel, early and sensitive biomarkers of AD neurodegeneration, risk modification intervention targets for early prevention, and tractable mechanisms for targeting therapeutics and designing future clinical trials.^9,10 The mid-life age group is of particular importance to this prevention approach, because it presents a critical and unique window for disease-course altering interventions, before the manifestation of substantial brain damage.

Accumulating studies of cognitively unimpaired middle-aged individuals are showing that modifiable risk factors for late-life dementia (as assessed by aggregate dementia scales, such as the Cardiovascular Risk Factors, Ageing and Dementia (CAIDE) risk score), ¹¹ are associated with worse cognition (e.g., poorer visual recognition, ¹⁰ lower episodic and relational memory ¹² ) and declining brain structure (e.g., lower hippocampal volume, ¹⁰ thinner cortex, larger hippocampal fissure ¹³ ) an estimated 23 years prior to dementia onset. Conversely, stimulating lifestyle activities are associated with better cognition (e.g., episodic and relational memory) in at-risk mid-life populations, and may protect against the risk of sporadic late-onset dementia, such as for those with a family history of AD. ¹⁴ A recent study found that inherited AD risk (i.e., Apolipoprotein E [APOE] ε4 genotype) modulated the association between sex, lifestyle factors and cognition: APOE ε4 + females showed a significant association between higher occupational attainment and stronger episodic and relational memory. ¹² These findings suggest that modifiable lifestyle activities can offset cognitive decrements due to inherited AD risk in mid-life, and support the targeting of modifiable lifestyle activities for the early prevention of AD.

Deep phenotyping studies like the ones described above are designed to provide group-level evidence of the impact of risk-factors but can also generate individual-level information about brain and behavior that may be relevant to research participants. Research has shown that 9 out of 10 people want to know their risk of brain disease, ¹⁵ and that once individuals know their own risk of AD, they adopt health behaviors faster. ¹⁶ However, current clinical guidelines recommend against disclosing individual-level information about non-modifiable risk factors (e.g., APOE ε4 carrier status, amyloid-β in cerebrospinal fluid) to cognitively unimpaired individuals, due to the low predictive value of these risk factors for future dementia. ¹⁵ By contrast, disclosing information about modifiable risk factors to research participants (e.g., those captured in dementia risk scores like CAIDE) has not been discussed in previous recommendations, ¹⁷ despite recent studies showing that these factors are strongly associated with reduced brain health in mid-life.^10,12,13

Modifiable risk factors are critically different from non-modifiable factors, insofar as they are actionable. Studies show that people want information more when they believe it will be useful in guiding their actions, and that actionability is the number one motive that drives people to seek health information. ¹⁸ Here, we argue that disclosing research-derived modifiable risk information can promote current brain health, independent of its individual-level predictive value for future dementia, and thus, ought to be made available to research participants.

Current ethical guidance

Much of the current ethical guidance concerning the disclosure of individual research results has arisen in the context of genetics research. Historically, three criteria have been used to determine whether disclosure of individual research results is required: i) analytical validity (the results are scientifically valid and confirmed); ii) clinical significance (they have significant implications for a participant's health); iii) clinical ‘actionability’ (a course of action to ameliorate or change the clinical course of the disease is readily available).^19,20 These criteria have typically been justified based on the so-called ‘duty to warn’, as well as the duty to avoid causing harm to research participants. Individual research results describing modifiable risk factors for dementia are unlikely to meet these conditions. While these results are scientifically valid, their predictive value at the individual level (i.e., the extent to which modifying a particular behavioral or lifestyle factor will reduce individual risk of dementia) remains uncertain. ² Moreover, the lack of established preventive interventions (i.e., the clinical ‘actionability’ of the results) for dementia diminishes the requirement to make research results available according to standard ethics guidelines.

Even when individual research results do not satisfy the above criteria, disclosure may nevertheless be permissible based on considerations of beneficence and non-maleficence, respect for participant autonomy, and justice. However, these same considerations have also been used to argue against disclosure; for an overview of this debate, see Downey et al. ²¹ As we argue below, considerations of beneficence and non-maleficence, respect for autonomy, and justice do not provide a conclusive case in favor of, or against, disclosing individual-level risk information about modifiable risk factors for dementia. However, if we reframe this information as concerning current brain health rather than future dementia risk, the case for disclosure is much more compelling.

Ethical considerations concerning disclosure of individual research results

Focus on optimizing current brain health

As the above discussion demonstrates, considerations of beneficence and non-maleficence, respect for autonomy, and justice do not provide conclusive justification for, or against, disclosing individual-level risk information, including modifiable risk, for future dementia to cognitively unimpaired research participants.

Therefore, we argue for a different approach to evaluating the disclosure of individual-level modifiable risk factors for AD. Rather than focusing on individual-level disease prediction and prevention, we suggest that the permissibility of disclosure be evaluated based on the potential impact of behavioral and lifestyle changes on current brain health. This aligns with a broader trend in dementia research towards a holistic approach to understanding and treating AD. Since the 1990s, the majority of research into treatments for AD has been based on the hypothesis that the disease is related to the accumulation of proteins in the brain (e.g., amyloid and tau), which in turn impairs cognitive function. Yet, while several drugs have been developed that successfully remove these proteins, they have not led to significant improvement in clinical outcomes, ³⁴ leading to questions over whether the protein-reduction strategy is the best way to treat AD. ⁵ Consequently, there has been an increased emphasis on modifying lifestyle factors for dementia risk reduction and preventative interventions.

The impact of lifestyle factors on any individual's likelihood of developing dementia is uncertain. ² Nevertheless, the disclosure of individual risk information is useful for maximizing brain health throughout the lifespan. For example, even if a middle-aged individual who shows high modifiable risk of future dementia (e.g., based on aggregate dementia risk scores, such as the CAIDE score excluding APOE ε4 status, or scores of lifestyle activities such as the Lifetime of Experiences Questionnaire, ³⁵ that may indicate social isolation, physical inactivity, or lack of intellectual stimulation) never develops dementia, they can benefit from adapting a healthier lifestyle in the present. Research has found that more frequent engagement in physically, socially and intellectually stimulating activities is associated with stronger episodic and relational memory in mid-life individuals with a family history of AD, ¹⁴ and higher occupational attainment is associated with episodic and relational memory in females who carried the APOE ε4 allele. ¹²

The focus on modifiable lifestyle factors for current brain health provides a clear justification for disclosure that is aligned with considerations of beneficence, respect for participant autonomy, and justice. The benefits of individual-level disclosure are not contingent on a reduction in dementia likelihood, and the likelihood of psychosocial harms is mitigated by the fact that individual research results are not framed as the presence of disease, or as deterministically leading to the development of dementia. Moreover, while these results are not necessarily predictive of future disease, they convey meaningful information on which a participant can base decision-making about their brain health. And because they are reasonably simple to convey, additional costs associated with feedback are not likely to restrict feedback only to large studies. Indeed, this might provide underrepresented groups access to information they wouldn’t otherwise receive; insofar as this is an enticement to participate, this might help to attract otherwise underserved participants to research.

Focusing on the impact of risk disclosure on brain health is also consistent with a more participant-centric model of research. Providing individual-level results acknowledges a participant's contribution to research and has been argued to improve transparency and trust in the research enterprise, ¹⁹ which in turn might lead to increased enrolment, and promotion of long-term relationships with research participants. This approach is particularly valuable for establishing more inclusive research with underrepresented groups, for whom relationships of trust with the biomedical community are currently lacking.

When individual research results are returned, the process of communicating them to participants must strive to promote understanding of the meaning, application, and limitations of this information. This requires training of staff on how to deliver this feedback so that the information is given in a clear and accurate way. Studies show that individuals vary greatly in extent to which they desire health information. ³⁶ Therefore, a transparent strategy for communicating results, e.g., via prior discussion during the study consent process, should be part of the research protocol, including what support and guidance will be provided to participants. Participants may need to go through a separate informed consent at study conclusion, prior to receiving results, to ensure that they have not changed their mind about disclosure, and to inform them about new or unexpected findings that were not discussed in the initial informed consent.

While we have argued for the benefits of making individual-level risk information available to research participants, it is important that disclosure of individual-level risk not be construed as a form of ‘victim blaming’. ³⁷ Similarly, disclosure should not overemphasize the impact of individual behavior change on dementia risk, nor under-emphasize the importance of structural factors that may not only impede risk-reducing lifestyle changes (e.g., food insecurity may make it difficult to adopt a healthier diet), but also increase dementia risk (e.g., living in a dense urban environment may increase exposure to pollution).^7,32,33,38 The whole-population approaches mentioned above aims to circumvent these structural barriers, and while an in-depth discussion of these is outside the scope of this paper, we emphasize the value of combining the ‘at-risk groups’ and ‘whole-population’ approaches for addressing the scale and impact of dementia across the globe.^7,38

Conclusion

In summary, when viewed from the perspective of disease prediction and prevention, it is difficult to assess whether making individual-level, modifiable risk information available to participants is ethically justified. This is due in large part to the fact that the impact of lifestyle factors on dementia risk at the individual level remains uncertain. By contrast, when viewed from the broader perspective of promoting current brain health, making this same information available is more clearly justified. Thus, our argument is primarily about how risk information is framed and communicated to participants, should they choose to receive it. At the same time, while disclosing modifiable risk information might lead to the adoption of lifestyle behaviors conducive to brain health, it is important that individual-level brain health promotion be seen as complementary rather than opposed to population-level approaches.