Perspectives on Communicating Biomarker-Based Assessments of Alzheimer’s Disease to Cognitively Healthy Individuals

INTRODUCTION

Research in Alzheimer’s disease increasingly focuses on identifying risk statuses indicated by biological markers of pathological changes associated with later dementia, notably levels of amyloid-β in cerebrospinal fluid or on PET brain imaging. However, although significant research is ongoing, the clinical validity and value of biomarker information for asymptomatic individuals is still open to question [1]. The communication of biomarker information to people without symptoms of dementia thus presents an ethical challenge for researchers and, should such approaches translate into clinical practice, for clinicians [2, 3]. Routine communication of risk information to people without symptoms in clinical practice is currently not recommended [4, 5] considering the lack of robust individualized prediction [6] and the potential psychosocial harms caused by learning one’s risk status in the absence of effective preventive or therapeutic options [7]. Attention has also been drawn to the potential of broader testing and screening programs associated with earlier stages of Alzheimer’s disease to result in premature diagnoses and over-treatment of age-related cognitive change [8].

Nevertheless, in research settings there are two reasons why it may be acceptable or appropriate to communicate biomarker information related to Alzheimer’s dementia risk to people without symptoms [9]. Some participants will want to know their estimated risk status, and communicating these individual research results may be an expression of respect for their autonomy and a form of reciprocity in some circumstances [10, 11], even if this information is not currently actionable and the exact meaning of risk prediction and specific biomarkers within risk prediction in an individual is unknown. It has been estimated that between 50% and 90% of people are interested in knowing the results of a ‘predictive’ or ‘reliable’ test of their risk of developing dementia, with numbers varying across countries [12–17]. However, when people’s understanding of the current uncertainty and limited generalizability and clinical validity associated with biomarkers increases, their interest in learning about the biomarker wanes [15]. In fields where tests are predictive rather than probabilistic, experience suggests that actual uptake may be lower than anticipated, particularly in the absence of treatment [18]. It is thus likely that levels of public interest in learning Alzheimer’s disease risk are overestimated.

A more compelling reason for disclosure relates to recruitment for clinical trials. Clinical trials are currently ongoing which target pathology in a population of relatively younger, asymptomatic individuals identified as being at ‘high risk’ of developing Alzheimer’s dementia on the basis of family histories, genetic information, and biomarker tests [19, 20]. Within these studies, prospective research participants should know why they have been invited to take part in a study in order to make informed decisions about research participation [21] as part of a transparent recruitment process [22].

There is a small but growing body of evidence related to the impact of disclosure which suggests that learning that one is at elevated risk of developing Alzheimer’s dementia in the absence of symptoms of dementia does not cause significant psychological or social harm [23]. However, although the immediate psychological impact of risk disclosure seems limited, there is an increase in test-related distress, and adverse psychological effects occur in a subgroup of research participants. Results from the Risk Evaluation and Education for Alzheimer’s Disease (REVEAL) studies suggest risk disclosure may have implications in relation to long-term care insurance and possibly to health-related behavior changes [24]. Although not directly comparable, there is also evidence that diagnoses of mild cognitive impairment and mild Alzheimer’s disease dementia are associated with lower quality of life [25]. Recent studies suggest a low risk of psychological harms associated with disclosure of Alzheimer’s disease biomarkers [26, 27].

However, the evidence remains limited: there has been little analysis of the impact of risk disclosure in relation to non-genetic biomarkers and there is sparse evidence on disclosure outside the United States despite geographical variation in interest in knowing [16] and important differences across countries in long-term care provision. Published work on ApoE genetic testing concentrates on individuals who are first-degree relatives of a person with Alzheimer’s dementia [28], who may already perceive themselves to be at increased risk due to this history [29]. Further, there is potentially important variation in the experience of family history, notably between cases in which dementia occurs late in life and close to death rather than earlier in life [30, 31]. In sum, the impacts of biomarker disclosure on healthy research participants remain to be studied. Importantly, the impact of disclosure is also likely to be affected by the manner in which information about amyloid biomarkers is presented to the research participant. Careful attention is needed to determine whether people want to learn information about biomarkers for Alzheimer’s dementia, what this information should include, and when and how it should be communicated [32].

There are no current shared guidelines for communicating amyloid biomarkers, and a small number of descriptions of current practice have been published to date. The most detailed of these are those elaborated in clinical research settings for the communication of amyloid PET imaging to asymptomatic individuals [27, 33] and to those with mild cognitive impairment [34]. Clinical communication models focus on shared decision-making with regard to amyloid testing and standardized language to describe the clinical significance of amyloid biomarkers [35–37].

The disclosure process described by Harkins et al. [33] is the most clearly elaborated and detailed approach within a research setting published in the literature to date. It was developed for initial use in the Anti-Amyloid Treatment in Asymptomatic Alzheimer’s Disease (A4) study [20] (NCT02008357), which recruits participants based on positive results on amyloid PET imaging. It draws on a Delphi consultation with experts in genetic testing and neuroimaging.

The A4 disclosure process starts with education in the form of an information brochure provided in advance of the initial consent visit. This sets out the state of knowledge about amyloid imaging, the range of possible results, and their implications. The information in the brochure is then discussed in detail during the consent process, along with the motivations of potential participants to join the study. Comprehension of the brochure is then assessed, and potential participants are screened for anxiety and depression to establish their suitability to receive amyloid biomarker results. The actual PET imaging takes place at a second visit, and disclosure at a third visit. Prior to the disclosure of results, study staff again ascertain a participant’s willingness to learn the results, their mood, and whether they have had recent life stress. Harkins et al. emphasize the importance here of involving staff who are skilled in communicating and recognizing distress. Disclosure itself occurs using standardized language which reflects that used in the information brochure and is accompanied by written information and with a family member or friend present if desired. Comprehension of the results is then assessed. Participants are followed up by phone three days after disclosure to assess wellbeing, distress, and the impact of disclosure, with a follow-up plan created based on responses. Finally, participants are followed up over the course of the study. The approach described by Burns and colleagues [27] for the Alzheimer’s Prevention Through Exercise (APEX) study (NCT02000583) is similar. Notable characteristics are an emphasis on the probabilistic nature of both ‘elevated’ and ‘nonelevated’ amyloid results, a policy not to share scan results with the participant’s physicians or enter them into the medical record, and follow-up by phone or e-mail at six weeks and six months.

The approach developed by Lingler et al. [34] focuses on the disclosure of amyloid status to people with mild cognitive impairment. Their guidance starts from guidance on genetic counselling and an iterative expert consultation, and was then tested and refined with mild cognitive impairment patients and caregivers. This approach also emphasizes the importance of pre-disclosure counselling and post-disclosure assessments of comprehension and the value of involving family or friends in the disclosure process. It also contributes important points on the content of information—highlighting the possible value of using brain images and the importance of clear graphics, as also increasingly recognized in the communication of genetic risk [38]. Finally, participants in the study by Lingler and colleagues expressed a desire for seamless communication with primary care providers.

The research reported in this paper was conducted to inform the development of the risk disclosure process during recruitment to trials within the European Prevention of Alzheimer’s Dementia (EPAD) project [39, 40], a longitudinal cohort study and clinical trial platform for the secondary prevention of Alzheimer’s dementia. The paper examines the perspectives of individuals with no symptoms of dementia, people with dementia, and caregivers on the communication of Alzheimer’s dementia risk. No biomarker information has been communicated to these individuals, and the paper briefly describes their expectations related to the implications of learning information related to Alzheimer’s disease risk status. The paper then sets out the preferences of participants related to the content, process, and follow-up of biomarker-based risk disclosure in research settings. These preferences are intended not to be prescriptive, but to reflect qualities of Alzheimer’s disease risk communication considered important by those who may be involved in it in the future, and which should be engaged with by researchers and clinicians in the development of disclosure processes.

METHODS

Focus groups were held in London and Barcelona between November 2015 and January 2016, and in Brussels in December 2015. Spanish participants were recruited through the BarcelonaBeta Brain Research Centre (BBRC) of the Pasqual Maragall Foundation [41]. Participants in the United Kingdom were recruited through the Prevent cohort study [42]. The group discussion held in Brussels was comprised of participants recruited from within Alzheimer Europe’s European Working Group of People with Dementia¹ (EWGPWD). The aim of this latter group was to obtain the perspective of people with dementia. This allowed the investigators to explore further the relationship between the diagnostic and risk disclosure processes. The study was approved by ethics committees in both Cambridge (NHS REC Reference 15/EE/0381 and University) and Barcelona.

Four groups were held in both London and Barcelona, three with people with a first-degree relative with Alzheimer’s dementia, one with people without this family history. In total, 48 people participated, 32 in Spain and 16 in the UK. Group sizes ranged from 3 to 10. The participants included 30 women and 18 men, all of whom identified as Caucasian. 34 of the participants (71%) had a family history of dementia. Further details are given in Table 1. The EWGPWD group included 6 people with dementia (4 women, 2 men, median age 64.5) and 4 partner caregivers (2 men, 2 women, median age 66.5). Demographic data other than age were not collected from this group.

The individuals recruited for the focus groups in London and Barcelona were purposively sampled to be indicative of the likely future population for Alzheimer’s disease clinical trials involving risk disclosure. They are research-engaged individuals, who are more likely to have a family history of dementia. While their views may not be representative of the population, they have immediate relevance to questions regarding the development of responsible disclosure processes in research. The participants from the EWGPWD, as members of a pre-existing, established group, represent a convenience sample.

The groups were conducted according to the same protocol at the sites in London and Barcelona (see Supplementary Table 1). An adapted protocol was used for the EWGPWD. Participants were provided with short background information on biomarker-oriented research into Alzheimer’s disease. The structured protocol then asked a series of questions related to 1) this background information; 2) participants’ interest in learning their Alzheimer’s disease risk, and the implications of this information, 3) how this differed for different types of risk information, and 4) preferences related to disclosure. This paper reports focuses on findings related to sections 2 and 4.

The focus groups were recorded and transcribed verbatim. Groups in London and the EWGPWD were conducted in English, while the Barcelona groups were conducted in Spanish and translated into English as part of the process of analysis. A coding framework was developed and used by RM and AD to identify key themes in the transcripts at each site. Data coded to these themes was then shared and used in further iterations of coding and analysis. Following an initial discussion of the themes emerging from the data at each site which identified comparable results, data from both sites was analyzed together.

The investigators identified preferences related to disclosure from initial analysis of the discussions in section 4 of the protocol related to the process of disclosing Alzheimer’s disease risk information. These were collated by RM and AD and written up for participants. This report was circulated by email to the participants in Spain and the United Kingdom for comment and respondent validation [43]. The participants from the EWGPWD reviewed and agreed upon the report in a face-to-face meeting.

The study has a number of limitations. First, the findings relate to expectations of the disclosure process, rather than reflections on experiences of disclosure. As such it involves some elements of conjecture on the part of participants. Focus groups are particularly well suited for these types of questions, as the focus on interaction enables people to draw on both their own experience and that of others as discussions develop. However, focus group methods also have limitations, not least the limited generalizability of the results and the difficulty and limited value of quantifying perspectives which emerge during group interaction. In the current work, it was also necessary to develop the protocol, facilitate the groups, and analyze the data while working across languages and cultures. This required a more structured approach to facilitation and protocol design to enable the production of comparable data than might have been adopted in a single-language setting, and further detailed work is needed to explore cultural differences related to disclosure and attitudes to dementia risk. Nevertheless, the current work provides a set of preferences related to disclosure generated and reviewed by participants which represent a starting point for consideration and further exploration.

RESULTS

The first section presents the results of the group discussions related to why participants would be interested in or willing to learn their Alzheimer’s disease risk status, and what they saw as the potential limitations and concerns associated with doing so. Results from the EWGPWD group discussion are based on slightly reformulated questions as the participants already have dementia. Consequently, in this section, the views expressed by this group should be interpreted not on the basis of how they are affected but rather on how they would have been affected in the past.

Section two presents the preferences related to disclosure processes in research settings.

DISCUSSION

Focus group participants’ discussions of the likely benefits and harms associated with learning information related to Alzheimer’s disease risk emphasize the perceived personal utility of the information. These findings correspond with those raised in similar research on both the hypothetical and actual impact of learning the results of biomarker tests related to Alzheimer’s disease risk [14, 15, 17, 44]. In line with the findings of REVEAL and similar studies, they suggest the short-term psychological impact of risk information may not generally result in harm, but that there may be subgroups who experience significant negative impacts.

The discussions reported here do, however, draw attention to the longer-term impacts of risk information, not least the potential for hypervigilance and long-term monitoring among people who already feel they are at risk of developing dementia. Risk information should thus be considered in terms of its potential contribution to ‘dementia worry’ [45, 46] and the danger of introducing a diagnostic misconception, whereby people who undergo research diagnostic procedures come to see themselves and are seen by others as patients although they would not have received any diagnosis in a standard clinical setting [47]. This is important in considering the implications and assumptions involved in the preferences of research participants related to the disclosure process.

Preferences concentrate on the content of the information, the quality of communication, and the follow-up process. Discussions of content emphasize the importance of attending to the language used in discussions of biomarker test results, and distinguishing between risks and diagnoses. This reflects the potential difficulties which have been highlighted elsewhere [48] associated with using terms such as ‘preclinical Alzheimer’s disease’ to refer to individuals who have no symptoms of dementia. In practice, it can be difficult for investigators to effectively communicate this distinction, and risk statuses do frequently become elided with diagnoses by both people at risk and clinicians [49]. For example, in studies involving the disclosure of mock PET results to people with mild cognitive impairment and care partners, even when researchers clearly stated that PET evidence of amyloid was not synonymous with a clinical diagnosis of Alzheimer’s disease, this relationship was still presumed by recipients [50].

Attention is thus needed to effective communication of both risk and ambiguity. Guidance on risk communication is well developed in relation to genetic testing, where data on relative and absolute risks are more readily available. Research groups have effectively used pictographs and other visual presentations of risk (e.g., video) to help research participants understand relative and absolute risks conveyed by genetic susceptibility markers for dementia [38, 51]. However, the lack of precise information about the risk associated with non-genetic biomarkers of Alzheimer’s disease (notably related to amyloid load), and the timescale to which the risk refers, makes such an approach difficult to implement. Research is ongoing to improve the accuracy of biomarker testing [39, 41, 42]. However, currently, this limitation should be addressed through the development within these studies of information or education materials to accompany disclosure processes [33, 40]. In line with the findings from qualitative work following amyloid disclosure, these may need to address how classifications of amyloid ‘positive’ or ‘negative’ are produced [32], and should also be transparent about the scientific uncertainties associated with biomarkers [52]. However, more work is also needed to inform the communication of scientific ambiguity or epistemic uncertainty [53] related to Alzheimer’s disease, as well as in our understanding of the concept of Alzheimer’s disease itself, given that the majority of individuals die with these pathologies in their brains and that, among the oldest old—the age of greatest risk for dementia—the risk of Alzheimer’s dementia associated with these pathologies is attenuated [54, 55].

This emphasis on uncertainty is important in considering discussions related to providing information on risk reduction. In practice, it appears rare for research participants to use genetic risk information, including that related to Alzheimer’s disease, to change their behaviors to reduce health risks [56]. Nevertheless, some individuals identified as being at higher risk will wish to act upon it [57]. Providing well-evidenced information on risk reduction as part of the disclosure process would recognize this perceived value and may also reduce the harms associated with learning about disease risks without having any actionable options available. However, while there is consistent evidence of beneficial associations between mid-life physical activity and other exposures, healthy aging, and disease outcomes [58], the available evidence about prevention does not yet clearly point to the efficacy of particular activities or behaviors in preventing the onset of dementia [59]. As in the discussion of risk information, communication related to risk reduction should highlight the current state of the evidence and persistent uncertainties, and be updated as further evidence becomes available [60].

Concerns related to expertise and the consistent interpretation of results echo those of participants in relation to amyloid disclosure in mild cognitive impairment [34]. They suggest an education and liaison role for Alzheimer’s disease researchers, healthcare professionals, and Alzheimer’s associations to ensure that the existence of research diagnostic classifications for Alzheimer’s disease [61] and associated technologies is recognized. The focus on education, follow-up, and in-person disclosure highlights the potential benefits of learning from the genetic counselling model, in line with the approach proposed by Harkins and colleagues [4]. In particular, such models highlight the importance of shared decision making and two-way communication [36, 62]. However, it is also important to remain cognizant of material differences between ‘state’ and genetic ‘trait’ information [28], notably the potential for levels of amyloid change over time, and the unclear nature of the relationship between amyloid load and the development of dementia. Importantly, this involves re-emphasizing the limits in the evidence base of biomarker-based Alzheimer’s disease risk information in order to reduce the potential for over-treatment and the reification of a diagnostic misconception [63].

The preferences described here highlight the importance of considering the nature and extent of short and longer-term follow-up after the communication of biomarker status. Involving wider health professionals, particularly primary care clinicians, would enable participant preferences in relation to monitoring to be met, possibly providing reassurance in the context of hypervigilance. However, such monitoring could also intensify the impact of risk information, complicate its relationship with a clinical diagnosis and possibly impact on life insurance and employment status. Regular check-ups with medical specialists would start to blur the clear boundaries between ‘research’ and ‘clinical’ criteria proposed in the 2011 NIA-AA definitions. In such circumstances, it remains to be established that the potential harms associated with unnecessarily treating individuals who may never develop dementia outweigh the benefits to those who would. It would also have resource implications for healthcare systems, although the costs of monitoring, surveillance, and preventive treatment of asymptomatic individuals with increased risks of Alzheimer’s dementia are difficult to estimate accurately [64]. As current research studies will not routinely refer at-risk participants to medical care, it is not expected that major financial effects on healthcare systems will occur imminently. However, research groups should be prepared to discuss and address expectations related to support and future care, rather than automatically introducing such monitoring with unknown consequences for participants, health systems costs, and harms through diversion of resources.

Risk communication should be a two-way process, supported through a clear procedure for dialogue, encouraging and responding to follow-up questions and identifying individual needs and concerns. Recommendations on disclosure emphasize the importance of short-term follow-up by the research team, for instance through a telephone call a few days post-disclosure [33, 34]. A member of the research team should be easily accessible for research participants post-disclosure, for instance by telephone. However, participants in this research also highlight the value of longer term follow-up, in line with the six week and six month contact described in the APEX study [27], and of broader support networks of family and friends [33, 34]. Again, though, this requires a careful approach which avoids reifying a risk status as a diagnosis, and emphasizes that people who learn they are at higher risk do not have any cognitive impairments, and are not and may never become Alzheimer’s dementia patients.

CONCLUSIONS

Interest in learning risk information regarding Alzheimer’s dementia seems common. Often, however, it is dependent on favorable estimations of the predictive value and personal utility of Alzheimer’s disease biomarkers: interest is lower when people learn that biomarkers are not conclusive or informative and options to reduce risk of disease are limited or absent. Respondents’ discussions dwelt more on the long-term implications of risk information and the production of hypervigilance than short-term effects.

Participants’ preferences related to the process of learning risk information underline the need for careful, responsible risk disclosure processes in research settings. They have immediate implications for practice for researchers, clinicians, and Alzheimer’s associations. Importantly, they highlight the fine line and inherent ambivalence in trying to offer an appropriate level of care and support and to avoid an over-interpretation of the clinical significance of biomarker-based risk information. Research groups should be careful to prevent the care and support they offer from reinforcing the diagnostic misconception and resulting in the harms and burdens associated with hypervigilance.

ACKNOWLEDGMENTS

We are particularly grateful to participants in the Prevent study and in the ALFA project, without whom this research would have not been possible. The Prevent study is funded by the Alzheimer’s Society, and the ALFA study receives funding from “la Caixa” Foundation. We would also like to thank the wider group of the Ethical Legal and Social Implications work package of the European Prevention of Alzheimer’s Dementia (EPAD) study and the reviewers for their constructive comments on this paper. The study was funded through EPAD, which receives support from the Innovative Medicines Initiative Joint Undertaking under grant agreement n° 115736, resources of which are composed of financial contribution from the European Union’s Seventh Framework Programme (FP7/2007-2013) and EFPIA companies’ in-kind contribution.

Authors’ disclosures available online (https://www.j-alz.com/manuscript-disclosures/17-0813r1).