Preliminary Development of a Multidomain Behavioral Intervention for Primary Prevention of Dementia in Midlife: A Qualitative Study of Focus Group Discussions

Abstract

Objectives:

45% of global cases of Alzheimer’s disease and related dementias (ADRD) are attributable to 14 risk factors, but population adherence to protective health behaviors is low. The goal of this study was to use qualitative methods to develop a personalized health education intervention (TEACH: Tailored Education for Aging and Cognitive Health), grounded in the Health Belief Model.

Methods:

Guided by a phenomenographic approach, we conducted six focus groups (N=3-6 participants each) of midlife adults, presenting fictional health information including images conveying ADRD risk and descriptions of personal health belief factors. Participants provided feedback about the images, their understanding of ADRD risk, and interpretation of health belief concepts. Qualitative data were analyzed using framework matrix analysis.

Results:

Groups found the presented risk information to be understandable and relevant to health behaviors. They provided feedback on visual aids and alternative language to improve clarity of the health belief descriptions. Participants were able to identify connections between health beliefs and personal behaviors.

Conclusions:

Results guided the TEACH intervention development project, including creating an explanatory framework for educating participants about ADRD risk factors and health beliefs. The long-term goal is to test a multi-domain intervention to promote sustained health behavior change for primary prevention of ADRD.

Introduction

The 2024 report of the Lancet Commission on dementia prevention stated that 45% of dementia cases globally may be attributed to 14 modifiable risk factors, including depressed mood, physical inactivity, diabetes, smoking, and excessive alcohol consumption.¹ Correctly timed interventions to modify these risk factors could substantially reduce the global burden of Alzheimer’s disease and related dementias (ADRD).² In addition to robust observational research demonstrating associations between health behaviors and dementia risk,³ there is a rapidly growing body of randomized controlled multidomain intervention trials for primary prevention of dementia. One of the best-known is the Finnish Geriatric Intervention Study to Prevent Cognitive Impairment and Disability (FINGER) trial, which randomized 1,260 older adults to a multidomain intervention that included diet, physical activity, cognitive activity, and vascular disease management versus a control healthy living education condition. After two years, participants in the multi-domain intervention had 25% larger improvement on cognitive composite measures than controls.⁴ A U.S.-based replication study, U.S. POINTER, recently found that a structured multidomain intervention improved cognition in older adults at risk of cognitive decline⁵. Other multidomain lifestyle interventions have had more modest effects,^6-8 and multiple large trials have shown no significant effects.^9-11 Despite these mixed findings, there is significant interest in multidomain behavioral interventions for primary prevention of dementia.

A significant weakness in existing interventions for dementia prevention is poor adherence (only 19% adherence in the FINGER intervention,¹² with those with low adherence not differing from controls in their cognitive outcomes¹³), which is particularly problematic because protective health behaviors must be sustained for years or decades to meaningfully reduce late-life risk. This raises concerns about scalability for multidomain interventions such as FINGER or U.S.-POINTER, which include multiple active coaching components delivered over several years.^14,15 Moving interventions into midlife, when risk factors have the most significant impact on late-life cognitive outcomes,¹ poses additional challenges for long-term behavioral maintenance. Few multidomain interventions have focused on this midlife period, with most enrolling cognitively unimpaired older adults or people with mild cognitive impairment.

The objective of this U.S. National Institutes of Health (NIH) Stage I behavioral intervention development study was to use qualitative methods to develop a theoretically grounded, personalized multidomain health behavior intervention (TEACH: Tailored Education for Aging and Cognitive Health) targeted to middle-aged adults. Crucially, TEACH is designed to shift from resource-intensive, supervised lifestyle programs to an upstream behavioral model targeting the psychological mechanisms that support long-term lifestyle modification. Built on an existing psychoeducation program¹⁶ and grounded in the Health Belief Model¹⁷ using constructs from the NIH Science of Behavior Change Research Network,¹⁸ TEACH will target health beliefs (i.e., perceived threat of disease, perceived benefits of change, perceived barriers, and self-efficacy) as mediators of health behavior change. The Health Belief Model was selected because it specifically targets how individuals weigh long-term health threats against the immediate cost of lifestyle change.

In our prior quantitative work, health belief factors including perceived future time remaining in one’s life, self-efficacy, ability to defer gratification, and consideration of future consequences were associated with lower ADRD risk and higher engagement in brain health behaviors.¹⁹ This supports our hypothesis that these intrapersonal health beliefs are important proximal targets for intervention. While prior qualitative research confirms that the public is receptive to psychoeducation about dementia risk reduction, knowledge gaps persist, particularly regarding metabolic and vascular risk factors.²⁰ Furthermore, individuals often view personalized risk profiling as empowering but also express apprehension about the psychological burden and potential distress of receiving a “high risk” result.²¹ This underscores a critical translation gap addressed by the current project: developing a framework for communicating ADRD risk paired with information about intrapersonal health belief factors that can be used to engage in risk reducing behaviors.

Our central hypothesis guiding the development of the TEACH intervention is that educating participants about their personalized ADRD risk will increase their perceived threat of dementia, while simultaneously assessing and targeting intrapersonal health beliefs will lower perceived barriers and increase self-efficacy. The proposed mechanism of action is that by modifying these proximal cognitive and behavioral targets, TEACH will successfully translate risk awareness into sustained, long-term adherence to protective health behaviors. While these quantitative findings established that these intrapersonal health beliefs are associated with ADRD-relevant health behaviors, several critical translation gaps remained that prompted this phase of intervention development.^22,23 Specifically, we sought to determine 1) how to translate complex constructs from the Science of Behavior Change (e.g., delay discounting, response inhibition) into layperson-friendly language; 2) visual aids to clearly communicate this personalized information; and 3) how midlife adults conceptualize relationships between these intrapersonal health belief factors and their daily habits. We therefore conducted focus groups to develop an effective framework for communicating health information to adults in a way that was understandable and applicable to their health.

Methods

Study Design and Intervention Development Framework

We adopted a qualitative research design using semi-structured focus groups to obtain an intensive, multi-perspective summary of end users’ experiences and language to inform intervention material refinement. This work is guided by the Person-Based Approach²⁴, which prioritizes iterative end user feedback to maximize intervention acceptability and feasibility prior to efficacy testing. This initial step of the broader TEACH project is situated within Stage 1A (intervention generation and refinement) of the U.S. National Institutes of Health (NIH) Stage Model for Behavioral Intervention Development.

Participants

Participants were recruited from the Rhode Island Alzheimer’s Disease Prevention Registry through Rhode Island Hospital and from the surrounding Rhode Island community between February and April 2023 via convenience sampling. The participants satisfied the following inclusion criteria ascertained by a screening phone call: a) being between the ages of 45-69 years old; b) being cognitively unimpaired (Minnesota Cognitive Acuity Scale > 52);²⁵ and c) being fluent in English. These criteria were chosen as this is the target population for a planned future pilot trial of the TEACH intervention. All study procedures were approved by the Rhode Island Hospital institutional review board (approved protocol #1895972). Participants completed informed consent procedures individually with a research staff member before joining the focus group. They were compensated for their participation.

Focus Groups

Six focus groups consisting of 3-6 participants were held in person at Rhode Island Hospital or community locations (e.g., senior center). Small group sizes (mini-focus groups) were selected to facilitate deep, highly interactive discussion of abstract health belief constructs and visual risk data. The total sample of N=6 groups is consistent with qualitative standards demonstrating that thematic saturation is typically achieved within 3-6 focus groups in relatively homogeneous samples.^26,27 No participants refused to participate or dropped out. Each focus group was approximately 90 minutes long. Focus groups followed a standard discussion guide, including probes to explore and seek clarification. Each focus group was attended by two female study investigators who are clinical neuropsychologists and experts in ADRD (JDD, LEK); one served as a facilitator to provide an overview of the group discussion and present questions and follow-up probes. They did not have prior communication with participants. A research assistant was present to record field notes and non-verbal communication.

During each focus group, participants were told that the overall goal of the study was to develop a brain health program for ADRD and to learn how to communicate about relevant risk factors and personal traits. Participants were presented via PowerPoint slides with health information about a fictional, gender-neutral person (“Taylor”). This included two primary domains of health information: 1) Taylor’s personal risk for ADRD and 2) their personal health beliefs. Regarding personal risk for ADRD, participants were shown four images meant to convey Taylor’s degree of ADRD risk for 14 modifiable (e.g., diet, physical activity, social network) and four non-modifiable (e.g., age, sex, family history) risk factors (Figure 1). These were adapted from the Lancet Commission on Dementia Prevention¹ and cross-referenced with domains assessed by the Australian National University Alzheimer’s Disease Risk Index (ANU-ADRI²⁸), a well-validated self-report tool. Minor modifications were made to ensure the risk profile was suitable for individual behavioral intervention; specifically, we excluded macro-environmental factors like “air pollution” (a Lancet Commission risk factor) that are not directly modifiable by individual behavior change, and included “pesticide exposure” as an actionable environmental exposure assessed by the ANU-ADRI. Structured questions assessed participants’ understanding of the risk images, perception of which risk factors Taylor most needed to change, and opinions about the images used to convey the risk information.

Figure 1.

Images conveying a fictional person’s risk of Alzheimer’s disease and related dementias, including non-modifiable and modifiable risk factors.

Regarding Taylor’s personal health beliefs, participants were shown descriptions of seven health belief constructs: future time perspective,²⁹ delay discounting,³⁰ deferment of gratification,³¹ consideration of future consequences,³² response inhibition,³³ executive control,³⁴ and self-efficacy³⁵ (Table 2). Each of these constructs corresponds to a measure from the Science of Behavior Change Research Network.¹⁸ Each focus group was shown information about four of the seven constructs, with the order counterbalanced across focus groups so each health belief construct was discussed at least three times. A layperson-friendly definition (8^th grade reading level or below) was presented for each construct. Participants were asked structured questions to assess their understanding of the definitions, elicit alternative words or phrases to describe the constructs, draw links between the health belief constructs and Taylor’s health behaviors, assess the perceived relevance of the construct in their own lives, and examine perceptions of the constructs as fixed versus changeable. They were also shown three different images to convey Taylor’s level of each health belief construct (Figure 2). Structured questions assessed their understanding of the images and whether the images facilitated understanding of the written definitions of each construct.

Table 2.

Descriptions of health belief constructs.

HBM Domain	Construct	Layperson-Friendly Description
Perceived threat	Future time perspective	Future time perspective refers to how much time you believe you have left in life and how you would like to spend that time.
	ADRD risk	N/A (conveyed by ADRD risk images)
Perceived benefits and barriers	Response inhibition	Response inhibition is your brain's ability to manage impulses.
	Delay discounting	Delay discounting measures how much you want something right now, even if you could hold off for something better later on.
	Executive control	Executive control is how well your brain can focus on a goal while ignoring things that could distract you from your goal.
	Deferment of gratification	Deferment of gratification is your ability to put off something you want now in order to wait for something even better in the future.
	Consideration of future consequences	Consideration of future consequences refers to how you think about the long-term consequences of your actions.
Self-efficacy	Self-efficacy	Self-efficacy is your belief in your own ability to reach a goal.

Note. HBM: Health Belief Model. ADRD: Alzheimer’s disease and related dementias.

Figure 2.

Images conveying fictional individuals’ levels of health belief traits.

Data saturation was assessed by team review of each focus group session to identify emerging themes. Saturation was considered reached when no new themes or variation in responses emerged in successive focus groups. This approach assured that the number of focus groups was sufficient to capture key perspectives to inform intervention development.

Qualitative Analysis

Each focus group session was digitally audio-recorded and professionally transcribed. All identifiers were scrubbed from the transcripts. Transcripts and field notes were reviewed by two members of the research team who are experienced in conducting focus groups and working with qualitative data (RKR, IA). They reviewed data independently. To minimize bias, the research team engaged in ongoing reflexivity, used a semi-structured guide with neutral prompts, and used investigator consensus meetings during framework analysis to resolve discrepant data interpretations. Transcripts were not returned to participants for comment, and participants did not provide feedback on the findings. Data were charted into a framework matrix for analysis using Microsoft Excel.^36,37 One rater (RKR), a social scientist/medical anthropologist, charted the data into a matrix by summarizing participant comments to each of the major questions posed in the focus groups. A framework matrix is a process for qualitative data reduction commonly used in health services research. This approach is particularly appropriate because we required an aggregated descriptive summary of participant responses to the images, risk factors, and health beliefs reviewed in the focus groups in order to design intervention material. A meta-summary of all individual responses was included. Responses were reviewed by the entire research team, which included three clinical psychologists and a social scientist/medical anthropologist. Major trends were identified to determine clarity of health information provided, alternative ways of displaying and summarizing information, and preferences for explanatory images. A research assistant took notes on team decisions to ensure credibility and reproducibility of data analysis.

Results

Participant demographics are reported in Table 1. No participants withdrew from the study after enrollment. Overall, focus groups demonstrated the understandability and relevance of information about ADRD risk and health belief constructs to non-expert participants. Participants expressed clear opinions and reactions to presented information, including both positive and negative feedback on the method in which the information was conveyed. They were able to describe links between health belief factors and behaviors such as dietary choices or physical activity. Participants noted some areas of overlap between the health belief constructs, though subtle distinctions between the concepts were observed. Many expressed a desire to learn their own personal health information after seeing the information provided about a fictional person.

Table 1.

Focus group participant demographics

	Group 1 (N=4)	Group 2 (N=3)	Group 3 (N=3)	Group 4 (N=6)	Group 5 (N=6)	Group 6 (N=4)
Age range	68-69	58-68	47-63	51-69	57-69	52-69
Education (% postgraduate)	25%	67%	67%	0%	33%	50%
Sex (% women)	50%	67%	100%	67%	67%	75%
Race
White	75%	100%	100%	100%	67%	100%
Black/African-American	0%	0%	0%	0%	17%	0%
Asian/Pacific Islander	0%	0%	0%	0%	17%	0%
American Indian/Alaska Native	0%	0%	0%	0%	0%	0%
Multiracial	25%	0%	0%	0%	0%	0%
Hispanic/Latino (%)	0%	0%	33%	17%	0%	0%

ADRD Risk Factors

Most participants were aware that some aspects of ADRD risk are modifiable. Diet and alcohol use were the most cited factors that were previously known, with some participants also noting physical inactivity as a known risk factor. Participants were less aware of social network, history of traumatic brain injury, and pesticide exposure as modifiable ADRD risk factors.

Some participants expressed that listing education as a non-modifiable risk factor was confusing, because people can seek additional education throughout the lifespan. For example, this participant contrasts education level (8^th grade vs. college) with a person’s level of relevant health knowledge:

"The only one that surprised me was education. I wasn't sure how that would actually fit into the picture. And I mean…if someone with an eighth-grade education would do worse than someone with a college education? Are we talking about that? Or is it education about health that you want to convey to people?"

[197]

ADRD Risk Images

Participants were shown four different ways of indicating “Taylor’s” dementia risk. The first image showed a color bar with risk factors arranged from low risk (green) to high risk (red) (Figure 1A). Most participants understood the color bar image and were able to correctly identify what risk factors the person needed to work on. Participants were generally able to distinguish between the non-modifiable risk factors and modifiable risk factors.

"Well, you have to—at a point, you can’t change family history, age, or sex, but many of the things within the colored rectangle, they are modifiable."

[113]

However, some participants found the rectangular color bar to be confusing. One participant wondered whether the horizontal axis indicated degree of risk while the vertical axis indicated the amount of each risk factor that Taylor had. A few participants found the words placed on the black silhouette of a person's body to be confusing or alarming ("Change it to a color. I like color. Black just seems so—I don’t know—negative" [194] and "The first body looks a little scary -- 'cause it's Halloween-ish." [155]), but this was not a majority opinion. Some participants also wanted more information on the image about how the risk factors were operationally defined as low, medium, or high risk.

The next two risk images showed the person’s degree of risk arranged in concentric circles like an archery target (Figure 1B and 1C). The same information was presented with two color schemes, first with the low risk (green) items in the center of the target and second with the higher risk (red) items in the center. There was some disagreement about which color scheme was preferable, with most preferring higher risk (red) factors to be presented in the center of the bullseye. One participant concluded that the color scheme might depend on what information the researchers most want people to focus on:

"On this target type of visual, my eyes go to the center. Whether it was green or red, so you kinda want it to be the red, unless you want it to be the green. You know, unless you want them to see what they're doing positively. It depends on…which side of the coin you want to look at."

[183]

The final image was a vertical list of modifiable risk factors presented next to the black silhouette of a body, with risk factors arranged by high risk (red), medium risk (yellow), and low risk (green) (Figure 1D). Most participants found this image to be clear and understandable: "This is cut and dry…but it's nice at the same time." [127]

However, some participants disliked the lack of gradient in the degree of risk when it was categorized as low, medium, or high. When asked to vote on their preferred image, 12 participants voted for image #4, seven for image #1, and two for image #2.

Health Belief Constructs

Participants expressed generally strong understanding of each of the seven health belief constructs. They were able to articulate differences between the concepts, though several were conceptually connected (i.e., consideration of future consequences and deferment of gratification; response inhibition and delay discounting), and some participants perceived them as overlapping or duplicative. For some constructs, particularly future time perspective, they had mixed opinions about whether having more of the trait would make a person more or less likely to engage in a health behavior. For example, some participants felt that feeling as though there is a long amount of time remaining in one’s life would be motivating to make healthier choices, whereas others thought that this would decrease their motivation because they would not feel urgency to make behavior changes now. Participants also indicated that there is a situational component to several of these constructs, wherein a person may have a high level of a particular trait but may not always optimally use that trait when making day-to-day decisions about health behavior changes. For example:

“You’re human and you’ve got your impulses, and whichever your impulses are, it’s hard to do it sometimes. But you have to. And then you fight with your mind….And you’re like…I won’t. Nope, nope. No. Yeah, yeah…and then you [have to] do it [again] the next day.”

[163]

Participants also provided alternative language to ensure descriptions were understandable. When asked which construct is most relevant to their own health, each of the seven concepts was considered primary by at least one participant. The most cited constructs were consideration of future consequences, executive control, and response inhibition.

Generalized self-efficacy

The concept of self-efficacy was generally well understood. Many participants suggested the alternative term "self-confidence" as more widely known, though several participants recognized subtle distinctions between self-efficacy and self-confidence.

"Some people have a lot of confidence in themselves, but…are really not competent." [179] "You make a really interesting point there. You can have self-efficacy without having confidence, and vice versa. Self-efficacy isn't really your belief in your ability to reach a goal. It's your ability to reach a goal."

[139]

Most participants expressed that self-efficacy is something that is under one's control and noted that higher self-efficacy scores are likely to translate to greater engagement in healthy behaviors.

Deferment of gratification

Participants found the concept of deferment of gratification to be intuitive. They were able to relate a high, medium, or low score to how a person may or may not engage in health behaviors. Many provided examples about dietary choices as well as physical activity and alcohol consumption. Some participants noted that a person’s ability to defer gratification may fluctuate based on situational factors. Opinions were mixed as to whether deferment of gratification is a fixed trait or malleable. Some people felt that it could be controlled, and others discussed circumstances in which it might not be, such as if a person has a neurodevelopmental disorder such as Attention Deficit/Hyperactivity Disorder.

“Listening to you say that makes me think of those people who have, for example, ADHD…you know, they’re just wired to basically have instant gratification and not deferment. And that’s just not within their control.”

[182]

Consideration of Future Consequences

Participants demonstrated a good understanding of the concept of consideration of future consequences. Several identified parallels with deferment of gratification, though subtle differences between the constructs were noted.

“That word ‘defer,’ to me, makes that a little different…this [consideration of future consequences] is less of an active term to me. Whereas deferment is more of an active type of term. So, I do think that this one is something that you could change because…you’re up in your brain thinking about things all the time. You could consider all sorts of things all the time.”

[113]

Participants also identified ways in which consideration of future consequences would affect health behaviors.

[If a person has a low score] “They would be less likely to follow up and do things to improve our health…they might just ignore whatever their doctor tells them or what other health professionals…’you’re going to die anyway’…I’ve heard that from my relatives. ‘Nobody lives forever, you know. Might as well enjoy myself.’”

[194]

Response inhibition

Participants expressed understanding of the concept of response inhibition, though several noted some overlap with other constructs, including deferment of gratification and consideration of future consequences. One participant noted that a high score on something with “inhibition” in its name might be hard for people to understand. Another participant suggested the phrasing “suppressing negative impulses” might be more understandable.

Participants had mixed opinions about whether response inhibition is a malleable trait. Some people expressed that response inhibition is something that one’s brain simply does and is outside of one’s direct control. Others partially agreed with that position but said that a person can be taught strategies to more effectively use response inhibition in certain situations:

“Everybody’s got impulses, positive and negative…but it’s the ability to suppress those impulses, you know?…You have to get yourself out of that environment that’s going to get you into trouble.”

Delay Discounting

The concept of delay discounting was confusing to some, though not all, participants. Several said that the definition was more ambiguous than the other health belief constructs. The word “discounting” was perceived as confusing by several participants. The interpretation of a high versus low score was also confusing to some:

“So delayed discounting shows how much you want something right now, even if you could hold off for something better later on. So are you saying he had a high score of wanting something right now? Or he had a high score of holding off for something better later on?”

[196]

Alternative language, such as “self control” was suggested. Participants noted overlap between this construct and response inhibition and delayed gratification.

Executive Control

The concept of executive control was well understood by most participants. Again, some similarity was noted between this construct and other health beliefs such as response inhibition. However, some conceptual differences between these were noted:

“Executive control…isn’t really the same as self-control, because executive control is more how your overall brain works and how it executes things…I think response inhibition and executive control are connected in some ways…think of response inhibition as being more of a short-term thing, with executive control as more long-term.”

[179]

“I think of executive control as getting yourself to do and to complete things, whereas response inhibition is getting yourself to not do things.”

[139]

Executive control abilities were considered an important factor in engagement in health behaviors by most participants. Participants differed in the degree to which they believed a person could modify their executive control abilities, with some believing that these differences are hard-wired into the brain whereas others stated that a person could learn strategies to improve executive control.

Future Time Perspective

Comments suggested that there were a variety of responses to this concept. Some participants felt that a high score on future time perspective (i.e., feeling like there is a long amount of time remaining in one’s life) would motivate a person to make health behavior changes. Others felt the opposite, that feeling that there is a large amount of time remaining would decrease motivation to make health behavior changes now.

“See, if I think I have all the time in the world, I don’t feel particularly motivated to have to do anything, ‘cause I always have an opportunity down the road to do it.”

[179]

“If I said, ‘Oh, my God. I only have six more months to live,’ I’m eating everything. I’m going – that’s it…If I said, ‘Oh, gosh, I’m gonna be around for 30 more years,’ I better start working on this now so that I have, you know, 29 really good years…to make sure I make it to the 30.”

[180]

Images to Convey Level of Each Health Belief Construct

Participants were also shown three sets of images to convey the level of each trait a fictional person has. The first set of images showed silhouettes of a person with a low, medium, or high level of blue color (Figure 2A). The second set of images showed a person walking on flat ground (intended to indicate a high amount of a trait that would facilitate health behavior change), a slight hill, or up a mountain (indicating a low amount of a trait that would facilitate health behavior change) (Figure 2B). The third set of images showed a speedometer indicating low, medium, or high amounts of a health belief trait (Figure 2C).

The first and third sets of images (silhouettes and speedometers, Figures 1A and 1C) were understood by most participants. The second set of images was less well understood, with greater disagreement among participants about how to interpret the images and connect them to levels of health belief traits. Some participants inferred that the hiker on the mountain had traits that would pose barriers to health behavior change: “Jordan over there is struggling. Jordan knows he’s got struggles and it’s…gonna be a mountain just to get through life.” [127]. Others viewed the hiker as someone who would be more likely to achieve their goals: “You think of people who are hiking as people who are, you know, determined and they have goals, and they set them, and they’re working hard.” [196]

Multiple individuals expressed that no images were needed to understand the health belief constructs and that the images distracted from the concepts themselves. Some participants expressed very negative reactions to the images. For example, reactions to the first set of images (Figure 1A) included, “It’s a bad graphic.” [196], “They’re kinda creepy.” [180], and “I see blue socks.” [133]. Reactions to the third set of images (Figure 1C) were similarly varied. Some participants commenting that the speedometer removed the human element of the health belief traits: “[You] dehumanized it, though…where’d Taylor go?” [161] and “It’s like a sports car, but we’re talking about a person here.” [140]. However, others found this image the most intuitive to understand: “If we’re going to go with a graphic, I like this one better because it’s not as distracting with people, like thinking about the people. This one is a…gauge.” [133] and “That was better than the bodies [from images 1 and 2; Figure 2A and 2B]” [155].

When asked to vote on a preferred set of images, participants were relatively evenly split, with nine preferring the first set of images (silhouettes), two preferring the second set (hiker), and eight preferring the third set (speedometer).

Considerations for Individual Score Disclosure

Most participants expressed interest in learning their personal scores on the ADRD risk assessment and health belief constructs. One person expressed that learning their personal information would make them feel “empowered.” Multiple participants expressed that they would want emotional support and actionable strategies to manage their ADRD risk. Some participants reflected that their response to a hypothetical disclosure would depend on their scores.

“It would be great if I scored high on everything. But if all of my scores were low I’d be like, ‘Ugh. Again.’ You know? So it depends on how it’s presented…I would need psychological and emotional support, and a big support team if I was told. Especially if you’re already somebody who is not well and not healthy…Not just, ‘you need to do this better.’ [but] how?”

[196]

“I’m assuming if someone’s giving you these results, they’re saying to you, ‘We know the risk factors, but we also know that…if we improve these, that your chances lessen.’…So this is our starting point. Let’s be positive. It’s our starting point, but we can move forward and improve those…if I was all red [high risk], I’d probably go home and cry, but then I’d know I have somewhere to go. So I’m at the bottom, but I can go up maybe, if I was all red.”

[127]

Discussion

These results represent the first step of a Stage I intervention development study for TEACH, an intervention being designed to educate people about modifiable risk factors for ADRD and health beliefs that may affect engagement in relevant health behaviors. Using focus groups, we examined participants’ understanding and reactions to the health information of a fictional character. This included disclosure of modifiable and non-modifiable dementia risk factors as well as health belief concepts that may impact engagement in modifiable health behaviors.

Focus groups demonstrated good understanding of ADRD risk information and found it to be relevant to health behavior which is encouraging as perception of risk is critical for making individual health behavior change per the Health Belief Model.¹⁷ This is consistent with prior research examining health education in the context of ADRD risk, though past work has primarily focused on disclosure of biological risk factors such as APOE ε4 or AD biomarker status. For example, the REVEAL study showed that 63% of participants recalled their APOE ε4 status after one year,³⁸ and those who were at higher genetic risk were significantly more likely to self-report health behavior change.³⁹ Other studies have found minimal or absent health behavior change after ADRD risk disclosure,^40,41 suggesting that knowledge of ADRD risk alone is insufficient to motivate behavior change.

Focus group participants were generally able to draw connections between personal health belief factors and health behaviors. Most individuals were able to generate specific examples linking a high or low score for a given health belief factor to a hypothetical health behavior. Participants described overlap between several of the constructs. For example, consideration of future consequences and deferment of gratification were perceived to similar, as were response inhibition and delay discounting. Several participants described subtle distinctions between these concepts. We have used participants’ own language to refine our definitions of health belief constructs and to develop the explanatory framework for educating individuals about these constructs (Table 2).

One of the health belief factors that generated the most varied responses was future time perspective. Some participants expressed that a higher future time perspective (i.e., more perceived future time remaining) would be a motivator for health behavior change, while others said that this would reduce motivation. Although a recent meta-analysis reported small-to-medium effect sizes between future time perspective and health behavior outcomes, cultural variability was a significant moderator of these effects.⁴² Thus, individual-level associations between future time perspective and engagement in health behaviors may be highly variable and partially dependent on age, gender, and cultural background. Importantly, this does not preclude us from including future time perspective in this type of behavioral intervention. The goal is to build individuals’ awareness of health belief traits and their relationships to health behaviors, regardless of whether an individual perceives a given trait as a barrier or facilitator of behavior change.

Using the feedback provided by the focus groups, we have created an explanatory framework for disclosing individuals’ personal risk for ADRD and educating them about their personal health beliefs. Based on the preferences of most respondents, we will use image #4 (the vertical list of risk factors in red, yellow, and green; Figure 1D) to convey participants' individual ADRD risk. Some participants expressed a preference to convey the magnitude of risk conferred by each risk factor. However, the study team agreed that this may be unrealistic and misleading in practice, as relative risk conferred by each factor interacts with person-level characteristics such as age, sex, socioeconomic status, and medical history. Thus, we will categorize risk factors into low, medium, and high risk and alphabetize the list of risk factors within each category. We have also clarified language used when conveying ADRD risk and health belief information based on participants’ feedback. For example, one theme that emerged was that the inclusion of education as a non-modifiable risk factor was confusing; our protocol for educating people about their results now includes a discussion about this reflecting early life, rather than lifelong, educational attainment. Although several of the health belief constructs were perceived to be overlapping, they were understood by the majority of participants. Thus, we will retain all the measures in the explanatory framework and make decisions about whether to consolidate or eliminate any after the next phase of qualitative data collection and analysis.

The next phase of this study is to disclose individuals’ personal ADRD risk and health belief information to them using our explanatory framework. Individuals will complete a health belief assessment, and a licensed psychologist will discuss their ADRD risk and personal health belief information with them. We will then conduct individual interviews using a phenomenographic method, which focuses on the variations in how people learn and understand concepts. We will examine the acceptability, appropriateness, and applicability of the presented information to participants’ engagement in health behaviors important for ADRD prevention. This will be used to further refine the explanatory framework for our intervention, which focuses on educating people about their health beliefs and how to change their behavior to reduce risk for ADRD. Using the results from this study, we have begun to develop intervention content that incorporates health belief factors into psychoeducation about health behavior changes for dementia prevention. For example, one exercise will ask participants to identify health belief constructs that may facilitate a behavior change and which may be barriers to change. Another will prompt participants to notice these health beliefs in daily life and identify connections to specific behavioral choices, using examples provided by participants in the present study. We will conduct a randomized controlled trial to assess the feasibility of delivering this intervention versus basic health education alone on proximal outcome measures including perceived threat of ADRD, dementia knowledge, and self-efficacy.

The data from this study represents a first step in this Stage I intervention development project. One limitation of this study is the lack of racial and ethnic diversity in our sample due to convenience sampling, including from the RI Alzheimer’s Disease Prevention Registry. We are conducting a follow-up study with a more representative sample that additionally examines the association between social determinants of health, these health belief factors, and health behaviors. We will use the additional qualitative data from this follow-up study to further refine the explanatory framework and address structural barriers to health behavior change. Additionally, while our inclusion criteria spanned ages 45-69, the mean age of our sample was 62 years (range 47-69), resulting in relative underrepresentation of perspectives from adults in early midlife who may experience different health belief dynamics or greater competing demands from employment and family caregiving. Another limitation is that we presented fictional data, rather than individuals’ personal health information. Reactions to personal health information may vary, and this is an important area for future research. We also did not formally assess participants’ health literacy, which may affect their understanding of ADRD risk factors and health belief traits. Additionally, our eligibility criteria were broad. This may limit generalizability, as midlife individuals who seek out a dementia prevention trial may have different health beliefs, higher baseline perceived threat of ADRD, or different information needs than this community-based sample. Finally, although the health belief traits are presented as separate constructs, they are intercorrelated.¹⁹ Future research may consider more detailed behavioral phenotyping of these health belief constructs and their inter-relationships to guide intervention development.

The qualitative insights from these focus groups establish a blueprint for the next phases of TEACH intervention development. Specifically, future work will involve individual, phenomenographic interviews where midlife adults learn their own ADRD risk and health belief data. This phase will be used to refine the intervention protocol with a goal of developing a personalized mapping framework. Rather than prescribing generic lifestyle changes, participants will be educated about their health belief profile, which will be used to inform individualized goal setting and specific behavioral strategies. The TEACH intervention will then be tested in a pilot RCT to evaluate its feasibility, acceptability, and target engagement.

Conclusions

We demonstrated that focus group participants understood images demonstrating a fictional person’s degree of risk of ADRD based on 14 modifiable risk factors. They could also describe differences between health belief concepts and provided example language to make these concepts more accessible to a non-expert audience. They made connections between health belief traits and specific health behaviors. Overall, participants expressed enthusiasm and interest in learning about their own personal risk profiles to lower risk for ADRD and improve their lifestyles. This information will be used to develop an explanatory method for disclosing participants’ personal health information to them as part of the TEACH intervention. Our long-term goal is to develop TEACH as a brief, remotely delivered intervention to increase long-term engagement in ADRD-relevant health behaviors in midlife.

Data Availability Statement

The data are not publicly available due to the nature of the qualitative data and risk that even de-identified transcripts could contain information that could compromise the privacy of research participants.