An online shared decision-making intervention for dementia prevention: A parallel group randomized pilot study

Groups

The intervention delivered as treatment as usual (TAU) was created to be a reasonable representation of what many older persons face in finding out about the importance of brain health and making lifestyle choices if they choose to try to improve it. A great deal of information is available online and in popular books, and authors have proposed a number of programs for brain health (16-18) that recommend multiple activities or interventions with little guidance on prioritizing among them or strategies for behavior change. The TAU intervention mirrored this situation, providing participants with information resources from which to choose activities and methods of behavior change but without specific guidance on activity or behavior change strategies.

The intervention based on the cognitive behavioral shared decision making model, or CBDSM (15), included the same psychoeducational component as delivered to the TAU group, but the individual online sessions were structured to be similar to sessions of cognitive behavioral therapy (19), with an emphasis on collaborative goal setting, education in self-monitoring, and problem solving. The model-based intervention aimed to help older adults make evidence-informed choices about lifestyle factors that could help reduce their risk for cognitive decline and dementia. Grounded in the cognitive behavioral model of behavior change, it draws on a shared decision-making approach to working with older adults as healthcare providers act as consultants and facilitators of goal choice while taking individual patient needs, preferences, and circumstances into account. The model specifies three core elements for the intervention: (a) psychoeducation on risk and protective factors organized according to three hypothesized mechanisms by which they affect brain health, (b) additional education on evidence-informed approaches to goal choice, and (c) instruction in behavior change strategies with ongoing support.

The three mechanisms by which activities are hypothesized to affect brain health are drawn in part from Livingston et al. (20), updated to add brain stimulation via neurohumoral effects as a mechanism (15). This organization simplifies and facilitates psychoeducational efforts and helps older adults understand how to make evidence-informed choices about lifestyle changes. Education on how to evaluate medical information, such as the interpretation of clinical vs statistical significance and common risks indicators, such as odds ratios, gives older adults the tools they need to understand research findings and make informed choices. Finally, synthesizing the shared decision making and cognitive behavioral models is apt given both approach’s emphasis on psychoeducation and collaboration in developing goals and pursuing behavior change.

Study design

All study procedures were completed remotely, by telephone, video conferencing or online data collection. Persons interested in the study contacted the investigators either by telephone or e-mail and were then screened by telephone using a standard script that assessed whether they met inclusion criteria. Those who were deemed potentially eligible were then asked to complete an assessment of their ability to participate in the online intervention by asking them to receive and open an e-mail with a link to a video conference and to open a link to an online questionnaire from the REDCap data capture software (21).

Participants who met these criteria were sent the informed consent document for review. In a second video conference, one of the investigators conducted the informed consent interview. Participants who chose to be in the study were asked to affirmatively indicate their consent for participation by opening another online questionnaire, checking a box on a form indicating their willingness to participate, and by providing an electronic signature. They were then asked to complete a baseline assessment comprising self-report questionnaires delivered by REDCap and a cognitive assessment using the CogniFit® (New York: Cognifit, Ltd.) platform. They were then scheduled for the first intervention session. These procedures were completed under a protocol approved by the Institutional Review Board of Nova Southeastern University.

Participants

Participants were recruited from a local lifelong learning program that provided programming online during the pandemic. Several persons were referred by other participants and from a local religious organization. Inclusion criteria included being 50 years of age or older, having an interest in improving their brain health, and a willingness to participate in the 12-week study. Participants were required to have an e-mail address at which they could receive study communications and links to video conferences, and were required to be able to use videoconferencing software (San Jose, CA: Zoom Media).

Measures

Measures used in this study included a variety of self-report measures assessing risk factors for cognitive decline and dementia, current health and emotional status, and cognitive functioning. At the conclusion of the study, participants completed a questionnaire that evaluated their views of the intervention with respect to its usefulness and usability and an interview that allowed them to provide feedback about it.

Participants completed a daily questionnaire assessing daily brain activities, the Cogstim index described below, such as exercise, healthy eating, and cognitive activity during weeks 1, 6, and 12 of the study. Finally, they also completed a second follow-up interview six weeks after completing the study.

Primary Outcome Measures

The Australian National University Alzheimer’s Disease Risk Index (22) is a measure developed to allow an evidence-based assessment of individuals’ risk for developing Alzheimer’s disease. It includes assessments of demographic factors such as age and gender as well as education, medical history and status, depression, and cognitive and social activity. It has been validated on several cohorts of older adults (23).

The Memory Self-Efficacy Scale (24) is a brief scale distilled from a longer scale designed to assess older persons’ self-report of memory difficulties. It correlates well with the original scale and was independently validated in relation to list and text recall tasks.

The Dementia Knowledge Assessment Scale (25) is a 25-item measure developed to assess knowledge of dementia. It has been subjected to psychometric evaluation and found to have several subscales (26). In this study, we used its seven item Risk and Health Promotion subscale to assess participants’ knowledge of risk factors for dementia and strategies for improving their brain health.

Secondary Outcome Measures

In addition to the primary outcome measures, participants completed several other self-report measures, including a measure of mood, the Center for Epidemiological Studies--Depression scale or CES-D (27), the 10-item version of the Perceived Stress Scale, or PSS (28), and the Cognitive Functioning and General Life Satisfaction scales drawn from the Patient-Reported Outcomes Measurement Information System, or PROMIS (29).

Cogstim Index

Given increasing interest in using apps to track health status and behaviors (30), as part of this pilot study we sought to test a simple questionnaire that could be used to monitor brain health activities on a daily basis. The availability of this sort of questionnaire would allow interested individuals to track their ongoing efforts to develop brain healthy lifestyles over time, potentially enhancing their motivation to change their lifestyle and maintain those changes (31-33).

We developed a brief set of questions to enable daily tracking of cognitive training, other mentally stimulating activities such as reading or working puzzles, physical activity, and diet. Questions on cognitive training, mentally stimulating and physical activity simply ask the person whether they completed any of those activities. If so, the questionnaire was programmed to ask how many minutes were devoted to the activity and for a rating of the activity’s intensity. Scores for each activity were calculated as minutes multiplied by intensity. The items proposed by Sofi (34) were used to assess level of adherence to the Mediterranean diet. They were used because of their brevity and strong evidence base. As the questionnaire was based on a previously reported model emphasizing cognitive activity, brain stimulation, and reducing inflammation referred to as the Cogstim model (15), this index is referred to here as the Cogstim Index.

Technology Acceptance Model Questionnaire

To assess participants’ views of the intervention, after completing the study they were asked to complete a questionnaire based on the Technology Acceptance Model (35). This scale provided estimates of participants’ perceptions of the usefulness, ease of use of the intervention, and likelihood of using it again. We hypothesized that if their views were positive, their average ratings of the intervention would be significantly more positive than the center rating point of a seven-point Likert-type rating scale.

Interviews

Semi-structured interviews were completed during the week after the 12th session and again six weeks later. Participants’ responses to specific questions, such as whether they enjoyed the study and found it beneficial, were recorded. Open ended questions included, for example, “What did you like about the study?” and “How could we make it better?” Participant responses were recorded verbatim and entered in a spreadsheet for tabulation.

Cognitive Training

Participants were given the opportunity to complete unlimited cognitive training using the commercial CogniFit® online software. This software was chosen because of substantial evidence showing that it is effective in improving cognitive function in older adults (36, 37) and because it could be used online by participants rather than in person. CogniFit® was also used for initial and follow-up assessments of cognitive functioning in several domains; the program provides an individually tailored training program based on participants’ initial assessments and ongoing progress. The program targets various cognitive functions, including attention, working memory, verbal and visual memory, and eye-hand coordination.

Procedure

After obtaining informed consent, participants were e-mailed links to a series of online questionnaires and the cognitive assessment with the request that they complete both prior to the next scheduled visit. After completion of the baseline assessments, participants were scheduled for the first weekly session, an introductory overview of the program and goal setting. During this session, the Treatment as Usual (TAU) group received a general goal setting worksheet and were encouraged to choose goals to work on during the study. They were not otherwise encouraged to set specific goals in order to simulate what we hypothesize is a typical situation in which older persons are provided with information and encouraged to change their behavior without specific shared decision-making guidance or behavior change support.

Both groups were oriented to the CogniFit® software program and advised of its availability for use. This was to provide an estimate of the extent to which older adults, given a choice, would engage in computer-based cognitive training.

The CBSDM group received an overview of the model and were encouraged to choose activities from one of the three hypothesized brain health promoting mechanisms. They also received an introduction to SMART goal setting (38) and a structured goal setting worksheet. They were asked to develop at least one goal prior to the next weekly session.

In subsequent sessions, the investigator presented a 10-15-minute PowerPoint presentation on one of several topics, including risk and protective factors, meditation, sleep, stress management, exercise, cognitive training, Mediterranean diet, sense of purpose, social engagement, and mentally stimulating activities. Presentations usually included presentation of a relevant research study followed by suggestions for behavior change based on the study’s findings. For both groups, the final session was a review of all presentation topics and discussion of follow-up strategies.

For the TAU group, each session began with a general orientation to the topic, proceeded to the presentation, and concluded with answering questions and any discussion raised by the participants. For the CBSDM group, each session began with an agenda-setting slide in which the investigator proposed to review goals and progress and then to the topic presentation. The session concluded with a review of goals and, if needed, problem solving about progress toward goals.

As originally envisioned, the intervention was to have been delivered in small groups, but during the study it became clear that group scheduling would delay completing the pilot study unacceptably. Throughout the study the intervention was delivered one-on-one with one of the investigators (RLO).

After the 12 weekly sessions, participants again completed self-report measures and the cognitive assessment. They completed an exit interview based on a semi-structured format eliciting their views of the intervention, what they liked or disliked about it, and any suggestions for improvement. They were contacted again six weeks later again to be asked these questions and about the extent to which they had achieved their goals and maintained gains.

Secondary Outcome Measures

For the PROMIS Life Satisfaction scale, effects for treatment, and the interaction of time with treatment group were not significant. The effect for time, however, approached conventional levels of significance and represented a large effect size (Figure 3; χ² [1] = 3.37, p = 0.07, d = 1.03 [95th CI -0.07-2.14]). For the PROMIS self-report of Cognitive Function scale, no effects were significant. Inspection of the interaction plot suggested improvement in both groups during the study that represented a large effect size (χ² [1] = 1.91, p = 0.16, d = 0.83 [95th CI -0.35-2.01].

• Download figure
• Open in new tab

Figure 3. PROMIS Life Satisfaction Scale

TAU = Treatment as Usual; CBSDM = Cognitive Behavioral Shared Decision-Making Model

For the Center for Epidemiological Studies – Depression Scale, effects for treatment group and the interaction of group with time were not significant, but the decline in depressive symptoms during the study approached significance and represented a large effect size (χ² [1] = 3.39, p = 0.07, d = 1.19 [95th CI -0.08-2.45]. For the Perceived Stress Scale, no model effects were statistically significant, although inspection of the interaction plot suggested a decrease in both groups that represented a medium effect size (χ² [1] = 0.79, p = 0.37, d = 0.46 [95th CI -0.55-1.46].

Cogstim Index

ITSA analyses for the overall questionnaire (summing scores for all domains) suggested increases in brain healthy activities for both groups in the second half of the study (weeks 6-12; t = 3.12, p = 0.004 for the CBSDM group and t = 2.73, p = 0.01 for the TAU group). For the physical activity score (Figure 4), the CBSDM group showed significant increases in the second half of the study (t = 2.44, p = 0.02) but the TAU group did not (t = 0.89, p = 0.38). The difference between the two groups, however, was not significant (t = 0.40, p = 0.69).

• Download figure
• Open in new tab

Figure 4. Interrupted Time Series Analysis for Physical Activity in Weeks 1, 6, and 12

CBSDM = Cognitive Behavioral Shared Decision Making Model

Cognitive Assessments

Evaluation of training effects on the Cognifit® cognitive measures before and after the study showed, in general, evidence of modest but nonsignificant improvements in cognitive domain scores. When time spent training was included in models, however, the effect of time approached significance and represented a large effect size (e.g., for attention, χ²[1] = 2.88, p = 0.09, d = 0.93 [95th CI -0.14-2.02]; for working memory, χ² [1] = 2.46, p = 0.12, d = 1.07 [95th CI -0.04-2.19]). This finding suggests that time spent in cognitive training had a positive impact on subsequent performance on cognitive measures.

Results of unplanned comparisons of those who completed at least one hour of cognitive training compared to those who did not were consistent with this interpretation of the impact of training on posttest scores, with a significant time by group effect (trained vs. not) on the motor coordination subtest that was associated with a very large effect size (χ² [1] = 9.57, p = 0.002, d = 2.44 [95th CI 0.89-3.98]). This finding supports the efficacy of training on this skill for those who engaged in one hour or more of training.

Acceptability and Usability

Participants’ ratings of the intervention on all TAM subscales of the Technology Acceptance Model scale were significantly different from the scale’s midpoints, with ratings of Usefulness, Ease of Use, and Intent to Use subscales in the future representing very large effect sizes (d = 3.50 [95th CI 1.94-5.04], 2.56 [95th CI 1.35-3.74], and 1.69 [95th CI 0.78-2.58, respectively). No between-group differences were significant.

Interviews

In interviews completed during the week after the intervention, all participants stated that they found the program helpful (n = 16) and that they enjoyed it (n = 16). All but one indicated that they would do it again. One participant said, “the program was extremely valuable,” while another stated, “I was made keenly aware of the need to keep my mind active.” Several indicated that they found weekly sessions helpful, especially the structure of linking discussions of topical presentations with specific research studies ensuring that there was an empirical foundation for lifestyle recommendations.

Although many of the participants indicated that they were aware of the information presented in the weekly sessions, they noted that hearing the topics again reinforced for them the importance of a brain healthy lifestyle and helped them maintain a focus on making lifestyle changes, as in “taking concrete steps” to develop better health habits. Finally, several of the participants in the CBSDM model group noted that the structure of the sessions, with goal setting and follow-up was helpful to them in initiating behavior changes.

In interviews six weeks after intervention completion, participants continued to state that they had found the program helpful and that they enjoyed it (16 of 16). When asked if they had continued to work on their goals, 15 of 16 stated that they were continuing to work on changing their behavior. Four participants indicated that they had achieved their brain health goals, two said they had not, and 10 said they had partially achieved their goals. Fourteen of 16 stated they would do the program again.