The prevalence of dementia is estimated to be between 5% and 7% among those aged 65 years and above [1,2]. Although the rates have fallen slightly in the United Kingdom over the past 20 years [2], the absolute number of cases is likely to increase owing to longer life spans [3]. Over the next 30 years, 10-15 million cases are predicted in Europe [3], and 132 million in the world by 2050 [4], with approximately 43% of prevalent cases needing high-level care (equivalent to nursing home care) [3]. It has been estimated that if interventions could delay both disease onset and progression by one year, there would be nearly 9.2 million fewer cases of the disease worldwide by 2050, almost all attributable to a reduction in persons needing a high level of care [3].

Cognitive impairment in the absence of dementia is common, but prevalence estimates vary considerably depending on the definition [5]. Mild cognitive impairment (MCI) is the most commonly used definition, conventionally defined as a deterioration in at least one nonmemory cognitive domain in addition to memory impairment, without severe functional impairment or loss of independence in Instrumental Activities of Daily Living [3,6]. The incidence of MCI is high in those aged above 65 years [7-9], ranging from 51 to 76.8 per 1000 person-years. There is not enough evidence yet that screening for MCI is either effective or cost-effective [10], but a systematic review of cohort studies suggests that between 5% and 10% of MCI cases progress into dementia annually [11]. However, defining cognitive impairment solely in terms of MCI misses a large group of individuals who are at a similar risk of developing dementia [12-14]. An alternative way of characterizing cognitive impairment is age-associated cognitive decline (AACD). AACD is described as 1 SD below normal cognitive functioning in any cognitive domain, with some investigators considering an additional criterion (self-report of a gradual decline in memory present for at least 6 months) [5]. The estimated prevalence of AACD depends on the number of cognitive domains assessed and whether additional criteria are used, which may explain the variable prevalence estimates—in one population-based study, up to 20% of those aged above 60 years had AACD using the simple criteria described by Ritchie et al [14] and as low as 1.4% in the UK Cognitive Functioning and Ageing Studies, where detailed criteria were used [5]. Progression to dementia is common no matter which definition is used, with rates almost comparable with that of MCI: 9% per annum in a population-based study (additional criterion not used) [14] and 10% in Cognitive Functioning and Ageing Studies (additional criterion used) [12,13]. Importantly, there is currently no diagnostic or treatment pathway for people with AACD. Given the scale of the problem, there are clearly insufficient resources on offer to help prevent cognitive decline or dementia in those with cognitive impairment or to prevent the development of cognitive impairment.

There is mounting evidence that healthy behaviors (particularly physical activity) and cognitive exercises improve cognitive functioning and activities of daily living, and a recent trial from other settings has demonstrated the potential effectiveness of combining healthy behavior and cognitive interventions [15].

Several large cohort studies and systematic reviews indicate that leisure time physical activity—even at moderate intensity levels—is protective, as is greater fish and fruit and vegetable consumption, with increased risks shown for obesity and high saturated fat intake [16-23]. The Caerphilly cohort is representative of the UK population and measures protective factors (not smoking, BMI levels, fruit and vegetable intake, physical activity, and moderate alcohol intake [22]), and after 30 years, it demonstrated markedly reduced risks of cognitive impairment (odds ratio [OR] for having 4 to 5 protective risk factors=0.36 and dementia OR=0.36). A trial of the Mediterranean diet for over four years demonstrated beneficial effects on cognitive decline among those with no cognitive deficits at baseline: changes in cognitive z-scores were 0.04 (95% CI −0.09 to 0.18) for the Mediterranean diet plus olive oil, 0.09 (−0.05 to 0.23 vs controls) for the Mediterranean diet plus nuts, and −0.17 (−0.32 to −0.01) for the control diet [24]. Interventions to increase physical activity are protective in the shorter term [25] with effects even at 6 months to a year on both cognitive decline, gray matter volume, and atrophy [26-28]. The population attributable risk (PAR) has been estimated for diabetes, hypertension, obesity, physical inactivity, depression, smoking, and low educational attainment, with the highest PAR found for physical inactivity (PAR 21.8%, 95% CI 6.1-37.7) [29].

A systematic review of cognitive exercise trials [30] documented a strong effect size in cognitive performance (weighted mean difference=1.07; 95% CI 0.32 to 1.83; n=3194) with effects maintained after 2 years. A more recent systematic review of cognitive and memory training among those with MCI documented substantial heterogeneity of interventions [31] but a promising effect of cognitive exercises (effect sizes ranging from 0.10 to 1.21). The large Advanced Cognitive Training for Independent and Vital Elderly study investigated the effect of training in several cognitive domains (memory, reasoning, and processing speed) [32]; the change in reasoning (effect size=0.26) was particularly important, resulting in significantly less functional decline in activities of daily living, which was maintained over 5 years. Booster training produced additional improvement in reasoning performance (effect size=0.28) [33].

Although there is evidence that suggests healthy behavior changes can be protective against cognitive decline, behavioral interventions are complex and resource intensive, if delivered by purely face-to-face methods. In contrast, the internet is now used extensively and successfully by older people for self-management of health [33,34]. Although many individuals may benefit from using an intervention independently, the additional impact of behavioral facilitation may be important in helping initiate and maintain behavior change [35,36], and the effectiveness and cost-effectiveness of a more intensive intervention may vary with the risk of developing dementia.

The primary aim of this study (the Active Brains study—feasibility trial ISRCTN 23758980) is to assess the feasibility and acceptability of our trial procedures and of a web-based digital intervention (Active Brains), which helps support people in making healthy changes (physical activity, brain training, and diet) to maintain cognitive function and prevent cognitive decline.

The randomized controlled study will use a parallel design. Participants will be divided into cognitively impaired and noncognitively impaired subgroups based on existing cognitive impairment (each subgroup is treated as a separate trial for randomization and reporting). A total of 180 participants from each group will be randomly allocated to 1 of 3 study groups (totaling 360 participants):

It is anticipated that the noncognitively impaired group will recruit more quickly than the cognitive impairment group, due to the low prevalence of MCI and AACD within the age group. Participants exceeding the allocated group size of 180 will not be randomized but will be offered to access the intervention (no support) as part of a cohort group.

Primary care recruitment will involve practice staff inviting patients from searches of UK practice lists based in England (GPs [General Practitioners] will screen practice lists to avoid inviting those who have an existing diagnosis of dementia and who are terminally ill or seriously mentally ill) between October 2018 and January 2019. Participants will also be recruited opportunistically during consultations with practitioners. Invitation letters, participant information sheets, reply slips (for those not interested to inform us why), and a getting started card with instructions on how to log on to the website to start the study will be sent from GP practices. Participants can also contact the research team directly if they have any questions. Following informed consent, GP practices will be asked to provide demographic data (gender, age, and postcode to help us establish the level of deprivation) for all participants who are invited to the study. We will compare the demographics of those invited but who do not participate in the study with recruited participants to examine any differences between these 2 groups. This will help the assessment of the likely generalizability of our findings.

Figure 1 outlines the participant identification and screening procedures. Participants recruited through primary care will be screened to include only participants aged between 60 and 85 years and to exclude persons with a diagnosis of dementia, terminal illness, serious mental illness, or persons within the same household.

When participants first access the Active Brains study website, they will be asked to sign up with a username, password, and their unique ID code (provided on their study invitation letter). They are then presented with the information sheet and web-based consent form to complete. If they consent, they then answer a series of screening questions to ensure that they fit the study inclusion criteria (shown in Textbox 1). Age, dementia diagnosis, and the same household questions will be repeated in the web-based screening. Additionally, web-based screening will include willingness and ability to access the internet, the Godin Leisure Time Exercise Questionnaire (this questionnaire [47] will be used to screen highly physically active participants as it is a short and simple measure, collecting only sufficient data for screening participants and tailoring intervention content. The International Physical Activity Questionnaire [48] will be used to assess physical activity as an outcome measure as it has higher granularity and assesses sedentary behavior, which is targeted in the intervention), the Instrumental Activities of Daily Living Questionnaire (IADL) [49], and 4 cognitive tasks (short computer games that evaluate cognitive skills). If eligible, participants will be asked to complete baseline measures (Table 1).

Subgroups of cognitively impaired participants will be identified by combinations of impairment in the Baddeley reasoning cognitive task, IADL, and memory cognitive task. Key subgroups are MCI, which will be defined as 1.5 SD below the norm in a nonmemory cognitive domain plus memory impairment [5,12,13], and AACD will be defined as 1 SD below the norm for the Baddeley Reasoning Test [50] and IADL.

Simple randomization will be used to assign participants to intervention arms using the study software’s computer-generated random numbers. Randomization occurs online after a participant completes baseline measures, ensuring blind randomization. They will have an equal chance of being in each of the 3 groups. Once randomized, participants will be informed of their group allocation (they will also be emailed this information). If participants are in one of the treatment arms, they will be taken directly to the Active Brains intervention. These participants can then use the Active Brains intervention as much as they would like to over the course of the study. The central facilitator will be notified by email when a participant is randomized to the support arm of the trial, so that they know to expect to provide support to the participant in the coming weeks. Research staff will not be blinded to the participant group but the statistician and health economists conducting the analyses will be.

At the outset, the project recruited 2 older adult patient and public involvement (PPI) representatives to inform the development of the intervention and trial. They contributed to the development of the protocol and early intervention drafts. When intervention development began, it was determined that further PPI input would be essential to make the intervention and study materials engaging and easy to use. Three older adult representatives were recruited. Their input was invaluable in providing direction and feedback on the content and structure of the intervention (eg, the name, logo, content, order of pages) as well as the consent processes, interview topic guides, screening, and study measures (eg, feedback on unsuitable phrasing, resolving questionnaire burden).

During and after the trial, the PPI representatives will inform the discussion and interpretation of the trial outcomes and process evaluations and provide a patient perspective on resolving problems with recruitment, retention, participant support, and follow-up.

The Active Brains intervention was developed to provide interactive tailored support to older adults in initiating and maintaining evidence-based behaviors that support cognitive and physical health using appropriate behavior change techniques. All intervention content has been iteratively developed with extensive input from the target user group to ensure that it is highly accessible and engaging. The Active Brains study website (containing the consent material and the study measures) is accessible to all participants, with only the intervention arm participants having access to the Active Brains intervention content. For the first 7 months, the intervention groups will have access to the Active Brains “Starter Section,” which provides support for users to initiate changes to their lifestyle in line with the intervention’s recommendations.

Within the first section of the Active Brains intervention, 3 modules will become available to users, released sequentially. Within these modules, users will have access to information addressing common concerns, instructions about recommended activities, guided goal setting and reviewing, personalized motivational feedback about their progress, and motivational reminder emails. The modules are as follows:

After 6 months, users will be given access to the Active Brain “Booster Section,” providing further advice on habit formation and additional resources.

Textbox 2 summarizes the Active Brains intervention using the Template for Intervention Description and Replication (TIDieR) checklist [71]. For more details on the development of the Active Brains intervention, see the study by Essery et al (unpublished data, 2020).

Participants will be encouraged to set and review goals every week when using Active Brains. When they unlock “Brain Training,” they will be encouraged to play the games 3 to 5 times per week for the first 6 months. The intervention will be available to the participants for a year. However, the Active Brains intervention aims to help participants establish healthy habits (eg, regular walking) so that participants are not dependent on the intervention in the long term—we anticipate that participants will not need to use the website every week for 6 months to gain benefits from the intervention. The level of engagement that is effective in improving outcomes will be investigated in the main trial.

Patients in the group receiving support from a central facilitator (in addition to the website) will be offered a brief (10 min) telephone support call 2 weeks after they begin the study. In this support call, they will discuss the cognitive or lifestyle changes that they are planning to try. Patients will be offered 2 more support contacts by phone (up to 10 min) or email to support them in making behavioral changes.

Central facilitators will be people trained in the congratulate, ask, reassure, and encourage (CARE) approach to provide support to patients, which is based on Self-Determination Theory [72] and designed to promote autonomous motivation for making behavioral changes. This approach has been successful in previous trials [73].

The comparison group will receive usual care from their GP practice, in addition to brief web-based advice about becoming more physically active, improving the diet, and staying mentally active.

Table 1 presents a list of the measures and when they will be presented to participants. Participants will receive follow-up reminders by email, post, and telephone and by contacting a nominated contact person (optional). If no contact is successful, the contact person will be asked to help put the research team in touch with the participant or to complete 3 short questionnaires about the participant (EQ-5D-5L proxy version 2 [55], IADL [49], and Informant Questionnaire on Cognitive Decline in the Elderly [70]).

Participants will register, be randomized, and complete baseline measures. Participants in the intervention arm will then be given access to the intervention, which they can use for one year. Participants in the support arms will be invited to have a support call at weeks 2, 6, and 10. A sample of participants will be invited to participate in a qualitative process interview at 6 months. All participants will be invited to complete outcome measures at 12 months.

This is a feasibility trial, so it was not powered to measure patient outcomes. A sample size of 180 for each of the 2 subgroups was chosen for several reasons. As Active Brains is a digital intervention with multiple possible pathways through it, a sufficiently large group would be required to observe different patterns of usage. Fully web-based interventions are easy to recruit and manage larger numbers of participants, making this possible. The proposed fully powered trial would need to recruit over 20,000 patients, so implementation issues and trial processes (eg, managing multiple supporters) needed to be tested at scale. A total of 360 participants would provide robust evidence for the feasibility of the trial.

Outcome measures will be collected by using the software at registration and at one year. It will be held securely in a database accessible only by the data controller and members of the study team. Each participant will be assigned a unique study ID when they register with the website. Anonymous outcome data will be held in a separate write-only database to the personally identifiable data to ensure blinding and guard against modification.

Personally identifiable data will be held in accordance with the General Data Protection Regulation [74]. It will only be accessible to members of the study team throughout the study to monitor usage and provide support. The data will be analyzed by a statistician who has had no access to these data.

There will be 2 qualitative process studies, one with participants and one with central support facilitators. Interviews in both these studies may be carried out at any time between 2 and 12 months after participants begin the study. Both of these studies will allow assessment of the acceptability and feasibility of the intervention and highlight any modifications to the intervention or study procedures that might be required before embarking on the later planned, fully powered trial.

Serious adverse events (SAEs) will be reported by both participants, practice staff, and possibly by the central support facilitators who will have contact with participants during the study. It is not anticipated that SAEs will be related to this research. SAEs are defined as any untoward medical occurrence or effect that at any dose results in death, is life threatening, requires hospitalization or prolongation of existing hospitalization (excluding hospitalization for pre-existing conditions or planned procedures), results in persistent or significant disability or incapacity, is a congenital anomaly or birth defect, or results in other important medical events. Any SAE occurring to a research participant will be reported to the ethics committee where, based on the initial judgement made by the chief investigator or an agreed deputy, the event was related to the administration of any of the research procedures and was an unexpected occurrence. Nonserious adverse events will not be collected.

GP practices and the central support facilitation staff will inform the research team and/or the Clinical Trials Unit of any SAEs within 24 hours of them being aware of the event occurring. GP practices and central support facilitation staff will be provided with a standard operating procedure and a form for SAE reporting. Patients may also report SAEs. Reports of SAEs will be provided to the ethics committee within 15 days of the chief investigator becoming aware of the event. In addition to the chief investigator making the initial judgement about the SAE, all SAEs will also be sent to the trial steering committee for adjudication.