Innovative Midlife Intervention for Dementia deterrence

Final Report Summary - IN-MINDD (Innovative Midlife Intervention for Dementia deterrence)

Executive Summary:

1.1 Executive summary

According to the World Alzheimer’s Report 2015, 46.8 million people worldwide are living with dementia in 2015 (with 10.5 million of these in Europe). It is estimated this will rise to 74.7 million cases by 2030 and to 131.5 million by 2050. In addition, the total estimated worldwide cost of dementia in 2015 is US$ 818 billion, rising to US$1 trillion in 2018 and US$2 trillion in 2030 (World Alzheimer’s Report 2015). Any intervention that can reduce the incidence of dementia (or delay its onset) will have obvious individual and societal benefits. The global costs of dementia also make prevention/delay a key focus for future health policy.

In the In-MINDD project we have:
• Combined a comprehensive literature review with the opinions of international dementia experts (in a Delphi study) to identify the most up to date model of modifiable risk factors for dementia which yields a personalised LIBRA (lifestyle for brain health) profile for individuals in midlife.
• Validated this model against large population-based datasets.
• Implemented the model as an online tool for individualised assessment of lifestyle and long term brain health in midlife – users (aged 40-60) receive an assessment and a personalised strategy focused on the adoption of a brain healthy lifestyle.
• Devised an online support environment to help users adhere to positive lifestyle change.
• Tested the In-MINDD system in primary care across 4 countries in a feasibility study.
• Communicated the message that dementia risk can be modified and a healthy lifestyle supports long term brain health to a wide audience, including the general public, health professionals, policymakers and other stakeholders.

The In-MINDD LIBRA score developed in the project is the most comprehensive, evidence-based dementia prediction score that is currently available. It is also novel in that the model targets people in midlife, maximising the potential for positive impact.

The In-MINDD profiler and support environment is another key output of the project. The system was tested in a feasibility randomised controlled trial (RCT), where we demonstrated that a low intensity, on-line intervention is both feasible and acceptable to participants and their primary care providers. Participants in both arms of the trial showed a small improvement in their LIBRA “Keep This Up” score. The improvement was slightly larger in the intervention arm, although this was not statistically significant. Some risk factor areas showed more promise in terms of improvement than others – in particular regular drinking, diet, physical activity and cholesterol. It may be that future refinements of the In-MINDD on-line support environment should focus on these areas.

After the end of the project, the In-MINDD system (In-MINDD 2.0) will be available online in English, French and Dutch to anyone who wishes to establish their LIBRA profile and make use of the support environment (target population 40 to 60 year olds). The system will be available for three years from January 2016, with administrative support from experts in each of the partner countries. In-MINDD 2.0 is designed to be modifiable. There is scope for use of the system as is, or for it to be adapted, for future research purposes.

In the course of our work, we have found relatively low awareness about dementia risk and the potential impact of lifestyle change among the general public and even among some general practitioners. The In-MINDD project has begun the task of raising awareness. We have engaged with several different stakeholders throughout the project, culminating in the in-MINDD Policy Symposium. We believe that the work we have undertaken has sown the seeds for future policy and awareness campaigns.

We believe that the In-MINDD model and the In-MINDD system show real promise for future use to promote long term brain health and help prevent dementia or delay its onset. The project has potential to have a significant impact at an individual level, at a societal level, within primary care and in future health policy.

Project Context and Objectives:

1.2 A summary description of project context and objectives

Background

Dementia is a serious loss of cognitive ability beyond what might be expected from normal ageing. According to the World Alzheimer Report 2015, over 46 million people currently live with dementia worldwide. This figure is set to increase to 131.5 million by 2050. Dementia is one of the most feared conditions related to ageing. It is also a very costly public health issue. The total estimated worldwide yearly cost of dementia is US $818 billion, and it will become a trillion dollar disease by 2018. Whilst dementia is currently incurable, modelling studies have estimated that if obesity rates were to drop by 5%, dementia prevalence rates would be lower by 6% and a decline in physical inactivity rates of 5% would reduce dementia by 11%. It is estimated that even a very small delay in disease onset of 1 year would have a significant impact on dementia incidence globally. Because of the high social and economic costs associated with dementia, this would be hugely beneficial.

Low to middle income countries are seeing huge rises in life expectancy which will increase the prevalence of dementia greatly (WHO 2012). These countries will not have sufficient resources to respond to the care requirements and large scale intervention to address systemic reform (including promotion of disease modifying factors), which will be required (Dias A, Ferri CP, Graham N, Ineichen B, Prince M and Uwakwe R. 2006).

The current European road map for the prevention of dementia indicates that prevention needs to start years before symptoms become apparent and needs to be multi-domain focused. Dementia has complex and interacting aetiologies. Although some of the principal risk factors are non-modifiable (e.g. age, genetics), there is increasing evidence of links between dementia and modifiable risk factors, related to lifestyle.

Prevention can be divided into primary, secondary and tertiary categories and how it is targeted depends on how one defines dementia. As yet there is no biomarker for dementia and many people have the pathology for a considerable time before the symptoms occur; indeed for some people with physiological signs, the clinical symptoms do not occur before death. The boundaries between non-pathological brain ageing and the dementias are unclear and contentious (Frisoni , Bocchetta , Chételat , et al. 2013, Savva, Wharton, Ince, et al. 2009). With a strictly physiological definition of dementia, primary prevention may be defined as the prevention of physiological hallmarks of dementia in the brain (WHO 2014). With a clinical definition of dementia, primary prevention begins when a person is asymptomatic of the disease and interventions are aimed at maintaining that status by removing damaging stimuli and strengthening resilience in a number of ways. The risk of dementia increases with age and most people with age onset dementia (those over 75) have a multi factorial aetiology. There are several rare causes of dementia with a single neuropathology. However, the prevention of these cases is highly intensive and so far research (although promising) has provided as yet, no vaccine against dementia (Lambracht-Washington and Risenberg 2013). In-MINDD targets the large numbers of people with age-onset dementia of mixed aetiology where risk factors will be addressed through multifactorial approaches. Globally these are important risk factors for all chronic diseases.

Since a ‘cure’ for the complex syndrome of dementia is an unlikely development by 2025, risk reduction is modelled to be the most effective approach to delay onset and potentially reduce new cases. Despite this potential, prevention still receives relatively little research funding as compared to drug therapy development and search for biomarkers. Dementia research still emphasises biomarkers, biochemical mechanisms, treatments, and cure. In the UK, the research impact report from the Alzheimer’s Society shows that 5% of research funding between 1990 and 2012 was dedicated to studies of risk factors and preventive strategies.

Raising general awareness of the dementia prevention message is another key challenge we face. It is important to bring the dementia public health message into the mainstream of major chronic disease prevention in order to give it the same legitimacy and standing in the eyes of the public and healthcare professionals alike.

Of course there are many complex barriers to the acceptance of the prevention message, not least, dementia is a condition that weighs heavily on the human psyche, triggering emotional responses of fear and dread. A public discourse of tragedy is perpetuated through storytelling and the media, and evidence suggests understandings of dementia are complex and often contradictory (McParland). Dementia prevention needs to somehow embrace this complexity and deliver a message that is clear, but not misleading, or over-simplified.

Having said that single neuropathology dementia prevention attempts are research intensive, that is not to suggest that there is a straight line between knowledge of modifiable dementia risk factors and changes in health behaviour. Nor is it a simple message to deliver on a population level. It is still possible to get dementia having none of the modifiable risk factors over a lifespan and it is also possible to have many risk factors and not develop dementia. It is however, at the population level that prevention can have an impact on the incidence of dementia at a national, or indeed a global, scale.

Previous studies have taken a range of approaches to addressing the challenge of dementia prevention. The FINGER study is a multi-domain intervention with weekly face to face contact in exercise and dietary education groups and cognitive training. preDiva concentrated on intensive vascular care through nurse led dietary and exercise advice and MAPT targeted specific dietary supplements, exercise and cognitive training. HAtice is a more recently commenced study that has no face to face contact but telephone coaching by motivational coaches who are nurses.

See Figure 1 Summary of Dementia Prevention RCTS by Dr Edo Richard In-MINDD Policy Symposium 1 October 2015.

The In-MINDD Project

The foregoing sets out the backdrop to the In-MINDD study and illustrates the importance of dementia prevention/delay, the challenges that face us and some of the other initiatives which have been developed in this area. In-MINDD was devised and implemented against this background.

The main objectives of the In-MINDD project have been:
• To develop a robust dementia risk model based on modifiable risk factors.
• To design a state of the art online system for use in primary care that provides an individualised profile, based on the modifiable factors in midlife that can increase and/or reduce an individual’s risk of developing dementia in later years.
• To promote individual response through the development of personalised strategies encouraging individuals to make lifestyle changes to protect their brain health and mitigate their potential risk of developing dementia in later life.
• To help adherence to lifestyle change by providing a supportive on-line environment and information.
• To raise general awareness (including among health professionals, policymakers and the general public) that it is possible to influence future dementia risk and long term brain health by adopting a healthy lifestyle in midlife.

The In-MINDD intervention provides comprehensive state-of-the-art assessment of dementia risk, multi-use of programmes designed for risk modification, flexibility for the end user and a tool that can be easily integrated into practice in primary care. The intervention is aimed at people in their fifth and sixth decades.

In-MINDD forms the outer edge of the continuum of scalability (compared to some other interventions outlined above) since it does not require on-going, regular input from healthcare professionals. It was our endeavour to investigate if such a low cost intervention could form a useful part of a strategy to support people to change their lifestyle and health behaviours.

The In-MINDD LIBRA score developed in the project is the most comprehensive, evidence-based dementia prediction score that is currently available. It is also novel in that the model targets people in midlife, maximising the potential for positive impact. In the project we have demonstrated that a low intensity, on-line intervention is both feasible and acceptable to participants and their primary care providers. We believe that the In-MINDD model and the In-MINDD system show real promise for future use to promote long term brain health and help prevent dementia or delay its onset.

Project Results:

1.3 A description of the main S&T results/foregrounds

1.3.1 WP1 Risk Prediction Algorithm

The main goal of WP1 was to build a multi-factorial model for dementia risk, in which a broad spectrum of factors would be incorporated. This model was to be validated against population-based data sets and then used as primary input for the development of the In-MINDD risk profiler and support environment in WP2. The In-MINDD system developed in WP2 would then be tested in a feasibility study (in WP3) to shed light on the efficacy and acceptability of the In-MINDD system and its potential for helping an individual alter their profile score.

Identifying the major midlife risk factors for dementia is crucial to effective preventive strategies such as the approach taken within In-MINDD. In WP1, we developed an algorithm that allows staging someone’s future risk for dementia and quantifying potential for prevention. In the following, we report on the development of this risk algorithm by means of a systematic literature review (extensively covered in Deliverable 1.1) and complementary Delphi expert consensus, as well as results from a validation study on the algorithm’s ability for predicting future risk, using two epidemiological datasets.

1.3.1.1 Risk factor inventory

In the past decades a substantial knowledge base was built in the domain of neuroepidemiology, devoted to defining and quantifying the risk and protective factors that are associated with cognitive decline and dementia. The first aim of WP1 was to make a comprehensive inventory of all such factors, based on the available epidemiological literature and corroborated by expert opinion.

For the literature review, a literature search was conducted in PubMed on all new publications between 29 October 2009 and 5 December 2012. The search strategy was primarily based on an existing high-quality review (Plassman et al. 2010). It was decided to use the studies identified in that review and update the search using a slightly expanded search strategy. Importantly, this review used search terms for a range of pre-selected risk factors. Though their list already was considered to be rather exhaustive, it was decided that a broader generic search term might yield even more results and would also lead to potentially novel risk factors that had not been identified before. The search yielded 3,127 abstracts, of which 320 were included (based on title and abstract), for further scrutiny.

A Delphi consensus procedure was set up to inform the risk factor model based on the literature by obtaining additional insight from scholars in the field of dementia research. A Delphi study is a qualitative exercise in which a panel of experts anonymously answer questions on a clearly predefined topic. A moderator summarizes the responses and feeds this information back to the Delphi experts for further input until consensus is reached. For the In-MINDD Delphi study, a dedicated website that contained the online-survey was built and hosted by Maastricht University (see D1.2). Eight international experts agreed to participate in the In-MINDD Delphi consensus survey (2 from UK, 2 France, 1 Norway, 1 Sweden, 1 Australia, and 1 USA).

The Delphi dementia expert survey and systematic literature search resulted in a large amount of information on potential dementia risk factors. This information was subsequently pooled, ranked and synthesized to identify the major best-evidence modifiable risk factors for dementia that might be used in dedicated intervention programs (such as In-MINDD) for preventive actions. There was good agreement between the Delphi experts and the literature review on which factors to include, which corroborates our confidence in the findings. In addition, by adopting a broad search strategy we succeeded in identifying several novel risk factors which were not included in previous reviews, such as renal dysfunction and heart disease. These novel risk factors were considered to be of value for the In-MINDD risk profiler, although the level of evidence was lower compared to the acknowledged risk factors that had been identified in earlier studies. It was decided that additional efforts were warranted to substantiate the predictive value of these risk factors in separate studies (seer §1.3.1.3 “Further Work on Risk Factors”).

See Table 1 Effects of modifiable risk factors for dementia based on combined evidence from the Delphi expert survey and available systematic reviews.

Having decided on the risk and preventive factors that should be targeted in the In-MINDD intervention, we used effect sizes from meta-analysis for each risk factor to construct an algorithm for future dementia risk/lifestyle for brain health. These effect sizes were used to weigh the contribution of individual risk factors to dementia risk. A complete description of the final risk model can be found in the D1.2 document.

The final list of 12 modifiable risk and protective factors selected consisted of:
• healthy diet; physical exercise; cognitive activity; low-moderate alcohol intake (all protective factors);
• smoking; depression; hypertension; high cholesterol; diabetes; obesity; coronary heart disease, and renal dysfunction (all factors that increase risk).

While all identified risk and protective factors are considered modifiable, some are more amenable to lifestyle changes, while for others (if present), sustained medical attention and control is more appropriate. The figure below presents a breakdown of risk factors in terms of the overall risk of dementia. It shows each risk factor’s share of the overall dementia risk and separates out those factors that need to be managed or controlled and those that are more dynamic. As can be seen, it was finally decided to drop inflammation from the algorithm because measures of inflammation markers (e.g. C-reactive protein levels) are not routinely measured in primary care at all In-MINDD sites and cannot be reliably estimated otherwise.

See Figure 2 Breakdown of In-MINDD risk factors into those factors that are more dynamic (left) and those that need to be managed or controlled (right). The latter are important in terms of chronic disease management, but the former are more amenable to active lifestyle changes initiated by the individual.

In In-MINDD, the final model adopted in WP1 and used for the construction of the risk profiler in WP2 thus consisted of the 12 risk factors above. In the profiler, the weights of different risk factors (quantified by means of questionnaires and available clinical information) produces a given individual’s lifestyle for brain health (LIBRA) score, expressing the gain in terms of brain health (“room for improvement”) that a person may obtain by adopting a (more) healthy lifestyle.

1.3.1.2 Validation

As part of our work in WP1 we took a next step by testing the risk algorithm in relevant longitudinal datasets. In validation studies using existing epidemiological datasets, the In-MINDD risk score was predictive for cognitive outcome later in life e.g. the risk score significantly predicted incident dementia and incident cognitive impairment in the Maastricht Aging Study (MAAS) dataset but did not predict normal age-related cognitive decline. In the DESCRIPA study, an even larger pooled European longitudinal dataset, we showed age-dependent effects of the LIBRA score with greater risk prediction in midlife than in late life. This finding strongly supports the In-MINDD strategy that targeting these risk factors in midlife has most potential.

1.3.1.3 Further Work on Risk Factors

In addition to the above work, In-MINDD researchers at UM have also explored the effects of individual LIBRA risk factors, particularly those on which there is a paucity of published research, more specifically renal dysfunction, coronary heart disease, hypertension, and obesity.

For hypertension, In-MINDD researchers could replicate previous findings of a higher risk for cognitive decline in people with midlife hypertension in data from the Maastricht Ageing Study (Köhler et al., 2014). Subtle decline was already visible in people with a recent diagnosis of hypertension, highlighting the need for early intervention. Importantly, people whose blood pressure was well controlled with medication showed less decline over 12 years than people with poor blood pressure control. These findings support the In-MINDD strategy that adequate risk management (“Keep it Up!”) is also important for a brain-healthy lifestyle.

Depression is another modifiable risk factor, and was even considered most important by our group of international experts in our Delphi study (Deckers et al, 2014). Less is however known about its relationship with other risk factors. In a study of 35,791 primary care patients, we found that the effect of depression on dementia risk is higher in people with existing hypertension or in people with a history of stroke. However, even in the absence of these and other risk factors, e.g. diabetes or myocardial infarction, depression was a substantive risk factor for dementia in later life (Köhler et al., 2015).

Some recent studies did not confirm that obesity is a risk factor for dementia (de Bruijn et al., 2015; Qizilbash et al., 2015). Since the current evidence from available studies still favours the idea that obesity contributes to poor brain health, it remains included in the LIBRA index. We are analysing data from the Maastricht Ageing Study to explore the relationship between obesity and cognitive decline, with a special interest in the impact methodological choices made when doing such studies have on the outcomes.

Other papers we expect to publish shortly relate to the effect of cardiovascular disease on cognitive decline and the evidence for renal dysfunction and for coronary heart disease (including myocardial infarction and angina pectoris) in two meta-analyses. The aim is to provide a more accurate account of their risk for dementia than individual studies can do. Results are expected to be submitted for publication in early 2016.

1.3.1.4 Complementary Work on Machine Learning

The LIBRA score in In-MINDD was developed on the basis of a systematic literature review and expert consensus. This approach guarantees a best evidence model for dementia risk prediction. However, in In-MINDD we also wanted to explore the potential for new knowledge to be generated by the application of machine learning tools to longitudinal datasets on ageing.

In WP1 researchers from DCU applied machine learning algorithms to the MAAS longitudinal dataset to explore interactions of the risk factors identified within WP1; the provision of a secondary validation for those factors and the potential for improving accuracy.

Work undertaken included:
• An extensive exploratory analysis of the underlying data to describe the distributions variables in the dataset.
• Model creation using Deep Learning (DL) and artificial neural networks (ANN) via a manually coded framework in Python to address challenges found within longitudinal data (e.g. missing data, repeated measures), which used an abstract and extensible design for the addition of future functionality and a JSON data-store to export and share models with others.
• The utilisation of cross-validation techniques to mitigate against bias in the training of models developed.
• An incorporation of supervised (models to predict) and unsupervised (models to describe-reduce data dimensionality, cluster, or identify anomalies) algorithms.
• The development of software to allow researchers to match ontology elements specific to a particular analysis with relevant variables.

Initial results are promising and appear to validate the risk factors in the LIBRA model via a neural network survival analysis. We also identified interactions between risk factors, which will need to be clinically validated (on-going work with UM). Further work is required, but we believe that our machine learning research has added further weight to the LIBRA model. We also believe that the advances made on the application of machine learning techniques to longitudinal datasets will ultimately provide researchers with more sophisticated technical ways in which to investigate the relative effect of different risk factors for dementia.

1.3.1.5 Conclusion

In the final analysis, the In-MINDD LIBRA score is the most comprehensive and evidence-based dementia prediction score that is currently available for dementia risk prediction. In contrast to earlier risk scores, it is based on the current scientific evidence and existing meta-analyses for individual risk factors, instead of the risk estimations based on data from a single study.

It is acknowledged that more research is needed to fully understand the complex action of dementia risk factors. More specifically, interactions between individual risk factors need to be addressed in more detail in future studies. It will be important to update the In-MINDD LIBRA model as and when such new evidence becomes available. The In-MINDD researchers will therefore continue to contribute to this knowledge base, even beyond the duration of the In-MINDD project itself.

Our work has already generated traction in the scientific community in that multiple academic centres who are currently engaged in longitudinal (cognitive) aging studies have expressed their interest in collaboration with UM to validate the model on their own independent data sets (e.g. English Longitudinal Study of Ageing in UK, CAIDE study in Finland, and Doetinchem study by RIVM in the Netherlands). Based on the upcoming new insights, the LIBRA algorithm may even be improved further, e.g. by incorporating more fine-grained components to the model such as interactions of individual risk factors with age, gender or APOE genotype, and factor * factor interactions. Taken together these future initiatives have the potential of implementing LIBRA in future prevention trials even beyond what has been established so far by the In-MINDD consortium.

1.3.2 WP2 Profiler and Online Support Environment

WP2 dealt with the design, creation and development of the In-MINDD online profiler and support environment (the In-MINDD system) to implement the dementia risk model and algorithm devised in WP 1.

The objectives of WP2 were:
• To convert the risk model from WP1 into an online knowledge base and rules engine, which accepts a person’s risk factor profile and generates a quantitative dementia risk quotient and a personalised risk reduction strategy.
• To ensure that the system can be used efficiently and effectively by health professionals in a clinical setting by involving all stakeholders in design and validation tasks.
• To deliver supportive online environments in which In-MINDD participants interact with one another and with clinical specialists to encourage adherence, trace progress and drive success.
• To deliver a final and validated profiler and online environment after completion of the feasibility study (In-MINDD 2.0).

The sections that follow summarise the main work undertaken and results achieved in WP2 throughout the project.

1.3.2.1 Specification of Transformation Rules

A robust dementia risk model based on the inventory of modifiable risk and preventive factors was developed in WP1. An algorithm was constructed by the In-MINDD team for calculating a person’s risk of developing dementia based on the presence or absence of the risk factors in the dementia risk model (see Deliverable 1.2). Task 2.2 in WP2 involved the specification of transformation rules to translate the algorithm into knowledge artefacts that are computer-readable and computer-understandable. This meant representing the information in the dementia risk model in a form that the In-MINDD profiler can utilise to generate a dementia risk quotient for individuals in mid-life. It also meant taking into account the feedback from GPs consulted about the likely availability of risk/preventive factors in WP1.

For each of the 12 risk/preventive factors, a separate transformation rule was developed. For each individual risk/preventive factor, a question was developed or readily available measure selected to elicit information from users about that risk/protective factor. For each question/measure, a specified criteria or cut-off value was used (based on extant epidemiological evidence) to determine the presence or absence of that risk/preventive factor. For example, with respect to alcohol consumption, the measure used was the number of standard drinks or units of alcohol that a person consumes in a typical week and the cut-off value is the recommended weekly intake of alcohol in the country in which the person lives. Using the algorithm developed in WP1 and the cut-off values identified, (all of which are dichotomous), a person’s personal dementia risk profile, referred to in the In-MINDD system as a Lifestyle for Brain Health (LIBRA) profile, is generated.

Using the rules base and knowledge system, the online tool translates the dementia risk model into a specific dementia risk quotient and parses the risk space in separate components (modifiable/non-modifiable). Substantial work went into the digital encoding to generate a dementia risk/brain healthy lifestyle profiler, which then required extensive development and testing by the In-MINDD team.

1.3.2.2 Implementation of the Online Risk Analysis System

The main output of this task was the online In-MINDD profiler, which has been described in detail in Deliverable 2.1 The front end, or the Graphical User Interface (GUI), allows users to interact with the profiler by inputting data related to a range of modifiable risk and protective factors via a range of visual widgets. The information gathered is then used to generate a profile of the user’s overall dementia risk/brain healthy lifestyle. Building on WP1, transformation rules were followed so that information in the dementia risk model could be utilised by the In-MINDD profiler to generate a dementia risk quotient (LIBRA score). An online self-administered questionnaire was developed with digital encoding to generate a dementia risk score from the data input by users via a Graphical User Interface.

Consultations with stakeholders at the early stages of the project informed the final vision for the structure of the In-MINDD profiler. The initial vision for In-MINDD was to develop a software system that would be embedded into standard practice management software systems designed for the collection and reporting of patient-level data in primary care practices. In addition to discussions about the likely availability in primary care of variables predictive of dementia risk, meetings were held with representatives from companies providing software to primary care practices in Ireland to explore the possibility of incorporating a dementia profiling tool into existing practice management software systems. GPs in Ireland and Scotland were also asked for their views about the integration of such a system into practice. The feedback from these meetings was that in practice the integration of the In-MINDD system into practice management systems would not be realistic or practical at present. In addition, it was pointed out that since primary care practices are often already overstretched it would be unrealistic to expect primary care practitioners to add an extra profiling tool to use in their consultations with patients. Feedback from primary care colleagues across all the In-MINDD partners was consistent with this. This led the In-MINDD research team to opt for the development of a stand-alone online profiler that would not be linked to existing practice management software systems and would largely be based on patient self-report with several exceptions, such as cholesterol and blood pressure.

The In-MINDD Profiler produced consists of a suite of software developed to run as a Web Application in the cloud. The online self-administered questionnaire is for individuals in midlife and takes approximately 15-25 minutes to complete. The questions included in the In-MINDD profiler were designed to collect the information from respondents in relation to those risk and protective factors for dementia that are included in the inventory of risk/protective factors identified in WP1. Background information about participants and family medical history is also collected. Since it is unlikely that users will have information about their medical health available to hand, clinical data from primary care practitioners is needed in order to create a complete dementia risk/brain healthy lifestyle profile based on the targeted risk and preventive factors. A record sheet, to be completed by primary care practitioners, was designed specifically to collate information (where available) about Height, Weight, blood Cholesterol, Blood Pressure, and whether there is a recorded diagnosis of Cardiovascular Disease, Chronic Kidney Disease, and/or Depression for each user. Questions were developed to capture information about participants’ smoking habits and alcohol consumption. The In-MINDD profiler also collects data on participants’ mood, physical activity, cognitive activity and diet via four existing validated instruments, which have been carefully selected and adapted, where necessary.

The information fed into the In-MINDD profiler is used to generate a dementia risk profile, which is presented to users and primary care practitioners as a “lifestyle for brain health score” (or LIBRA score), in the form of an ‘exploded doughnut’ style chart, as described in detail in Section 4.4 of Deliverable 2.1.

Researchers in each of the partner countries met with general practices to get feedback on the profiler and its development. Overall, the view from general practices regarding the profiler was largely positive. Feedback from general practices related to issues around health literacy and length and ease of completion of some questions in the profiler, were taken on board and changes to address these issues were made. Additional feedback from follow-up meetings led to in-country refinements to ensure that the language and choices offered to participants completing the profiler were appropriate to the country, e.g. in relation to alcohol units or descriptions of food.

A Profiler prototype was initially developed, so that partners could validate, test and give feedback. This prototype was developed using standard software tools in order to produce draft versions of the input screens.

Response validation was an issue considered when developing the profiler. The In-MINDD profiler has been programmed to validate text and numeric values entered into open fields to ensure it meets a specified format (e.g. yyyy), falls within a specified numeric range (e.g. weight, height, cholesterol levels), or to identify specific questions that respondents must answer. It has been programmed to display error messages to prompt and guide users to answer correctly. Skip logic was used in the programming to automatically skip questions that are irrelevant to some respondents. Another consideration was security, which is a key component of each of Google’s cloud computing elements. Actions have been taken to ensure the confidentiality of users, including anonymising and coding the data and the development of clear guidelines to ensure privacy and confidentiality of participants.

To enhance the visual appearance and ‘user friendliness’ of the In-MINDD profiler for users, attention was given to the following:
• Layout, physical placement, spacing and alignment of items in the In-MINDD profiler and ensuring that screens are clutter free.
• Colour, including background colour for blocks of text, has been used to help users with navigation through the In-MINDD profiler.
• Judicious use of images for question comprehension. For example, graphics were used to illustrate what a ‘standard drink’/’unit of alcohol’ is in each of the partner countries to assist users in calculating how many standard drinks or units of alcohol they consume per week and to illustrate portion sizes.
• Wizard/support dashboard for question comprehension.
• Bolding or italicizing was used to direct participants’ attention to important words.

1.3.2.3 Development of the Support Environment

The In-MINDD online support environment was designed so that immediately on completing the profiler, the user can access their personalised LIBRA profile and score in the In-MINDD on-line support environment. In addition, the In-MINDD support environment provides important personalised information on risk factor reduction and strategies to protect brain health. Users receive a personalised plan based on their profile. The approach adopted to do this was based on a ‘health enhancing’ approach, as opposed to ‘reduction of disease’ approach.

In keeping with this approach, the In-MINDD support environment presents individuals with a personalised Lifestyle for Brain Health (LIBRA) profile, using an ‘exploded doughnut’ style chart, comprising three parts:
(i) “Keep this up” - the sum of dementia risk factors which are absent for the individual.
(ii) “Remember to manage well” – the sum of manageable risk factors for which the individual has risk. Three risk factors within the LIBRA model (i.e. CVD, diabetes and chronic kidney disease), whilst not amenable to modification by way of active lifestyle change once a user has developed the condition, are important factors for chronic disease management.
(iii) “Room for improvement” – the sum of the risk factors which are present for the individual and which can be modified through lifestyle change.

In addition to the LIBRA profile and score, the support environment also provides the following information and tools to help the user adhere to their new programme.
• A collection of information resources accessed via a website portal: These resources relate to the 12 modifiable risk/protective factors for dementia. The online resources include information on modifiable risk factors for dementia, which were developed by the In-MINDD research team and endorsed by experts. This information is personalised for each person according to their individual lifestyle for brain health (LIBRA) profile, i.e. according to whether the risk/protective factor is absent or present. Individuals are also directed to existing online relevant resources that are related to their LIBRA profile, utilising specific websites developed by trusted sources available within each partner country, to make sure that the resources are country specific and appropriate e.g. in Scotland, websites created and maintained by the NHS or by recognised charity groups such as the British Heart Foundation. Analogous sites were identified and used for Ireland, France and the Netherlands.
• Goal Setting: Since several lines of evidence indicate that setting goals is an important strategy for helping people to make lifestyle changes, the support environment incorporates goal setting, which again is personalised for each individual according to their LIBRA profile.
• Ask the Expert: The support environment includes an ‘Ask the Experts’ function, whereby experts were available to answer any questions that people using the system may have about dementia, dementia risk and about promoting brain health to help protect against developing dementia in the future. Expert responses to these questions are based on the best available evidence and answers are presented in a way that is supportive and health enhancing. This feature is interactive and provides a human face to the In-MINDD team. The experts respond to individual questions, and also add the questions and answers to the FAQs page. The questions are translated into respective languages and are available to users in the four partner countries.
• Blog: The In-MINDD blog is another feature that allows people using the In-MINDD support environment to share information on topics related to lifestyle changes to protect brain health and reduce risk of developing dementia in later life and to offer tips to other readers of the blog. The In-MINDD blog is a single page of entries restricted to the In-MINDD research team and participants allocated to the In-MINDD arm of the feasibility study. The blog is moderated to ensure it is constructive and beneficial to readers. One person was assigned as the moderator of the blog in each of the four partner countries. A set of rules outlining what is appropriate and inappropriate to post and what should or shouldn’t be in posts was drawn up. A set of quality guidelines and a procedure for complaints was also produced.
• Apps pages: The In-MINDD research team in each country selected mobile apps for tablet computers and phones to be added to the Apps page in the respective countries and languages, which may be useful in helping users to make and adhere to lifestyle changes. Download links were provided to the respective App stores (Android/iOS).
• Email reminders/alerts: During the Feasibility study, users were supported by monthly prompts and email alerts encouraging them to revisit the support environment, set goals and adhere to lifestyle changes to reach goals selected.

All elements of the support environment were localised into all project languages. The use of the support environment, in terms of number of visits, time spent and resources accessed, was tracked during the feasibility study (WP3), using Google Analytics.

1.3.2.4 Technical Functional Validation and Refinement

This task was concerned with involving key stakeholders in design and validation to ensure that the In-MINDD system could be used by health professionals in a clinical setting. Key elements of the task were to specify and test the functionality of the In-MINDD system, in order to ensure that it fits with and supports everyday practice and to understand how both practitioners and patients use the tool. The intention was to apply the resultant information to refine and develop the final version for feasibility testing in WP3.

Deliverable D2.2 (Validation Report) describes the technical validation work on the In-MINDD system carried out in WP2. It also describes the ethical approvals that were required to undertake this work. For this task, a three-phased approach was taken to functionally specify and test the In-MINDD profiler and support environment, namely, context setting, design in process and usability testing.

Phase 1 (context setting) sought to investigate and understand the general context into which the In-MINDD profiler and support environment would be designed to sit, used and implemented and allowed us the opportunity to consult with key stakeholders about the overall vision for the In-MINDD system. The research for phase one was carried out in Ireland, through interviews with key stakeholders. The applicability and acceptability of findings to and for other partner countries was then explored. As outlined in 1.3.2.2 above, the main feedback from the meetings with stakeholders was that the integration of the In-MINDD system into practice management systems, as originally envisaged, would not be realistic or practical at present. This led the In-MINDD research team to opt for the development of a stand-alone online profiler that would largely be based on patient self-report, with exceptions for certain clinical information.

For Phase 2 (design in process), stakeholder and organisational requirements were specified for the In-MINDD profiler and support environment. In each of the partner countries, researchers met with general practices to discuss the dementia risk model and algorithm, and the availability in general practice of certain variables within the model, to ensure that the profiler, when built, would be appropriate for use in primary care. More in-depth work involving the participation of GP practices in the co-design phase of the In-MINDD system (profiler and support environment) and its development was carried out in Scotland and Ireland. In Ireland the researchers also involved primary care patients in this phase. GPs interviewed were on the whole positive towards the In-MINDD project with some amendments being made to the In-MINDD system based on feedback from them. Focus groups revealed that there was a low level of knowledge and understanding about dementia generally among potential end users in Ireland. Awareness of modifiable risk factors for dementia was also low, and whilst focus group participants expressed much uncertainty about the role of various risk factors for dementia, they nevertheless expressed huge interest in becoming informed about this and were pleased to find out that there are steps that can be taken to protect brain health and potentially reduce future risk of developing dementia. The response to the profiler and online support environment from practice patients was overwhelmingly positive, and although the sample was not statistically representative, it is worth noting the positive views of all participants. Service users showed great interest in the support environment, were generally positive about its content and raised several points that usefully informed further development.

In Phase 3 (usability testing) the In-MINDD profiler and support environment were evaluated against requirements. This was mainly carried out in Ireland with primary care patients who would potentially be the end users of the system to check for bugs and robustness and evaluate the usability of the profiler. Findings from usability tests were fed back to the DCU IT team. Some issues concerning layout and alignment of the profiler questions were raised, but overall the participants were positive about the system. They found it easy to use and navigate. Any bugs identified during the usability testing were fixed and additional error checks were put in place to prevent users inputting invalid dates/numbers.

See Figure 3 LIBRA profile from initial In-MINDD system.

1.3.2.5 Development of Final Prototype

The In-MINDD profiler and support environment was developed for use and testing in WP3 (the Feasibility Randomised Controlled Trial). Taking into account:
i. lessons learned in developing In-MINDD, as described in D2.1;
ii. feedback and inputs from end users involved in its design and development (see D2.2);
iii. observations and field notes from researchers’ meetings with participants in the feasibility trial;
iv. the team’s desire to make the In-MINDD system available to the general public after the end of the project and also allow for potential use in future research and
v. a root and branch review of the system by the team with specialist IT subcontractor
a final prototype, referred to as “In-MINDD 2.0” was developed.

Changes were made to both the In-MINDD profiler and support environment. These changes included a redesign of the user interface (UI) of the profiler, making the system fully responsive and making it extensible. Making the system fully responsive means that an optimal viewing and interaction experience is provided to users, that the interface is easy to read, simple to navigate with little requirement on the part of the users to resize, pane, and scroll, and that it can be used across a wide range of devices (from desktop computer monitors to mobile phones). Making it extensible means that the profiler could - if required in the future - be extended e.g. allow for new risk factors to be added, changes to be made to the algorithm, addition of new languages etc. by the In-MINDD team administrators.

Changes to the support environment in In-MINDD 2.0 included the improvement of how LIBRA profile results and screen information resources are presented. There were also improvements to the goal setting tool. There is a new user area (available to registered users only) where users can view their profile, compare their current profile with their previous profile, select and view goals and keep a diary of activities they engaged in towards reaching the selected goal(s).

Both the profiler and the support environment of In-MINDD 1.0 were originally web-based resources accessible on desktop computers and laptops only. The front end of the user interface of In-MINDD 2.0 has been designed to be fully mobile responsive. This provides for an optimal screen environment for users across different devices.

In-MINDD 2.0 includes a bespoke Content Management System (CMS). Through the CMS, administrators of In-MINDD 2.0 (drawn from the In-MINDD team) have the ability to update any of the items/content of the In-MINDD profiler and support environment. The CMS is very user friendly and offers complete flexibility via: Page Manager, Image/Media Manager, Multi Language Module, Content Modules and Application Data Management.

The final prototype provides general users wishing to use the In-MINDD system with two main options, to use the system as either (1) a Guest User or (2) a Registered User. The storage of users’ data is linked to the users’ status, which is described in detail in D2.3. For both Guest Users and General Registered Users, In-MINDD 2.0 has security features to ensure that confidentiality is assured and that the system complies fully with data protection legislation.

As outlined in section 1.4.2 below, a plan for maintaining the In-MINDD system over a three year period commencing January 2016 has been devised, in addition to a framework for deciding about future use and development of the system.

See Figure 4 Screenshot of In-MINDD 2.0 Results page and example of personalised Lifestyle for Brain Health Score and Profile.

1.3.2.6 Complementary Work on Machine Learning

In addition to the above tasks, work was also done in WP2 to support the advances in machine learning techniques described in 1.3.1. In particular, the In-MINDD ontology was developed in WP2 for use in the application of machine learning and data mining to longitudinal datasets. Further details of our work on machine learning techniques are set out in 1.3.1 above.

1.3.3 WP3 Feasibility Study of the IN MINDD Profiler and Environment in Practice

The aim of WP3 was to test the feasibility, acceptability and impact of the In-MINDD profiler and on-line support environment (the In-MINDD “system” or “intervention”) by conducting a feasibility randomised controlled trial (RCT). The specific objectives for WP3 were:
1. To compare the effectiveness of the In-MINDD intervention compared with care as usual in reducing the overall risk of developing dementia and on individual risk factors.
2. To explore how patients use the In-MINDD intervention in terms of access to and time spent on the on-line environment.
3. To explore the feasibility of using the In-MINDD risk profiler and on-line support environment from the perspective of patients.
4. To explore patients understanding of, attitudes towards and experiences of obtaining a Lifestyle for Brain Health (LIBRA) profile.
5. To explore the supports and barriers to embedding behaviour changes into everyday life, and whether or not the In-MINDD intervention helps or hinders patients make and maintain changes to individual health-related behaviours.
6. To explore the feasibility of using the In-MINDD risk profiler from the perspective of primary care practitioners.
7. To understand practitioner views of the utility of the Lifestyle for Brain Health profile and how they relate that information to patients.

The In-MINDD RCT was not a definitive outcomes-based trial; rather, we sought to capture and understand the acceptability and usefulness of the In-MINDD intervention for both research participants (people in primary care settings) and practitioners. It was, therefore, analogous to the modelling stage described in the Medical Research Council’s Framework for Complex Interventions. Although the study was powered to detect a difference in overall dementia risk (difference in overall risk comparing risk score at end to risk score at start of RCT) between those randomised to the In-MINDD intervention and the control group (if such a difference existed), we were also particularly interested in exploring and understanding how patients and practitioners comprehend dementia risk, how this impacted on their understanding and willingness to use the In-MINDD intervention, and whether they were able and willing to make and maintain the lifestyle changes suggested by the In-MINDD intervention. This qualitative process evaluation was a crucial aspect of the In-MINDD trial and was underpinned by Normalisation Process Theory (NPT). This use of theory to understand why complex interventions are embedded, or not, into daily life and work is a crucial aspect of implementation research.

WP3 followed on sequentially from WP1 and WP2 and was crucially dependent on those work packages. Key tasks for WP3 were:
• Data collection forms and ethics.
• GP practice recruitment.
• Patient recruitment.
• Conduct of the feasibility trial.
• Analysis and reporting.

1.3.3.1 Study protocol

Underpinning all of these tasks was the development of a study protocol. Led by GU, this involved all of the research partners, who each designated a key individual to lead on protocol development; protocol developments and iterations were discussed at face-to-face Consortium meetings in Nice in November 2013 and in Glasgow in March 2014, as well as through videoconferences and email. The protocol was developed according to the SPIRIT guidelines (Chan A-W, Tetzlaff JM, Altman DG, et al. SPIRIT 2013: new guidance for content of clinical trial protocols. The Lancet 2013;381(9861):91-92 and Chan A-W, Tetzlaff JM, Gotzsche PC, et al. SPIRIT 2013 explanation and elaboration: guidance for protocols of clinical trials. BMJ 2013;34). A full copy is available on request. The study protocol version 5.4 was agreed on 11 June 2014.

The In-MINDD RCT was registered with the ISRCTN Registry: ISRCTN 98553005 (DOI: 10.1186/ISRCTN98553005). A protocol paper was published by the journal Pilot and Feasibility Studies (O'Donnell, 2015).

See Figure 5 In-MINDD Trial Protocol page on ISRCTN Site.
1.3.3.2 Data collection tools and ethical approvals.

The main data collection tool for the RCT was the In-MINDD profiler which was developed as part of WP2, and led by DCU. There were, however, other data collection tools which were required to ensure the smooth conduct of the trial and to collect qualitative data. These included an on-line Google document to record and monitor participant recruitment; a form to collect clinical data from general practitioners; and a set of interview schedules to guide the qualitative work.

The on-line Google document was developed by GU, in collaboration with partners. This document recorded participant details, the arm of the trial they were randomised to and monitoring data collected during the trial, such as when the trial was completed and any adverse events.

The clinical form was developed collaboratively by all partners. This was used to collect clinical data from practices, including verification of height and weight; blood pressure; blood cholesterol level; and whether the participant had ever been diagnosed with cardiovascular disease, diabetes and/or chronic kidney disease.

An essential part of the feasibility RCT was the qualitative process evaluation. GU led on the development of interview schedules and focus group guides to use with participants and with health care professionals. These were informed by NPT and covered a range of topics including: participants’ understanding of the role of modifiable risk factors in dementia; their views of the information they received after completing the profiler; their use of the on-line support environment; and how they did/did not incorporate recommendations into their daily life. For health care professionals, we explored their views of dementia; the use of risk scores; preventive strategies for dementia; and the In-MINDD profiler.

All of this work was completed for the feasibility trial going live in October 2014.

Ethical approvals were applied for in each country and final approvals were obtained in July 2014 for DCU; July 2014 for GU; September 2014 for UM; and August 2014 for UNS. Additional approvals were required by UM and UNS when they recruited additional general practices. In GU, NHS research management approval was obtained after ethical approval. These activities were detailed in D5.1. One issue was the variation in the number of approvals required across partner countries, as well as additional approvals, for example NHS Management approval in Scotland. These are detailed in Table 2.

See Table 2 Ethics committees.

1.3.3.3 Practice Recruitment

The aim of this task was to identify and recruit up to 6 practices in each participating country; two practices had already been recruited in both Ireland and Scotland during WP2 and these practices also participated in WP3. Recruitment of the other practices commenced early in 2014 and was a major focus of activity from May to August 2014 in all four partner countries.

A pragmatic strategy was adopted to account for the different primary care systems in each participating country. The initial aim of recruiting 6 practices per country was met in all locations except the Netherlands, who recruited 5 practices; these practices were, however, larger than the Dutch average, and so had a substantial patient population to recruit from. This was partly due to the requirement to apply for new ethical approval for each individual practice, which proved to be very time consuming (as reported in D3.1) but also due to a lack of interest in participation amongst practices (5 practices recruited from a total of 40 approached). France and Ireland recruited more practices, as they felt that the individual GP participating in each practice would find it easier to recruit a smaller number of participants.

In Scotland, the team drew first on a registry of research interested practices held by the Scottish Primary Care Research Network and, second, on direct links with local practices. This was successful, as most practices approached agreed to participate. Ireland also drew on existing links initially, but then had to identify and recruit practices from a wide geographical area. One practice later withdrew due to lack of time.

Practices covered a range of locations (urban, semi-urban, suburban/rural) and socioeconomic characteristics. Several of the practices recruited in both Ireland and in Scotland covered areas of socioeconomic deprivation.

1.3.3.4 Patient Recruitment

Patient recruitment into the trial was an integral part of WP3 and so overlaps entirely with Task 3.4 the feasibility study. Patient recruitment was initially planned to begin in July 2014. However, due mostly to delays in obtaining ethical approval (and, in Scotland NHS management approval) the start of recruitment was delayed to October 2014 in all countries.

In-MINDD planned to recruit 600 participants in total; 150 per country. However, with no exact precedents of prior work to inform this decision, we had chosen this number based on previous experience of other primary care-based studies with recruited patients. In the end, In-MINDD recruited 451 participants over a 38 week period, 75.2% of our projected target. Country breakdown and rates of recruitment are shown in Table 3.

See Table 3 Recruitment by country.

Recruitment was slow at the beginning of the trial. As a result, we had to develop pragmatic strategies to broaden and increase recruitment. These strategies included: recruiting more general practices; relaxing the application of inclusion criteria when first identifying potential participants (inclusion criteria were verified by the researcher at the baseline meeting instead); widening the application of search criteria to include any registered patients in the eligible age range who had any one of the risk factors of interest; recruiting participants from a workplace setting, rather than through general practices. In the latter case, participants were then seen by a GP in the workplace medical centre in order to collect the clinical data required for the profiler. Finally, the recruitment period was extended from 4 months to 7 months.

Overall, this broadening of the recruitment strategies was positive helping to maintain, and in some settings increase, the rate of recruitment.

1.3.3.5 Feasibility RCT and Analysis and Reporting

A total of 451 participants were recruited into the feasibility trial between October 2014 and June 2015: The breakdown by country is shown above.

Baseline characteristics of the 451 participants were described in Deliverable 3.2. Overall, the mean age of participants was 52.7 years and there were more females than males. Over two-thirds were married, in a partnership or co-habiting relationship and the majority were living with relatives, which included spouse/partner and/or children. Participants were also well educated, with 253 (56.4%) educated to college or university level. Over 70% were in paid employment, about half were in managerial or professional jobs. Baseline characteristics were similar across the two arms of the study (Table 4).

See Table 4 Baseline characteristics by In-MINDD treatment group (n, %).

There was some variation across countries. For example, in Scotland there were more males than females participating in the study; in Ireland, there were more females than in any other country. There was also some variation in educational attainment across countries. However, overall, there were no major differences in the participant populations across the countries.

Baseline medical history and health-related risk factors

Overall, the mean BMI was 27.9 (SD 5.4) with 31.3% of the study population obese and 37.3% with high cholesterol. There was a high prevalence of the population who reported being told they had high blood pressure at some point in their life. This was supported by almost 50.0% having a diagnosis of hypertension in their medical notes and 13.0% having a diagnosis of CVD. In addition, about 10.0% had a diagnosis of diabetes. Approximately one-third of the study population were obese, i.e. had a BMI of over 30. Data were also collected on health-related risk factors. Almost 20.0% were current smokers; just over 40.0% drank several times a week.

Baseline medical history and risk factors were similar across both arms of the trial (Table 5).

See Table 5 Baseline medical history and health-related risk factors.

Impact of the In-MINDD intervention on LIBRA “Keep This Up” score

The primary outcome was change from baseline in the LIBRA “Keep This Up” score. On 1 December 2015, 313 (69.4%) of participants had completed the trial and re-completed the profiler at 6 month follow-up: 159 in the intervention group, 154 in the control group. Table 6 shows the change in LIBRA “Keep This Up” score.

See Table 6 LIBRA “Keep This Up” score at baseline, follow-up and change between baseline and follow-up, by In-MINDD treatment group.

The mean LIBRA “Keep This Up” score improved slightly more for participants completing the In-MINDD intervention arm (mean change = 1.8) than for those in the control arm (mean change = 1.1) although the mean difference of 0.638 was not statistically significant (p=0.539 95% CI -1.389 – 2.664).

Impact of the In-MINDD intervention on individual risk factors

Table 7 shows the impact of the intervention on individual risk factors. Overall, the majority of participants showed no change in the level of risk factors across the trial period. There were, however, some points of note.

See Table 7 Changes in risk factors across the In-MINDD intervention arms (n, %).

Alcohol consumption improved more amongst participants in the intervention arm, with 8 showing improvement in the intervention arm versus 2 in the control arm (p=0.090). This was due to an improvement in regular drinking patterns in this intervention arm. There was also more improvement in diet and physical activity in the intervention arm, although this was not reflected in a change in obesity score. However, it is likely that the intervention period of 6 months was too short to demonstrate notable changes in obesity category. There was also greater improvement in both cholesterol overall and in high cholesterol. The reasons for this are unclear; it may be that participants who received a LIBRA profile which flagged cholesterol as a risk factor attended their GP for advice and treatment. However, GPs did not report a marked increase in workload as a result of In-MINDD. Alternatively, people may have commenced self-management approaches, such as taking cholesterol-lowering drinks, which may have had an impact.

Some areas showed no shift in favour of improvement, for example smoking. Given that only one-fifth of participants were current smokers, if may be that these individuals were reluctant to make any changes – this is borne out in the qualitative work reported later. Depression and cognitive activity also showed little improvement. This is unsurprising for depression, which was based on a past diagnosis of depression and so was not particularly amenable to change in the intervention. Cognitive activity may take longer to shift than could be monitored in a 6-month period.

Taken together, these results suggest that more sustained targeting of some areas, such as alcohol consumption, diet, physical activity and cholesterol level may yield more benefit in the longer term.

Use of the on-line support environment

Google Analytics was used to monitor use of the on-line support environment in terms of the pages accessed. This was monitored by country. There are some caveats to interpreting these data. As Google Analytics information is anonymised, it was not possible to link use to user ID. This meant that, as control participants completed the trial period and were granted access to the on-line site, their usage would also be monitored. We therefore chose to limit the period of analysis from 1 October 2014 until 30 June 2015. Overall, there were 748 views of the pages on risk factors in the 9 months from October 2014 to June 2015 (Figure 6). The most frequently viewed pages were diet (163 views, 21.8%), blood pressure (103 views, 13.8%) and physical exercise (86 views, 11.5%). The least frequently viewed pages were smoking (26 views, 3.5%), heart disease (31 views, 4.1%) and chronic kidney disease (13 views, 1.7%). Pages on obesity, mood and cognitive activity were also accessed less (at between 5.0 to 7.0%).

See Figure 6 Use of the on-line environment by risk factor.

As well as accessing information about the risk factors, it was also possible to set goals relating to the modifiable risk factors identified by the profiler as requiring active management. Goal options were viewed a total of 242 times. Although we cannot tell if goals were set and adhered to, it is clear that the goals viewed – in terms of frequency – mirrored the view of the risk factor information pages, suggesting that some participants were moving from information on risk to goal setting. The most popular goal options viewed related to diet (82 goal option views, 33.9%), physical exercise (38 goal options viewed, 15.7%) and blood pressure (35 views, 14.5%). Those goal options viewed least focussed on smoking (6 views, 2.5%) and cognitive activity (8, 3.3%).

Qualitative findings

The findings from this work have been reported in Deliverable 3.2 and are summarised here.

i. Awareness of modifiable risk factors for dementia:

Overall, we found that participants were largely unaware of the link between mid-life modifiable risk factors and the risk of developing dementia in later life. The exceptions were those who had personal experience of dementia, often in relation to a parent, but also other family members, friends or colleagues. Indeed, a significant number of those interviewed had experience of a family member with dementia.

When it came to thinking about potential risk factors for dementia, genetics and hereditary factors were often mentioned as leading to dementia, as was stress and mental health. This was true in all countries, for example:

‘Well, I always thought that there are certain genes played a role. Bad luck, I think so [too]. I also think it’s a matter of luck or bad luck. And certainly living conditions will play a role in it’ (ID3304028, The Netherlands, Female, Intervention arm, Baseline interview).

There were mixed views about modifiable health-behavioural risk factors, such as smoking and physical activity. Some interviewees had at least heard that there might be a role for these factors in increasing one’s risk of developing dementia in later life.

‘Interviewer: Did you have any idea about the kind of risk factors for dementia?
Interviewee: Yes, I suppose I did, you know, in terms of lack of exercise, obesity and links with education attainment, cognitive stimulation, so those kinds of things…’ (ID2201013, Scotland, Female, Intervention arm, 6 month interview).

With so many participants unaware of the role of modifiable risk factors in dementia, it raised the question of why they had decided to participate in In-MINDD. Knowing someone, especially a family member or friend, who had or currently had dementia, was a key factor. This was true regardless of gender of the participant. Others were identified by their GP as someone who might benefit from being involved. This was particularly true of French participants, who were identified and approached directly by their GP.

Interviewee: I did it because I trust my doctor….. But then I realised that it was also interesting for me…’ (ID441029, France, Female, Intervention arm, Baseline interview)‘.

For others, it was a general interest in seeing what their risk factors were or a general view that it was good to try and minimise their risk of developing dementia in the future.

ii. What did they change and was that easy?

Many participants interviewed had received “room for improvement” scores in relation to smoking, diet, and/or physical activity. Areas targeted included diet and cholesterol, as previously described. Changes to diet seemed to be particularly popular; this was apparent in all countries. It was also clear, however, that changes to diet had to involve spouses/partners and family members.

‘I also wanted to eat better …. With my husband, we talked about it and we will try to cook more vegetables. We are already large consumers of fruit, but it’s true that we could eat more vegetables. And it’s good for the kids too.’ (ID4408047, France, Female, Intervention arm, Baseline interview).

Some also discussed wanting to increase their physical activity, although maintaining this appeared to be more difficult than maintaining dietary changes. The key appeared to be small steps and finding way of fitting exercise into daily routines. The goal setting option was seen as helpful and a number of interviewees talked about using it, particularly in relation to exercise and diet targets. However, others had not engaged with this feature and it was not clear that they would. The context in which people live and work was also important to their ability to make, and maintain, lifestyle changes. Interviewees discussed issues such as workplace cafes and food availability, access to and time to be able to exercise, and the availability of social clubs. Availability of food at work could undermine plans to create healthier food to take to work, as this was seen to be time consuming.

Social activities were being addressed by some. One spoke of the need to socialise more as they felt they tended to get isolated due to working at home. Others had joined local clubs.

Changing behaviour was however acknowledged as requiring many approaches, of which In-MIND was just one aspect.

‘… it’s just about a broad approach [to making change] where you get it from several different angles and its just engagement. …. I ended up at the end of the year, I suppose, much better informed than I would be at the beginning of the year and it wasn’t particularly this [In-MINDD] but a culmination of things.’ (ID1103001, Ireland, Male, Intervention arm, 6 month interview).

iii. Reflecting on participating in In-MINDD

Participating in In-MINDD had clearly raised participants’ awareness of the association between modifiable health behavioural risk factors and the risk of developing dementia in later life, moving them beyond just thinking about age and genetics. Those allocated to the intervention arm reported an increase in their awareness of the links between general health and the risk of developing dementia in later life.

‘I think lifestyle now would definitely be a factor and genetic because when I looked on the website and I see how you can help yourself, you know to try and be active and be aware, you know ehm …. I’ve signed up now for a computer class which starts next Tuesday in the library …’ (ID1105004, Ireland, Female, Intervention arm, 6 month interview).

Some in the control arm commented on how completing the profiler made them reflect on their health behaviours.

‘… well when I was doing the filling in [of] the questionnaire thing, you know some of the questions … and then I was realising “Oh my God! Do I really smoke that much?” or “Do I really drink that much?” and it kind of it makes you think where normally you just plod along …’ (ID 1101078, Ireland, Female, Control arm).

The numerical value received in the LIBRA profile was generally well received and most seemed to understand the score and what contributed to that. A low score was often reassuring. For some, taking part in In-MINDD reinforced the view they already held about general health, particularly in relation to exercise, diet and keeping mentally active. For others completing the profiler make more explicit to them the risk that certain lifestyle behaviours pose:

‘… I hope there is something that you can actually do even at my age of 54 that I can maybe avoid severe dementia maybe if I make lifestyle changes just now’ (ID2204009, Scotland, Male, Intervention arm, 6 month interview).

However, all of this had to take place within a supportive context. For example, one Dutch participant spoke about the difficulties of reducing their smoking when they worked in an environment where lots of people around them smoked. This participant also spoke at both baseline and 6 month interviews about the increased stress in her life due to living beside a difficult neighbour. These wider issues can, therefore, impact on someone’s ability to make lifestyle changes.

1.3.3.6 Summary

We have demonstrated that a low intensity, on-line intervention is both feasible and acceptable to both participants and their primary care providers. Overall, the objectives for WP3 were met. As well as the substantive results, we have learned much about what does – and does not – work in relation to recruiting participants from general practice to a study focussed on prevention of dementia which requires a degree of digital literacy in order to complete the profiler and use the on-line support environment. The participants were, generally, well-educated and in employment; in future, adaptations are required to target those with lower levels of education and health literacy, to ensure that we do not enhance the “digital divide”.

Participants in both arms of the trial showed a small improvement in their LIBRA “Keep This Up” score. The improvement was slightly larger in the intervention arm, although this was not statistically significant. Some risk factor areas showed more promise in terms of improvement than others – in particular regular drinking, diet, physical activity and cholesterol. It may be that future refinements of the In-MINDD on-line support environment should focus on these areas. These areas were certainly the pages that were visited most often by participants.

In addition, we highlighted relatively low levels of awareness of the link between modifiable risk factors in mid-life and the risk of developing dementia in later life. This highlights a clear need to engage more with patients and the public – and with health care professionals working in primary care.

Potential Impact:

1.4 Impact, dissemination and exploitation

1.4.1 The In-MINDD Model and System

In the In-MINDD project we combined a systematic literature review with an expert consensus survey to generate a dementia risk model, referred to as a lifestyle for brain health (LIBRA) model. This model was then used as the basis for the In-MINDD system – an online tool that:
i. assesses individual lifestyle for brain health in midlife;
ii. devises personalised strategies for lifestyle change to promote brain health and potentially reduce a person’s risk of developing dementia and
iii. provides on-line support to help users adhere to their programme for change.

We tested the In-MINDD system in a feasibility study and gained new knowledge about the system, the LIBRA model, behaviour change, public awareness and attitudes to dementia and to brain health. We demonstrated that a low intensity, on-line intervention is both feasible and acceptable to participants and their primary care providers. Participants in both arms of the trial showed a small improvement in their LIBRA “Keep This Up” score. The improvement was slightly larger in the intervention arm, although this was not statistically significant. Some risk factor areas showed more promise in terms of improvement than others – in particular regular drinking, diet, physical activity and cholesterol. It may be that future refinements of the In-MINDD on-line support environment should focus on these areas.

In the final stages of the project we further refined the In-MINDD system and developed the final In-MINDD prototype (In-MINDD 2.0) with an improved user interface that is intuitive, responsive and extensible. In-MINDD 2.0 also has a tailor-made content management system and is available across different platforms (computer, tablet and phone).

We believe that the In-MINDD model and system show real promise for future use to promote long term brain health and help prevent dementia or delay its onset.

1.4.2 The Future for In-MINDD

A plan for maintaining the In-MINDD system over a three year period commencing January 2016 has been developed. After the end of the project In-MINDD 2.0 will be available online in English, French and Dutch to anyone who wishes to establish their LIBRA profile and adopt positive lifestyle changes to improve their long term brain health (target audience 40 to 60 years). Users will be able to access to the support environment, set goals, compare their scores over time etc. DCU have agreed to take responsibility for overseeing administration and maintenance for a period of three years commencing January 2016 with a nominated administrator in each partner country taking responsibility at a local level for administration and maintenance of the system. The inclusion of a bespoke CMS for In-MINDD 2.0 means that the In-MINDD team administrators can make any changes to the questions required and changes to the In-MINDD profiler and content of the support environment, without having to rely on a software developer, which is an effective and cost-efficient way of managing the system.

Apart from making the In-MINDD 2.0 available to the general public after the end of the project, the In-MINDD team are considering the potential for the research partners to collaborate on the further development of the In-MINDD system. In-MINDD 2.0 is designed to be extensible. There is scope for use of the system as is, or for it to be adapted, for future research purposes. This possibility will be explored over time by the partners. Possible opportunities include developing the interactive elements of the system, increasing personalisation (‘tailoring’), making more accessible to those with lower health literacy, tailoring for people with risk factors for dementia and other chronic illness, targeting more specifically the risk factors we have discovered are most amenable to change, exploring use for populations who are already experiencing memory problems and the possibility of coupling the online system with motivational interviewing or persuasive technologies. Opportunities to use the learning from the feasibility trial to develop a full-scale effectiveness trial will also be explored.

In addition to considering the consortium’s own future use of In-MINDD, the project team has discussed the question of providing access to the In-MINDD system itself and/or its underlying code to any other researchers who may wish to use or adapt the system in the future e.g. for further research. The team is open to the possibility of providing access to the system where it is reasonable and practicable to do so. However, this will be subject to any interested researchers gaining appropriate ethics approvals for the proposed research and to additional ethical, legal or other considerations (including any IP or other issues relating to the four existing validated instruments used as part of the profiler) as the In-MINDD Legacy Committee may decide.

The In-MINDD Legacy Committee will be responsible for overseeing the use of the In-MINDD system after the end of the project. The Legacy Committee includes the principal investigators from each of the academic partners. Their job will be to consider and decide on potential future uses of the system, including the assessment of applications to use the system from external researchers. The Legacy Committee will also consider whether (at a future date) the underlying code for the system should be made open access. The Legacy Committee shall review the position regarding In-MINDD and its legacy 3 years after the end of the project. The consortium believes that the establishment of the Legacy Committee is the best way of facilitating access to the In-MINDD system, while also safeguarding the rights attaching to external content included in the system, ethical imperatives and the legacy of the In-MINDD project for the future.

See Figure 7 Screenshot of In-MINDD 2.0 on tablet with 7” screen.

1.4.3 Advances in Machine Learning Techniques

The LIBRA score in In-MINDD was developed on the basis of a systematic literature review and expert consensus. This approach guarantees a best evidence model for dementia risk prediction. However, in In-MINDD we also wanted to explore the potential for new knowledge to be generated by the application of machine learning tools to longitudinal datasets on ageing.

During the project researchers in DCU applied machine learning algorithms to the MAAS longitudinal dataset. Initial results are promising and appear to validate the risk factors in the LIBRA model via a neural network survival analysis. We also identified interactions between risk factors, which will need to be clinically validated (on-going work with UM). Further work is required, but we believe that our machine learning research has added further weight to the LIBRA model. We also believe that the advances made on the application of machine learning techniques to longitudinal datasets will ultimately provide researchers with more sophisticated technical ways in which to investigate the relative effect of different risk factors for dementia.

1.4.4 Impact Overview

Dementia is one of the most feared of age-related conditions, yet we have found relatively low awareness about dementia risk and the potential impact of lifestyle change among the general public and even among some general practitioners. The In-MINDD project has begun the task of raising awareness. We have engaged with several different stakeholders throughout the project, culminating in the in-MINDD Policy Symposium. We believe that the work we have undertaken has sown the seeds for future policy and awareness campaigns.

In-MINDD has potential for impact at a number of different levels. In-MINDD can impact on an individual level, helping individuals in midlife to improve their lifestyle to support long-term brain health; empowering people to prevent or delay the onset of dementia, as well as other related conditions and improving their health and quality of life. Change at an individual level can be achieved by individuals accessing the In-MINDD system online and also by individual’s becoming more aware about dementia risk and lifestyle change.

In-MINDD can impact primary care practitioners. Over time we believe In-MINDD will facilitate increased awareness and understanding around primary prevention of dementia. It will provide a tool for use in primary care with patients with at least one modifiable risk factor for dementia, or patients worried about their memory and future dementia risk.

In-MINDD also has the potential for significant impact at a societal level. Given the very significant costs of health and social care for people living with dementia, a reduction in the prevalence of dementia would lead to very large savings in health and social care costs over the long-term. While the overall impact of the project may not be apparent for many years, there is potential for reductions in individual risk factors (e.g. improved mood, adherence to healthy diet, greater physical activity, enhanced cognitive activity, moderate alcohol consumption) to be seen much more quickly. The risk factors for dementia are also relevant for other major sources of ill-health including cardio-vascular disease, diabetes, and depression. The healthy lifestyle promoted by In-MINDD has the potential to reduce the burden on the health care system from conditions associated with related risk factors.

Finally, the project, whilst meeting the policy objectives expressed in European and national healthy ageing policy, also has significant potential to inform future policy, by providing new knowledge, information and tools to meet the challenges of dementia in ageing societies.

1.4.5 Dissemination

Effective dissemination of the In-MINDD message has been essential to achievement of the project’s core objectives, which include:
- raising awareness of the link between lifestyle and long term brain health;
- impacting healthcare policy and systems;
- improving the health and well-being of an ageing population and
- reducing associated health care costs.

The team undertook a comprehensive dissemination programme over the course of the project in WP4, aimed at academics, medical and allied health professionals, policymakers, advocacy organisations, the media and the general public. Our aim has been to educate, and raise awareness about, modifiable risk factors for dementia, the In-MINDD risk model, the In-MINDD system and the importance of living a healthy lifestyle in midlife. Below we describe the dissemination activities undertaken including online, peer reviewed publications, conference presentations, project materials, networking, media and the In-MINDD Policy Symposium.

1.4.5.1 Website and Online

The In-MINDD website was established at http://www.inmindd.eu/ at the start of the project, following extensive consultation and collaboration between the partners. The website was regularly updated during the course of the project e.g. with news items of interest, projects newsletters, flyers etc. and details of project publications.

The website includes the following pages:
• Homepage (providing a brief summary of the project and the project partners and links to more information, news items etc.).
• Project page (setting out more detail about the project’s aims and objectives).
• News (charting recent developments and events).
• Media (including key information about the project, project media contacts, general information about dementia, press releases, project promotional materials and media downloads).
• Partners (detailing each beneficiary, with links to beneficiary websites).
• Related Projects (hosting links to related projects of interest).
• Publications (providing details of project publications and other articles by partners that are relevant to In-MINDD).
• Contact Form (offering a mechanism for interested parties to contact the project team).

See Figure 8 Screenshot of In-MINDD homepage http://www.inmindd.eu/.

Traffic to the website grew significantly over the course of the project. The table below shows a summary of visits each year to the website since it was established in 2012.

See Table 8 Annual Summary of Visits to http://www.inmindd.eu/ since the establishment of the project, taken from http://inmindd.eu/awstats/.

As part of our on-line strategy, In-MINDD also established a twitter account (https://twitter.com/fp7inmindd) and a Facebook page (https://www.facebook.com/In-MINDD-500669519994517/info/?tab=page_info) and built our social media profile over the course of the project, with regular tweets and posts.

1.4.5.2 Media

Delivering the In-MINDD message to the general public through the media has been one of the key aims of WP4. Several project press releases were issued by the partners and In-MINDD also featured in press releases issued by third parties. Our website includes a dedicated media page.

Our work with the media has generated good press coverage, including the following:
• In Ireland the project was featured in the Irish Times (22 Oct 2012).
• In-MINDD was the subject of a radio report and interview (Dr Kate Irving DCU) on the Irish national broadcaster (RTE) on 19 Sept 2012.
• Partner UNS gave an interview on local radio station "France Bleu Azur" on July 24, 2013, when a researcher from UNS explained the principles of the In-MINDD study, the partners involved and details of the feasibility study taking place in France.
• In March 2014 In-MINDD partner UM published an article referencing In-MINDD in Geron (national journal devoted to ageing & society, special issue on brain plasticity and brain training).
• The project was featured in the Dutch regional paper Dagblad De Limburger (widely read in the Limburg province). Prof Verhey and Dr Köhler from In-MINDD partner UM were interviewed for the piece. The article entitled “How do you delay dementia?” (Dutch only) was published on 1 October 2014.
• The In-MINDD Project was featured in an Irish Times article, entitled “EU teams unite in supporting healthy ageing“. The article, published on 6th May 2014, was written by Maire Geoghegan-Quinn, (former European Commissioner for Research, Innovation and Science) and discussed the In-MINDD project as well as other projects which support healthy ageing.

1.4.5.3 Researchers, Healthcare Professionals and Stakeholders

Engagement with other researchers, healthcare professionals and stakeholders has been a key focus for In-MINDD. One of the core elements of our dissemination strategy has been to reach these groups through peer review publications and conference presentations.

Publications

The In-MINDD team has achieved a number of significant publications in peer review journals. To date we have achieved 12 publications (articles in peer review journals/paper in proceedings from a Conference or Workshop/section in an edited book or series). Some of the most significant publications achieved so far include the following:
• “Target risk factors for dementia prevention: a systematic review and Delphi consensus study on the evidence from observational studies” (Kay Deckers et al.) International Journal of Geriatric Psychiatry 2015 Mar;30(3):234-46. DOI 10.1002/gps.4245.
• “Promoting modifiable risk factors for dementia: is there a role for general practice?” (K. O’Donnell et al.) British Journal of General Practice 2015 Nov; 65 (640): 567 -568 . http://dx.doi.org/10.3399/bjgp15X687241.
• “Temporal evolution of cognitive changes in incident hypertension”, (Dr Sebastian Köhler et al) Hypertension. 2014 Feb;63(2):245-51. doi: 10.1161/HYPERTENSIONAHA.113.02096. Epub 2013 Dec.
• “Depression, Vascular Factors, and Risk of Dementia in Primary Care: A Retrospective Cohort Study” (S. Kohler et al) Journal of the American Geriatrics Society Vol 63, Issue 4, pages 692–698, DOI 10.1111/jgs.13357.
● “Reducing dementia risk by targeting modifiable risk factors in mid-life: study protocol for the Innovative Midlife Intervention for Dementia Deterrence (In-MINDD) randomised controlled feasibility trial” (Catherine A O’Donnell et al) Pilot and Feasibility Studies 2015, 1:40 DOI: 10.1186/s40814-015-0035-x.
• “Mapping longitudinal studies to risk factors in an ontology for dementia”, (Mark Roantree et al). Health Informatics Journal, DOI 10.1177/1460458214564092.

As some key project results will only be available at the project end (e.g. full analysis of the In-MINDD feasibility study), it has not been possible to publish on all of the project results as yet. However, we are intending to pursue an additional 19 publications. Work on many of these publications is already well advanced.

Presentations

Over the course of the project, members of the In-MINDD team have presented at numerous national and international conferences and events. A total of 73 dissemination activities to date (including numerous presentations) are listed on the Dissemination Register on the EU portal.

Some of the most significant presentations made include the following:
• A presentation by Dr Kate Irving (DCU) at the 5th Annual Translational Medicine Conference in Northern Ireland (2/3 May 2013), which focused on Healthy Ageing.
• Members of the In-MINDD team from DCU and UM presented at the Alzheimer Europe’s 2013 Conference in Malta (10th - 12th October 2013). The presentations made were entitled “Risk factors for dementia: A systematic literature review and Delphi expert consensus” and “Low cognitive activity as a risk factor in cognitive decline”.
• Dr Sebastian Köhler (UM) presented at the 13th meeting of the International College of Geriatric Psychoneuropharmacology (ICGP) in Pittsburgh, USA from 30 October – 2 November 2013. Dr Köhler presented on the In-MINDD project, discussing risk factors and new data regarding depression and dementia.
• Prof Kate O’Donnell from In-MINDD partner GU presented at the “Tackling smoking and dementia summit” in Edinburgh on 4 February 2014. Prof O’Donnell’s presentation was entitled: “Reducing your risk of dementia: are there modifiable risk factors? The European In-MINDD study”.
• Research generated by the In-MINDD Project was presented by Prof Frans Verhey of partner UM at the 22nd European Congress of Psychiatry, in Munich, Germany, on 3 March, 2014. The presentation was entitled “Identifying comorbidity as a strategy for the prevention of dementia.”
• Several In-MINDD team members presented at the 24th Alzheimer Europe Conference in Glasgow from 20-22 October 2014. The Conference focused on “Dignity and autonomy in dementia”. Presentations by In-MINDD partners included the following:
o Dr Maria Pierce (DCU) presented “Primary Prevention of Dementia: Potential for Alignment with Health Promotion Policy”.
o Lisa McGarrigle (DCU) poster presentation “Investigating the Construct Validity of a Model of Cognitive Reserve Based on Protective Factors in Dementia”.
o Dr Susan Browne (GU) presented “Innovative Midlife Intervention for Dementia Deterrence (In-MINDD): a Feasibility Randomised Controlled Trial.”
• Members of the team from In-MINDD partner UM presented at the International Psychogeriatric Association (IPA) Conference in Beijing on 23-25 October 2014. Dr Sebastian Kohler chaired a symposium on “Depression and Dementia—What Are The Underlying Pathways?” and also presented on depression and dementia, exploring evidence from epidemiological studies for the role of vascular factors. Kay Deckers presented a poster entitled “Renal function and dementia risk: a systematic review/meta-analysis“.
• Members of the team from Glasgow University (GU) attended and presented at the North American Primary Care Group Annual Meeting in New York on 21 – 24 November 2014. Dr Susan Browne presented “Innovative Midlife Intervention for Dementia Deterrence (In-MINDD): A Feasibility Randomised Controlled Trial” and Prof Kate O’Donnell delivered a workshop on “Using Normalization Process Theory to Understand Why Complex Interventions Are (or Are Not) Implemented”. The event presented a valuable opportunity to spread the In-MINDD message to a broad audience and engage with colleagues in the US and Canada in particular.
• On 25 June 2015 Kay Deckers from the University of Maastricht presented at the European Conference of Epidemiology 2015. The presentation was entitled "Renal dysfunction and risk for cognitive impairment: a systematic review and meta-analysis".
• Jim O’Donoghue from the Insight Centre for Data Analytics, School of Computing, DCU, presented at the International Joint Conference on Neural Networks (IJCNN) in Killarney, Ireland on 14 July 2015. The presentation was entitled “A Configurable Deep Network for High-Dimensional Clinical Trial Data” and will be published in the conference proceedings. Jim O’Donoghue also presented “A Framework for Selecting Deep Learning Hyper-Parameters” at the 30th British International Conference on Databases (BICOD) on 7 July 2015. The BICOD paper is already published in the Data Science Chapter of the Lecture Notes in Computer Science series.
• Seb Köhler (University of Maastricht) presented “Development and Validation of a New Index for Quantifying a Brain-Healthy Lifestyle: The Innovative Midlife Intervention for Dementia Deterrence (In-Mindd) Study” at the International Psychogeriatric Association International (IPA) Congress 205 in Berlin on 15 October 2015.

Other

In addition to conference presentations and publications in peer reviewed journals, the project has also engaged with other researchers, healthcare professionals and stakeholders in a number of different ways e.g. though University publications; GPs and Alzheimer Associations (short articles showcasing In-MINDD have been featured in the journals and on the websites of national GP Associations and Alzheimer Associations); EU Workshops; collaboration with other projects; engagement with the European Dementia Prevention Initiative (EDPI); exhibits and events and the project mailing list.

Through the In-MINDD project, all partners have extended their links at policy and practice level, with closer links in all countries to local Alzheimer Associations and GP representative bodies. Other examples include in Scotland closer links to The Alliance (a national organisation which brings together groups and individuals focussed on health and social care for chronic conditions) and closer links in Ireland to policy makers in the Department of Health and the Health Service Executive and other stakeholders, such as the Institute of Public Health in Ireland and the Dementia Service Information and Development Service. These links will afford new opportunities to disseminate the findings of In-MINDD to a wider audience.

1.4.5.4 Project Materials

A number of flyers and brochures have been generated over the course of the project. These have been uploaded to the website and printed in paper form for distribution at several conferences and events. There are English versions of all the project materials and also French and Dutch versions of leaflets and posters. We also created a roll-up banner for use at the project’s Final Symposium.

See Figure 9 The In-MINDD public brochure front and back (English version) designed to fold.

See Figure 10 The In-MINDD roll-up banner.

In addition to project leaflets and posters, we also issued three In-MINDD newsletters over the course of the project. These were uploaded to the website, printed in hard copy for distribution and emailed to the project mailing list. As at the end of October 2015, the In-MINDD mailing list comprised over 125 contacts made up of GPs, medical professionals, researchers, Alzheimer’s associations, healthcare providers, health promotion agencies, other projects, policymakers, state bodies and other interested parties. The mailing list has provided a key distribution network for the In-MINDD message.

See Figure 11 The Final Project Newsletter Autumn 2015.

1.4.5.5 In-MINDD Policy Symposium

As policymakers are a key audience for the project, the In-MINDD final conference was structured as a policy symposium, involving policymakers and stakeholders across the EU. The In-MINDD Policy Symposium was held on 1 and 2 October 2015 in Dublin. In total over 50 dementia experts and stakeholders from across Europe attended. The event provided an opportunity to disseminate the project results, but was also a form of action research in itself, providing key players with an opportunity to explore and debate dementia prevention strategies for the future. The Conference sessions included some lively debate and engagement between different stakeholders. The key themes discussed were captured in the notes from the Symposium and included:
• The strength of the evidence base for dementia prevention/delay, risk factors and lifestyle change.
• The need to include a dementia prevention message within other more established chronic disease prevention and awareness policies and programmes.
• The necessity to devise specific awareness and prevention initiatives for dementia risk.
• The need to spread a positive message about long-term brain health (rather than a negative message about risk) and to better define what “brain health” means.
• The importance of harnessing communities at grass roots level in future initiatives.
• The aspiration of reaching all age-groups and acknowledgement that all may benefit from hearing the message about dementia prevention and lifestyle change.
• The possible benefits of lifestyle change that may accrue even to those in the early stages of dementia.
• The advantages of taking a blended approach for interventions, where online tools are combined with face to face contact.

1.4.5.6 Joint Action Plan for Research

The In-MINDD team produced the project’s Joint Action Plan for Research, Memorandum of Understanding and Virtual Research Community at the end of the project in order to provide a framework for the future collaboration of the partners and further innovations in the area of dementia prevention. Full details of our plans are set out in D4.2 which includes details of the following:
• Priority research themes with potential for future collaboration between the partners to build on In-MINDD.
• Plans for our virtual research community (through proposed engagement with the EDPI).
• The consortium’s plans for the future use of the In-MINDD system.
• The group’s intentions for the storage and use of the data generated during the project.
• Establishment of the In-MINDD Legacy Committee.
• Potential future opportunities for researcher exchange.

The team believes that In-MINDD’s joint action plan for research sets out an appropriate framework for the legacy of In-MINDD and an effective blueprint for further research and future collaborations between the partners.

1.4.6 Conclusion

We achieved significant success in our dissemination campaign across several different media, reaching a broad audience through our on-line strategy, articles in peer review journals, conference presentations, project materials, mailing list, media coverage and Policy Symposium. We anticipate that the work of the project will also lead to a number of additional, high impact project publications in the coming weeks and months.

We have also devised a plan for the future use and availability of the In-MINDD system after the end of the project and have drawn up an effective framework for the In-MINDD legacy and for future collaborations.

We believe that the In-MINDD project, and our work within WP4, has made an important contribution to knowledge about, and awareness of, dementia risk reduction and that we have sown the seeds for future policy advances in respect of dementia risk.

We believe that the In-MINDD model and system show real promise for future use to promote long term brain health and to help prevent dementia or delay its onset.

List of Websites:
1.5 Website and contact details

The address for the project website is http://www.inmindd.eu/

Dr Kate Irving (Dublin City University) is the Coordinator of the Project (contact email: kate.irving@dcu.ie).

Final Report Summary - IN-MINDD (Innovative Midlife Intervention for Dementia deterrence)

Verwandte Dokumente

Diese Seite teilen

Herunterladen