Motivation as a mechanism underpinning exercise-based falls prevention programmes for older adults with cognitive impairment: a realist review

Objectives This review aimed to identify mechanisms underlying participation in falls prevention interventions, in older adults with cognitive impairment. In particular we studied the role of motivation. Design A realist review of the literature. Data sources EMBASE, MEDLINE, CINAHL, the Cochrane Library, PsycINFO and PEDRO. Eligibility criteria Publications reporting exercise-based interventions for people with cognitive impairment, including dementia, living in the community. Data extraction and synthesis A ‘rough programme theory’ (a preliminary model of how an intervention works) was developed, tested against findings from the published literature and refined. Data were collected according to elements of the programme theory and not isolated to outcomes. Motivation emerged as a key element, and was prioritised for further study. Results An individual will access mechanisms to support participation when they think that exercise will be beneficial to them. Supportive mechanisms include having a ‘gate-keeper’, such as a carer or therapist, who shares responsibility for the perception of exercise as beneficial. Lack of access to support decreases adherence and participation in exercise. Motivational mechanisms were particularly relevant for older adults with mild-to-moderate dementia, where the exercise intervention was multicomponent, in a preferred setting, at the correct intensity and level of progression, correctly supported and considered, and flexibly delivered. Conclusion Motivation is a key element enabling participation in exercise-based interventions for people with cognitive impairment. Many of the mechanisms identified in this review have parallels in motivational theory. Clinically relevant recommendations were derived and will be used to further develop and test a motivationally considered exercise-based falls intervention for people with mild dementia. PROSPERO registration number CRD42015030169.

The aim of this review is to identify and evidence the mechanisms underlying falls prevention interventions, in older adults with cognitive impairment, to aid intervention development and implementation. The research question is how does an exercise intervention work for people with cognitive impairment, in what circumstances, and why.
Design A realist review of the literature.

Setting and Participants
Publications reporting exercise-based interventions for adults with dementia (any level of cognitive impairment) in the community were included.

Interventions
Exercise-based interventions.

Outcome measures
A "rough programme theory" (a preliminary model of how an intervention works) was tested against published literature identified from a comprehensive search strategy. Data was collected according to elements of the programme theory and not isolated to outcomes.

Results
Motivation is a core element of the programme theory describing participation in exercise-based interventions for people with cognitive impairment. When the individual thinks that exercise will be beneficial they will access supportive mechanisms to enable it. Supportive mechanisms include having a "gate-keeper", such as a therapist or carer, who shares or takes responsibility for the perception of exercise as beneficial, enabling the person with dementia to access and participate in exercise programmes. A lack of access to support had a detrimental effect on adherence and

Introduction
Falls prevention is a complex intervention. In the general older adult population there is conclusive and robust evidence that falls risk can be reduced [1] however the same certainty cannot be provided to people with cognitive impairment [2]. People with cognitive impairment, such as dementia, have a high risk of falling [3][4][5], which frequently results in injury or hospital admission.
However, clinical guidelines are unable to recommend evidence-based interventions due to the paucity of research in this field [6].
Exercise has been shown to reduce falls risk [1]. Anecdotal evidence suggests that older adults with dementia are poor adherers to exercise programmes due to problems with memory, executive function and co-morbidities. Research is needed to explore the contextual factors and mechanisms associated with exercise engagement in older adults with dementia.
Realist synthesis is increasingly used for evaluating evidence for complex health and social interventions [7]. A realist review explores how underlying mechanisms (M) might be 'triggered' in the context (C) of a particular therapy in a particular population to produce an intended or unintended outcome (O). Mechanisms are further subdivided between resources, and responses [8]. Theory is generated and described through this Context-Mechanism-Outcome (CMO) heuristic [9]. CMO Configurations (CMOCs) can be interlinked, creating chains of possibilities and generating theories to explain why a particular outcome occurs with a specific intervention [10]. Interlinking CMOCs can be clustered together to form middle-range theories (MRT), and in turn, the programme theory or model of how an intervention works [7]. Realist methods encourage the F o r p e e r r e v i e w o n l y 5 | P a g e incorporation of data from a range of sources, accommodating complexity that is inherent in health research [11].
Traditional systematic reviews examine the effectiveness of a defined intervention ('does it work?'), as opposed to exploring the underlying mechanisms, which, in theory, may be more generalizable when studying complex interventions in heterogeneous populations. A detailed rationale for completing a realist review in this field has been published and reported in the lead authors PhD thesis [12,13]. There are a limited number of studies in this field, which have used different research methods [14]. Dementia populations vary according to level of impairment (mild to severe) and diagnosis (e.g., Alzheimer's disease, frontotemporal dementia). Therefore, what is relevant to one individual might be different for another. Developing a theoretical framework to rationalise and explain the key principles behind an intervention will aid its development and implementation [15] and take into account the complexity of the population.
The objective of this review was (i) to identify the underlying programme theory for motivation to undertake exercise-based interventions targeted at individuals who have been identified as at risk of falling and who have cognitive impairment, and (ii) to explore how and why that intervention reduces falls in that population. The aim was to produce a list of recommendations that could be used clinically or for further intervention development.

Study Design
The review followed the stages identified by Pawson et al [16] including; i) articulating key rough programme theories to be explored, ii) searching for relevant evidence, iii) appraising the quality of F o r p e e r r e v i e w o n l y 6 | P a g e evidence, iv) extracting the data, and v) synthesising evidence. A detailed protocol has been published [12] and guidance for publication of realist synthesis has been followed [17] and reported.

Scoping
Initial scoping was based on prior knowledge of the falls literature [14] and extensive clinical experience. The purpose was to clarify the aims of the review, develop an initial rough programme theory, and direct the search strategy for the main review.

Search processes
A phased literature search was conducted. An electronic search was completed of databases: EMBASE, MEDLINE, CINAHL, the Cochrane Library, PsycINFO, and PEDRO. Keywords and MESH headings were adapted according to the database used and included: accidental falls, falls rehabilitation, falls prevention, exercise, dementia, cognitive impairment.
The "iterative and interactive" [18] search process evolved during the review, utilising "cited-by" and manual reference list search to find additional primary evidence that was required to refine a particular aspect of the programme theory. The "cited-by" search was completed using Google Scholar for studies that had cited one of the qualitative papers [19]. The reference list of a systematic review [18] was manually searched. Material was chosen for; i) focus on dementia population, ii) qualitative methods, and iii) reporting experiences of completing an exercise or physical-activity intervention. The search results were screened by the researcher (VB) who documented the number of articles retrieved during each search stage using EndNote reference management software.

Selection and appraisal of documents
As is acceptable in realist reviews, material was included regardless of study method but had to focus on an exercise intervention, be published in English and involve community-based participants Titles and abstracts were screened according to relevance of the material to the synthesis aims [20] (Supplementary Table 1). Relevance and rigour were assessed to define the quality of the included material using a series of judgements [18].
The full-text for eligible studies was assessed for quality and extraction of data. A random sample of 10% of the materials was selected and assessed by the stakeholder group comprising rehabilitation and medical clinicians and academics.

Data extraction
Data were extracted based on relevance to the aims of the review and the rough programme theory.
Data were sought that substantiated, refined or refuted the theories. Data regarding contextual characteristics were documented. Relevant material was highlighted, labelled, and recorded [21].
NVivo software and Excel was used to record and code the extracted data.

Analysis and synthesis process
Extracted material was coded as context, mechanism or outcome, and judgements regarding how this influenced the CMOCs recorded through annotations. Codes were initially allocated to each MRT within the rough programme theory, and as each article was processed, these codes were iteratively adapted according to the new material. Material that was relevant to more than one MRT were coded accordingly with links across theories.
Three waves of searching, analysis and synthesis occurred to direct the next stage of the review.
Emerging findings were documented and then discussed with a stakeholder group.

Document flow diagram
The initial search identified 1954 papers ( Figure 1). Sixty-one papers were screened for inclusion.
Sixteen papers from the initial search were not included as theoretical saturation had been reached (e.g., no new findings were emerging with the consideration of new papers). The iterative search identified a further four papers.
Contextual information, including the levels of cognitive impairment and the type, dose, and setting of the interventions, are summarised (Table 1).
"Perception of benefit" response-mechanism The perception or feeling of benefit emerged from seven of the included papers [19,23,25,26,34,36,41]. Perceiving the benefit of an exercise-based intervention could be either a responsemechanism or context, moving within the CMOC depending upon the individual and other context components [19]. Interpretation of Suttanon et al [19] indicates the perception of benefit could be a response-mechanism that is operating when the person with cognitive impairment has "prior experience of being active, participating in exercises, and perceiving benefits of general exercise" (p1180, [19]). These context features are relevant for both the participant and carer [19,25].
Understanding an individual's previous experience of exercise and their perceptions of it can allow tailoring of approaches [19]. That perception or belief of the benefits of exercise may also be a characteristic or feature (context) of the person with cognitive impairment or their carer, which therefore encourages them to participate in the exercise intervention [17].
Recognition of improvements or changes in physiological-responses (for example in function or physical ability) reinforce an individual's perception of benefit [23]. Identification of benefit is important for both participation and the maintenance of an intervention [25,26]. Huger et al [34] identified that persons with cognitive impairment can experience "multiple problems" which could include lack of comprehension. This would influence ability to identify benefits from completing exercise, and whilst this statement seems particularly negative, it could be interpreted as a context component for certain individuals, rather than a general characteristic of all older adults with cognitive impairment.
Synergy is required between carers' understanding and support, ability to address barriers to exercise, and the participants' comprehension [25,41]. Perceiving the health benefits for others also generates the support and encouragement that prompts participation (e.g., an individual walking his dog [25]). The feeling of encouragement could also come from being able to compare The carer's perception and belief in the benefit of exercise must out-weigh the risk, care burden or adaption required to complete the exercise [25,36]. Negative connotations associated with exercising (e.g., reminder of inability to do previously enjoyable activities or potential deterioration), or adaptions or changes to routines or daily lives that are required to support the physical activity, are destructive to the perception of benefit for both the person providing the support and the person with dementia [25]. Concern can be both facilitator and barrier to engagement in exercise, requiring a judgement between not staying mobile and healthy, against concern about getting lost or falling [36].
There was a range of benefits perceived from completing exercise which were not limited to health outcomes. A contentious benefit was an attempt to re-establish previous activities or the "person" that came before the dementia diagnosis or progression of dementia [25]. Cedervall and Aberg [25] reported this perception as coming from the person providing the support. The consideration of how exercise might influence dementia or benefit falls risk was not directly reported [36]. An older adult (C 1 ) with dementia (C 2 ) who is being supported (M response1 ) by a professional person (M resource1 ) who can time-manage (C 7 ), is knowledgeable (C 8 ), firm but encouraging (C 9 ), is kind/friendly/supportive (C 10 ), who understands dementia (C 11 ) and can develop a rapport with the individual (C 12 ), will do an exercise programme (O 1 ).
F o r p e e r r e v i e w o n l y 13 | P a g e supportive strategies (C 18 ), and/or assistance (C 19 ) will feel supported (M response1 ) to complete (O 1 ) an exercise programme (M resource1 ) or routine physical activity (M resource2 ).
An older adult (C 1 ) with more severe dementia (C 20 ) will require more support (M response1 ) to successfully participate (O) in exercise programme (M resource1 ) or routine physical activity (M resource2 ).
An older adult (C 1 ) with dementia/AD (C 2 ) who has a carer (C 13 ) who receives information (C 21 ) and ongoing support (C 22 ) from the therapist/staff (M resource1 ) to enable them to support (M response1 ) the participation/completion of an exercise programme of the person with dementia (O 1 ).
An older adult (C 1 ) with dementia (C 2 ) who wants to exercise in a group (C 23 ), will feel supported (M response1 ) to complete (O 1 ) a group exercise programme (M resouce1 ) or group physical activity (M resource2 ).
An older adult (C 1 ) with dementia (C 2 ) who has a poorer ability to understand and learn new information (C 24 ), who has not exercised previously (C 25 ), has ill-health (C 26 ), or has regular holidays (C 27 ), will not access the required support needed to exercise (M resource1 ) and therefore will not feel supported (M response1 ) to complete an exercise programme(O 1 ).
The support could be provided through supervision [27,34], practical measures [25], strategies such as making or maintaining routines [26] or through emotional support [26,34]. There were many references to who provided the support and how it was given.
Supervision was discussed as a component of support [27]. Supervision by trained personnel "met the special needs of persons with cognitive impairment" (p153, [27]) by; giving clear and repeated instructions, optimally progressing the programme, and providing the amount of supervision required depending upon their ability to understand and learn new information [27]. Training instructors or supervisors provide more than just formal support during an intervention [34] and were influential in the commencement, participation and maintenance of exercise [19]. Key characteristics of the professional person were identified [19]. An ability to "understand my problem" (p172, [36]) also emerged as important, particularly in regards to dementia. This facilitated rapport development between supporter and person with dementia, which included a relationship built on personal information [38] and trust [41].
Carers were a frequently discussed and important component with the support they provided [40].
The role of the carer was described by Malthouse and Fox [36] as "facilitators to activity" and "gatekeepers". There were many ways in which the carers provided support including the avoidance of

Summary of findings
Motivation is a core element of the programme theory describing participation in exercise-based interventions for people with cognitive impairment. When an older person with mild to moderate cognitive impairment has the perception that exercise will be beneficial they can utilise supportive mechanisms and contexts to complete an exercise programme. Support as a motivational mechanism requires a "gate-keeper", such as a therapist or carer, who shares or takes responsibility for the perception of exercise as beneficial, thereby enabling the person with dementia to access and participate in exercise programmes. Having a perception of benefit is both a mechanism and contextual feature within this programme theory. A lack of access to support had a detrimental effect on adherence and participation in exercise.

Strengths and limitations
This review progress the falls prevention research by using a novel method and approach. The main strength of this review is the successful completion of realist rationale in a historically positivist research field which demands causal probabilities but which contains practice and policy directed by evidence. The realist review methodology was well-suited to the research question. Consideration of the mechanisms underpinning exercise-based interventions allowed development and extrapolation of the theoretical rationale. Exploring and documenting context components allows individualisation.
Transparency is encouraged in realist methods of enquiry and is evident in this review regarding the interpretation of material being influenced by the researcher. The researcher is acknowledged in this review and, whilst considered from a traditional systematic review method as producing the potential for bias, has assisted in the theory development and interpretations. Detailed recognition of underlying or "hidden" mechanisms and understanding of the CMOCs was strengthened by the the review not expanding to consider meso (institutional) or macro (government and policy) levels of social structure [43]. The review also did not base the initial rough programme theory on any overarching motivational theories (such as Self-Determination Theory [44,45]). Theoretical frameworks are typically consulted to structure realist reviews (e.g., search strategies and data analysis) [43,46]. This was not done in this review and is potentially a limitation. However, the programme theory and its subsequent recommendations are very specific and should be viewed as a considerable strength.

Recommendation Who
Older adults with mild to moderate cognitive impairment If a person with dementia has the belief that exercise is advantageous, a positive attitude to exercise, the ability to understand the benefits of exercise or is able to identify the physical or functional changes from doing exercise, then they will perceive the benefit of doing exercise If a person with dementia perceives the benefit, they will participate in exercise-based intervention • provide clear and repeated instructions • optimally progress the exercises • provide the amount of supervision required by that individual and their needs • understand the needs of persons with dementia If support is being provided by a carer, then the intervention should provide information and on-going support to enable them to continue Carers supporting an intervention should; • perceive and understand the benefit of the person with dementia doing exercise • provide transport or consider practical arrangements for access to the intervention • have a positive attitude • implement supportive strategies and/or assistance in the manner required by the person with dementia If the carers or supporters perception of the benefits of doing exercise out-weighs the risk, concern, or burden of extra care duties, then the intervention will be encouraged Benefits of exercise perceived by the carer or supporter for the person with dementia include; mood, behaviour, weight, flexibility, ageing, enjoyment of everyday life

Patient involvement
Patients were not involved.

Data sharing statement
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.

Instructions to authors
Complete this checklist by entering the page numbers from your manuscript where readers will find each of the items listed below.
Your article may not currently address all the items on the checklist. Please modify your text to include the missing information. If you are certain that an item does not apply, please write "n/a" and provide a short explanation.
Upload your completed checklist as an extra file when you submit to a journal.
In your methods section, say that you used the PRISMA reporting guidelines, and cite them as: and report characteristics (e.g., years considered, language, publication status) used as criteria for eligibility, giving rational 6-7 Information sources #7 Describe all information sources in the search (e.g., databases with dates of coverage, contact with study authors to identify additional studies) and date last searched.

6-7
Search #8 Present full electronic search strategy for at least one database, including any limits used, such that it could be repeated.

6-7
Study selection #9 State the process for selecting studies (i.e., for screening, for determining eligibility, for inclusion in the systematic review, and, if applicable, for inclusion in the meta-analysis).

6-7
Data collection process #10 Describe the method of data extraction from reports (e.g., piloted forms, independently by two reviewers) and any processes for obtaining and confirming data from investigators.

6-7
Data items #11 List and define all variables for which data were sought (e.g., PICOS, funding sources), and any assumptions and simplifications made.

6-7
Risk of bias in individual studies #12 Describe methods used for assessing risk of bias in individual studies (including specification of whether this was done at the study or outcome level, or both), and how this information is to be used in any data synthesis. Limitations #25 Discuss limitations at study and outcome level (e.g., risk of bias), and at review level (e.g., incomplete retrieval of identified research, reporting bias).

17-18
Conclusions #26 Provide a general interpretation of the results in the context of other evidence, and implications for future research.

21
Funding #27 Describe sources of funding or other support (e.g., supply of data) for the systematic review; role of funders for the systematic review.

23-24
The PRISMA checklist is distributed under the terms of the Creative Commons Attribution License CC-BY. This checklist was completed on 20. June 2018 using http://www.goodreports.org/, a tool made by the EQUATOR Network in collaboration with Penelope.ai 17 Results

18
An individual will access mechanisms to support participation when they think that exercise will be 19 beneficial to them. Supportive mechanisms include having a "gate-keeper", such as a carer or 20 therapist, who shares responsibility for the perception of exercise as beneficial. Lack of access to 21 support decreases adherence and participation in exercise. Motivational mechanisms were 22 particularly relevant for older adults with mild to moderate dementia, where the exercise intervention 23 was multicomponent, in a preferred setting, at the correct intensity and level of progression, correctly 24 supported and considered, and flexibly delivered. (in contrast to systematic reviews which identify whether an intervention works or not), 17 enabling exploration of contextual factors and underpinning mechanisms associated with 18 exercise, and thus the necessary conditions for participation.

19
 We developed a programme theory explaining the role of motivation in exercise participation, 20 and recommendations for clinicians to support exercise components of falls intervention 21 programmes for older adults with mild cognitive impairment.

22
 Some elements of the programme theory were not well supported by evidence, limiting the 23 depth and detail of the recommendations, in particular around the role of exercise specifically 24 in falls prevention.

Introduction
2 Falls prevention represents a complex intervention due to the multifactorial causes of falls. There is 3 robust evidence that falls risk can be reduced in the general older adult population [1], but for people 4 with dementia the effectiveness of falls prevention is uncertain [2]. People with cognitive impairment 5 have a high risk of falling [3][4][5], which frequently results in injury or hospital admission, but clinical 6 guidelines cannot recommend evidence-based interventions [6]. 7 Exercise, at the correct dose and intensity, reduces falls risk [1]. Motivation is defined as 'the 8 energisation and direction of behavior' [7]. A key challenge is how to motivate older adults to achieve 9 sufficient exercise participation and adherence to obtain such benefits, and this is especially so for 10 people living with dementia. Exercise interventions with older adults with dementia have reported 11 varying levels of adherence [8][9][10][11]. People with dementia undertake less physical activity compared 12 to those without dementia [12]. A range of factors (e.g., problems with memory, executive function, 13 carer burden, and co-morbidities) can influence exercise motivation [13]. However, people living with 14 dementia populations vary according to level of impairment (mild to severe), diagnosis (e.g.,

15
Alzheimer's disease, frontotemporal dementia) and support (e.g., carer availability subdivided between resources, and responses [15]. Theory is generated and described through this 2 Context-Mechanism-Outcome (CMO) heuristic [16]. CMO Configurations (CMOCs) can be linked, 3 creating chains of possibilities and generating theories to explain why a particular outcome occurs 4 with a specific intervention [17]. Interlinking CMOCs can be clustered together to form 'middle-range 5 theories' (MRT), and in turn, a 'programme theory' or model of how an intervention works [14].  Table 3). Quality appraisal of the articles was completed in 8 accordance with standard practice of realist reviews [25]. Relevance and rigour were assessed using 9 a series of judgements to define the quality (Supplementary Table 3).

10
The full-text for eligible studies was assessed for quality and extraction of data by one researcher (VB).

11
A random sample of 10% of the materials was selected and assessed by a stakeholder group 12 comprising rehabilitation and medical clinicians and academics.

13
Data extraction 14 Data were extracted based on relevance to the aims of the review and the rough programme theory. 15 Data were sought that substantiated, refined or refuted the theories, and described contextual 16 characteristics. Relevant material was highlighted, labelled, and recorded [28]. NVivo software and 17 Excel was used to record and code the extracted data.

18
Analysis and synthesis process 19 Extracted material was coded as context, mechanism or outcome, and judgements regarding how this 20 influenced the CMOCs recorded through annotations. Codes were initially allocated to each MRT 21 within the rough programme theory, and as each article was processed, these codes were iteratively 22 adapted according to the new material. Material that was relevant to more than one MRT were coded 23 accordingly with links across theories.

5
Document flow diagram 6 The initial search identified 1954 papers ( Figure 1). Sixty-one papers were screened for inclusion.

7
Sixteen papers from the initial search were not included as theoretical saturation had been reached 8 (e.g., no new findings were emerging with the consideration of new papers). The iterative search 9 identified a further four papers.

9
"Perception of benefit" response-mechanism The perception or feeling of benefit (Box 1 [20]) emerged from seven studies [13, 26, 30, 32, 33, 41, 4 43]. Perceiving the benefit of an exercise-based intervention could be either a response-mechanism 5 or context, depending upon the individual and other context components [26]. The perception of 6 benefit could be a response-mechanism that is operating when the person with cognitive impairment 7 has "prior experience of being active, participating in exercises, and perceiving benefits of general  [13,32]. Perceiving the health benefits for others also generates remaining mobile and healthy, against concern about getting lost or falling [43].

12
There was a range of perceived benefits from completing exercise which were not limited to health 13 outcomes. A contentious benefit was an attempt to re-establish previous activities or the "person"     [13] 23 suggest that because persons with dementia require care and support, they are more influenced by 24 support as a variable within an intervention. Carers themselves also required support, with 25 information identified as a resource-mechanism to enable the support to take place [13,26].  Intervention that is provided flexibly (C 8 ), for 6-12 months (C 9 ), 2-3 times a week (C 10 ), for minimum 15-20 minutes or whatever can be done or fit in with routine (C 11 ).

Discussion
2 Summary of findings 3 The review revealed motivation to be a core element of the programme theory underlying falls 4 prevention interventions in older adults with cognitive impairment. Within the motivation component 5 of the programme, two key mechanisms, perceived benefit and support, were shown to influence the 6 extent to which an older adult with cognitive impairment is motivated to undertake an exercised-base 7 intervention. When an older person with mild to moderate cognitive impairment believes that 8 exercise will be beneficial they can utilise supportive mechanisms and contexts to complete an 9 exercise programme. Support as a motivational mechanism requires a "gate-keeper", such as a 10 therapist or carer, who shares or takes responsibility for the perception of exercise as beneficial,

22
Transparency is encouraged in realist methods. The potential influence of the researcher in 23 interpretation is acknowledged and, whilst being a potential source of bias, also assisted in the theory 24 development and interpretations. Recognition of underlying or "hidden" mechanisms and

Focus Recommendation
Older adults with mild to moderate cognitive impairment If a person with dementia has the belief that exercise is advantageous, a positive attitude to exercise, the ability to understand the benefits of exercise or is able to identify the physical or functional changes from doing exercise, then they will perceive the benefit of doing exercise

Who
If a person with dementia perceives the benefit, they will participate in exercise-based intervention

What
Multicomponent exercise-based intervention that;  combines physical (including strength/resistance, balance, endurance/mobility, aerobic) and cognitive exercises  appropriately intensive and progressive  supported by suitable staff (who can interact, communicate and connect) and materials  considers speed of initiation, length of intervention, encouragement of active lifestyle and enjoyment  is delivered in a flexible manner for at least 15-20 minutes (or whatever can become or fit in with routine) 2-3 times a week for 6-12 months  can be delivered at home (for those wanting or needing 1:1 support from the intervention staff) or in a group (for those wanting carer respite, increase in habitual physical activity or socialising aspects) Support can provide encouragement for completing an exercise-based intervention Sources of support can include but are not exclusively supplied by; trained intervention staff, carer, spouse, family member If support is being provided by trained intervention staff, then they should have professional competence including;  time-management  knowledgeable  firm but encouraging  kind, friendly and supportive  understanding of the issues experienced by persons with dementia  rapport development Trained intervention staff supporting an intervention should;  provide clear and repeated instructions  optimally progress the exercises  provide the amount of supervision required by that individual and their needs  understand the needs of persons with dementia If support is being provided by a carer, then the intervention should provide information and on-going support to enable them to continue Carers supporting an intervention should;  perceive and understand the benefit of the person with dementia doing exercise  provide transport or consider practical arrangements for access to the intervention  have a belief in the benefit of exercise  implement supportive strategies and/or assistance in the manner required by the person with dementia If the carers or supporters perception of the benefits of doing exercise out-weighs the risk, concern, or burden of extra care duties, then the intervention will be encouraged

Circumstances
Benefits of exercise perceived by the carer or supporter for the person with dementia include; mood, behaviour, weight, flexibility, ageing, enjoyment of everyday life                      (accidental fall OR fall* OR fall risk OR postural balance) AND (mild cognitive impairment OR dementia OR Alzheimer disease OR cognit* impair*) AND (Aged OR elderly OR elder OR frail OR old* OR seniors OR geriatric OR older adult*) AND (Exercise OR "physical exercise" OR "exercise therapy" OR "physical activity" OR balance OR "resist* training" OR strength)  Could this be about the strength and balance exercise component of falls rehabilitation in older adults with cognitive impairment in the community?

Abstracts
Could this material provide useful information about completing the strength and balance exercise component of falls rehabilitation in older adults with cognitive impairment in the community?

Quality Appraisal of Articles
Is the material cohesive? Does it tell a comprehensive story or is there a juxtaposition of ideas or isolated statements?
What is the value of the evidence? Does it contribute to the topic area?

19
 We developed a programme theory explaining the role of motivation in exercise participation, 20 and recommendations for clinicians to support exercise components of falls intervention 21 programmes for older adults with mild to moderate cognitive impairment.

22
 Some elements of the programme theory were not well supported by evidence, limiting the 23 depth and detail of the recommendations, in particular around the role of exercise specifically 24 in falls prevention.  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60   F  o  r  p  e  e  r  r  e  v  i  e  w  o  n  l  y  Introduction   5 Falls prevention represents a complex intervention due to the multifactorial causes of falls. There is 6 robust evidence that some interventions can reduce falls risk in the general older adult population [1, 7 2], but for people with dementia the effectiveness of falls prevention is uncertain [3,4]. People with 8 cognitive impairment have a high risk of falling [5][6][7], which frequently results in injury or hospital 9 admission, but clinical guidelines cannot recommend evidence-based interventions [8].
10 Exercise, at the correct dose and intensity, reduces falls risk [1]. Motivation is defined as 'the 11 energisation and direction of behavior' [9]. A key challenge is how to motivate older adults to achieve 12 sufficient exercise participation and adherence to obtain such benefits, and this is especially so for 13 people living with dementia. Exercise interventions with older adults with dementia have reported 14 varying levels of adherence [10][11][12][13]. People with dementia undertake less physical activity compared 15 to those without dementia [14]. A range of factors (e.g., problems with memory, executive function, 16 carer burden, and co-morbidities) can influence exercise motivation [15]. However, people living with 17 dementia populations vary according to level of impairment (mild to severe), diagnosis (e.g.,

18
Alzheimer's disease, frontotemporal dementia) and support (e.g., carer availability  subdivided between resources, and responses [17]. Theory is generated and described through this 18 The objective of this review was (i) to identify the underlying programme theory for participation in 19 exercise-based falls prevention interventions in older adults with cognitive impairment, and (ii) to 20 explore how and why that intervention reduces falls. The aim was to produce a list of 21 recommendations that could be used clinically or to inform further intervention development. Excel was used to record and code the extracted data. Analysis and synthesis process 2 Extracted material was coded as context, mechanism or outcome, and judgements regarding how this 3 influenced the CMOCs recorded through annotations. Codes were initially allocated to each MRT 4 within the rough programme theory, and as each article was processed, these codes were iteratively 5 adapted according to the new material. Material that was relevant to more than one MRT were coded 6 accordingly with links across theories. 7 Three waves of searching, analysis and synthesis occurred to direct the next stage of the review.

8
Emerging findings were documented and then discussed with the stakeholder group.

9
Patient and public involvement 10 Patients were not involved.

4
Perceiving the benefit of an exercise-based intervention could be either a response-mechanism or 5 context, depending upon the individual and other context components [28]. The perception of benefit 6 could be a response-mechanism that is operating when the person with cognitive impairment has 7 "prior experience of being active, participating in exercises, and perceiving benefits of general exercise" 8 (p1180, [28]) and is applicable to both the participant and carer [28,33]. Understanding an individual's 9 previous experience of exercise and their perceptions of it can allow tailoring of approaches [28].
10 Perception or belief in the benefits of exercise may also be a characteristic or feature (context) of the 11 person with cognitive impairment or their carer, which encourages them to participate in the exercise 12 intervention [17].

21
Synergy is required between carers' understanding and support, ability to address barriers to exercise, 22 and the participants' comprehension [15,33]. Perceiving the health benefits for others also generates 17 Support

An older adult (C 1 ) with dementia (C 2 ) who is supervised by trained personnel (M resource1 ) who give clear/repeated instructions (C 3 ), optimally progress the exercises (C 4 ), provide the amount of supervision required by that individual and their needs (C 5 ), and understands the needs of persons with dementia (C 6 )
will feel supported (M response1 ) to complete (O 1 ) an exercise programme (M resource2 ).

An older adult (C 1 ) with dementia (C 2 ) who is being supported (M response1 ) by a professional person (M resource1 )
who can time-manage (C 7 ), is knowledgeable (C 8 ), firm but encouraging (C 9 ), is kind/friendly/supportive (C 10 ), who understands dementia (C 11 ) and can develop a rapport with the individual (C 12 ), will do an exercise programme (O 1 ). between supporter and person with dementia, which included a relationship built on personal 5 information [46] and trust [15]. 6 Carer involvement was frequently reported and was an important component regarding the support 7 they provided [48]. The role of the carer was described by Malthouse and Fox [44] as "facilitators to 8 activity" and "gate-keepers". There were many ways in which carers provided support including the  [36,45]. In others they did not [15,44,45,48], 17 indicating the complexity of the caring role and the feelings associated with it.

18
Carers provided varying levels of support that were tailored to the individual [28,33]. Carer 19 involvement was integral to programme delivery in one study [28]. The carer and their support was 20 more influential the more severe the cognitive impairment and may account for why people with 21 severe dementia were still able to engage in interventions [48]. However, it was highlighted how 22 complex the support component is, particularly as impairments progress [48]. Van Alphen et al [15] 23 suggest that because persons with dementia require care and support, they are more influenced by 24 support as a variable within an intervention. Carers themselves also required support, with 25 information identified as a resource-mechanism to enable the support to take place [15,28].  Intervention that is provided flexibly (C 8 ), for 6-12 months (C 9 ), 2-3 times a week (C 10 ), for minimum 15-20 minutes or whatever can be done or fit in with routine (C 11 ).

17
The "dose" of an intervention is a combination of frequency, duration, and intensity. A range of doses 18 was reported (Table 1)

Discussion
2 Summary of findings 3 The review revealed motivation to be a core element of the programme theory underlying falls 4 prevention interventions in older adults with cognitive impairment. Within the motivation component 5 of the programme, two key mechanisms, perceived benefit and support, were shown to influence the 6 extent to which an older adult with cognitive impairment is motivated to undertake an exercised-base 7 intervention. When an older person with mild to moderate cognitive impairment believes that 8 exercise will be beneficial they can utilise supportive mechanisms and contexts to complete an 9 exercise programme. Support as a motivational mechanism requires a "gate-keeper", such as a 10 therapist or carer, who shares or takes responsibility for the perception of exercise as beneficial, 11 thereby enabling the person with dementia to access and participate in exercise programmes. A 12 perception of benefit is both a mechanism and contextual feature within this programme theory. Lack 13 of access to support had a detrimental effect on adherence and participation in exercise.
14 Strengths and limitations 15 This review progresses falls prevention research by using a novel approach. The main strength of this 16 review is the successful completion of realist rationale in a historically positivist research field which 17 prioritises causal probabilities over generalisability; an intervention may benefit a group on average, 18 but we can be unsure if a given individual will benefit or be harmed. The realist review methodology 19 was well-suited to the research question. Consideration of the mechanisms underpinning exercise-20 based interventions allowed development and extrapolation of the theoretical rationale. Exploring 21 and documenting context components allows individualisation.

22
Transparency is encouraged in realist methods. The potential influence of the researcher in 23 interpretation is acknowledged and, whilst being a potential source of bias, also assisted in the theory 24 development and interpretations. Recognition of underlying or "hidden" mechanisms and All of the materials included within the review described participants who had either completed 5 regular physical-activity or the exercise-based intervention under study. The perspective of those not 6 completing an exercise-based intervention must be considering for further programme theory 7 refinement particularly considering the motivational mechanisms. 8 Further research could focus on the assessment and/or measurements of these mechanisms, for 9 example by investigating the assessment of perceived benefit through use of measurements or scales.

Focus Recommendation
Older adults with mild to moderate cognitive impairment If a person with dementia has the belief that exercise is advantageous, a positive attitude to exercise, the ability to understand the benefits of exercise or is able to identify the physical or functional changes from doing exercise, then they will perceive the benefit of doing exercise

Who
If a person with dementia perceives the benefit, they will participate in exercise-based intervention

What
Multicomponent exercise-based intervention that;  combines physical (including strength/resistance, balance, endurance/mobility, aerobic) and cognitive exercises  appropriately intensive and progressive  supported by suitable staff (who can interact, communicate and connect) and materials  considers speed of initiation, length of intervention, encouragement of active lifestyle and enjoyment  is delivered in a flexible manner for at least 15-20 minutes (or whatever can become or fit in with routine) 2-3 times a week for 6-12 months  can be delivered at home (for those wanting or needing 1:1 support from the intervention staff) or in a group (for those wanting carer respite, increase in habitual physical activity or socialising aspects) Support can provide encouragement for completing an exercise-based intervention Sources of support can include but are not exclusively supplied by; trained intervention staff, carer, spouse, family member If support is being provided by trained intervention staff, then they should have professional competence including;  time-management  knowledgeable  firm but encouraging  kind, friendly and supportive  understanding of the issues experienced by persons with dementia  rapport development Trained intervention staff supporting an intervention should;  provide clear and repeated instructions  optimally progress the exercises  provide the amount of supervision required by that individual and their needs  understand the needs of persons with dementia If support is being provided by a carer, then the intervention should provide information and on-going support to enable them to continue Carers supporting an intervention should;  perceive and understand the benefit of the person with dementia doing exercise  provide transport or consider practical arrangements for access to the intervention  have a belief in the benefit of exercise  implement supportive strategies and/or assistance in the manner required by the person with dementia If the carers or supporters perception of the benefits of doing exercise out-weighs the risk, concern, or burden of extra care duties, then the intervention will be encouraged

Circumstances
Benefits of exercise perceived by the carer or supporter for the person with dementia include; mood, behaviour, weight, flexibility, ageing, enjoyment of everyday life   CMO: context-mechanism-outcome; CMOC: context-mechanism-outcome configuration;   Describe the analysis and synthesis processes in detail. This section should include information on the constructs analyzed and describe the analytic process.   (accidental fall OR fall* OR fall risk OR postural balance) AND (mild cognitive impairment OR dementia OR Alzheimer disease OR cognit* impair*) AND (Aged OR elderly OR elder OR frail OR old* OR seniors OR geriatric OR older adult*) AND (Exercise OR "physical exercise" OR "exercise therapy" OR "physical activity" OR balance OR "resist* training" OR strength)  Could this be about the strength and balance exercise component of falls rehabilitation in older adults with cognitive impairment in the community?

Abstracts
Could this material provide useful information about completing the strength and balance exercise component of falls rehabilitation in older adults with cognitive impairment in the community?
Quality Appraisal of Articles Is the material cohesive? Does it tell a comprehensive story or is there a juxtaposition of ideas or isolated statements?
What is the value of the evidence? Does it contribute to the topic area?
What is the material's position in relation to the programme theory and general topic area?