Visually Impaired OLder people's Exercise programme for falls prevenTion (VIOLET): a feasibility study

Article history

The research reported in this issue of the journal was funded by the PHR programme as project number 12/181/07. The contractual start date was in September 2014. The final report began editorial review in January 2017 and was accepted for publication in July 2017. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The PHR editors and production house have tried to ensure the accuracy of the authors’ report and would like to thank the reviewers for their constructive comments on the final report document. However, they do not accept liability for damages or losses arising from material published in this report.

Declared competing interests of authors

Dawn Skelton receives personal fees from Later Life Training Ltd outside the submitted work and is a Director of Later Life Training, a company that runs as a not-for-profit organisation delivering Falls Management Exercise (FaME) Postural Stability Instructor (PSI) training to health and fitness professionals across the UK.

Copyright statement

© Queen’s Printer and Controller of HMSO 2018. This work was produced by Adams et al. under the terms of a commissioning contract issued by the Secretary of State for Health and Social Care. This issue may be freely reproduced for the purposes of private research and study and extracts (or indeed, the full report) may be included in professional journals provided that suitable acknowledgement is made and the reproduction is not associated with any form of advertising. Applications for commercial reproduction should be addressed to: NIHR Journals Library, National Institute for Health Research, Evaluation, Trials and Studies Coordinating Centre, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK.

2019 Queen’s Printer and Controller of HMSO

The literature search

To contextualise and evidence the Visually Impaired OLder people’s Exercise programme for falls prevenTion (VIOLET) study, a narrative review of the relevant literature was carried out. A literature review was undertaken in January 2016 to support the publication of the study protocol. The search terms ‘falls’, ‘fear of falling’, ‘visual impairment’, ‘sight loss’, ‘elderly’, ‘older people’, ‘older adult’ and their combinations were used to search the Cumulative Index to Nursing and Allied Health Literature (CINAHL), Cochrane, MEDLINE (Medical Literature Analysis and Retrieval System Online, or MEDLARS Online) and PubMed databases. The grey literature was also searched to include significant government and other reports, such as work by the Royal National Institute for Blind People (RNIB) and reference lists of key studies provided a further literature source. The results are discussed below.

Falls and visual impairment

In the UK, one in eight people aged > 75 years and one in three aged > 90 years live with significant sight loss. 1 Visual impairment (VI) is associated with a loss of function in activities of daily living,2,3 and visually impaired older people (VIOP) are more likely than sighted peers to move into residential settings, be physically dependent and have poor quality of life (QoL). 4–6 Gleeson et al. 7 state that VI is an independent risk factor for falls. 5,8,9 Poor vision has been related to the recurrence of falls. 10 Waterman et al. 11 cite research identifying a number of falls risk factors associated with VIOP including muscle weakness, impaired balance and – specific to sight – poor visual acuity and contrast sensitivity, decreased depth perception and reduced visual field. 12,13 However, VI is not independently associated with a higher incidence of falls, despite being a risk factor. 14

Among the general population, falls in older people are common and can be life limiting and debilitating, and estimates suggest that, each year, between 30% and 62% of older people fall. 15,16 Falls are seldom due to a single cause6,12 and are associated with considerable morbidity and mortality,17 with approximately 10% of falls resulting in fractures. 18 The economic costs of falls are estimated at 0.07–0.20% of gross domestic product, with falls and their consequences accounting for 0.85–1.5% of Western economies’ total health-care expenditure. 19 In 2015 in the UK alone, falls in the over-65-year-olds were estimated to cost £4.6M per day. 20

Older people with VI have a (1.7-times) higher risk of falling than the general population, more hospital and nursing home admissions and report more contact with their general practitioner (GP) than those who do not have VI. 21,22 Although VI is not independently associated with a higher incidence of falls,14 Boyce22 suggests that in the UK 8% of falls-related hospital admissions are likely to occur in visually impaired people, accounting for 21% of the total cost of treating accidental falls. Such estimates are problematic in that costs to GPs may not be included in treatment costs, reasons for falling may not be known or recorded and aggregate estimates are likely to mask wide geographical variation.

Evidence suggests that multifactorial falls intervention programmes are effective in reducing falls among older people, often tackling underlying health problems, initiating strength and balance training and offering home modifications. 23–25 In 2004, the National Institute for Health and Care Excellence (NICE) guidelines on prevention and assessment of falls in older people25 suggested that by 2005 all NHS boards in the UK should provide a falls service. This echoed the earlier call in 2001 from the National Service Framework for Older People for specialist falls services in both hospital and community settings. 26 A 2008 survey noted that just half of participating falls clinics in the UK assessed for VI and there was little information about how to modify falls prevention programmes for VIOP. 27

Fear of falling and visual impairment

Fear of falling (FoF), with or without a fall, is also common in VIOP and can lead to a cycle of restricting daily activity and mobility, loss of confidence, diminishing physical and mental assets and reduced social participation and overall QoL. 28–30 A UK report by Visibility, a Glasgow-based vision charity, found that older people are highly likely to avoid activity because of VI. 31 Anxiety and depression are also common in those with VI and this may also lead to reduced activity. 1 In a recent randomised controlled trial (RCT) of a cognitive–behavioural therapy-based intervention to reduce FoF in older people, Parry et al. 17 suggest that FoF is likely to affect between 3% and 85% of community-dwelling older people who fall and up to 50% of those who have never fallen. The authors also cite research highlighting that FoF may lead to avoidance of activity, increasing frailty and risk of further falls independent of physical impairment. 32–36

Fear of falling is a significant predictor of a future fall. 37 Clinical and laboratory observations have shown that concerns about falling have a negative impact on older people’s gait patterns. For example, studies of gait and balance in older people’s use of elevated walkways showed both dysfunctional gait adjustments and disproportionately slow walking speeds38 as well as postural balance abnormalities39 when compared with younger subjects. It is also clear that there is a distinction between proper caution around activity and overcaution/a fear–avoidance cycle that perpetuates disability. 40,41 Indeed, a recent review suggests that FoF contributes to falls risk independent of actual gait or balance problems. 42

A recent Cochrane review investigating three-dimensional exercise – yoga, tai chi and strength and resistance training – in community-dwelling older people found that exercise can reduce FoF, but only in the short term, with evidence for long-term efficacy lacking. 43 The authors suggest that future exercise for older people intervention trials should have core outcomes that include FoF. Parry et al. 17 cite a systematic review that identified 12 high-quality RCTs, each with FoF as an outcome,35 but only one of these trials aimed to reduce FoF. 44 The interventions included a range of settings, from community tai chi to home-based multifactorial interventions, and all reported reduced FoF. However, Parry et al. 17 also highlight a more recent multifactorial intervention study, based in a geriatric outpatient setting, that did not find such a benefit. 45

There are also limited health economic data about FoF interventions. 40,42

Measuring the number of falls as a primary outcome does not take into account the more complex impact of an intervention that, by reducing FoF, may increase participants’ confidence in their ability to walk safely and continue to enjoy everyday activities. 46

In relation to VIOP, FoF can also be a barrier to uptake of exercise programmes. Research eliciting the views of VIOP28,31 suggests that perceived risk, stigma, lack of awareness among health professionals, lack of appropriate supporting materials and FoF may be barriers to exercise attendance. Enabling factors could include peer acceptability, appropriate supporting material with demonstration, sensitive explanation, carer involvement and individually tailored interventions.

Fear of falling is thus a vital outcome in falls intervention research, and there are important reasons for its inclusion in studies in VIOP. Any such intervention needs to include targeting of FoF, ameliorating its adverse outcomes and appropriate quantification, FoF being a key outcome measure.

Interventions for falls in visually impaired older people

Gleeson et al. 7 highlight well-designed exercise programmes that reduce falls in the general population44,47 but suggest that such programmes have not been successful in community-dwelling older adults with VI. For example, a recent Cochrane systematic review reports that home safety assessment and multicomponent home-based exercise programmes reduce the rate of falls and risk of falling in community-dwelling older people. 48 However, the New Zealand-based VIP trial49 did not demonstrate such results with VIOP. This study assessed the efficacy and cost-effectiveness of home safety and home exercise interventions for preventing falls and falls-related injuries in community-dwelling people aged ≥ 75 years with severe VI. The VIP trial included 391 participants with visual acuity of 6/24 or worse, of whom 361 (92%) completed 1 year of follow-up. The home safety assessment and modification programme was specifically designed for people with severe VIs and the 1-year home exercise intervention used the Otago exercise programme. 50 This programme includes specific muscle strengthening and balance retraining exercises that increase in difficulty and a walking plan, modified for those with severe visual acuity loss. The home safety modifications, delivered by an occupational therapist (OT), resulted in a significant reduction in the risk of falling while the group receiving home exercise, managed by a physiotherapist, showed a non-significant increase in the risk of falls. However, the authors highlighted that within the exercise programme stricter adherence to the prescription was associated with fewer falls.

Waterman et al. 11 suggest that the VIP study participants who received both home safety and home exercise interventions may have been unwittingly given confusing or conflicting messages from the OT and physiotherapist, this having a negative impact on intervention adherence and study outcome. 28 Adherence to the Otago programme was significantly lower in VIOP than in older people without significant VI in this study (only 18% of VIOP completed all home exercise sessions over a 1-year period)49 and this may have been due to lack of confidence in exercising at home without supervision.

The VIP2UK trial11 was adapted from the VIP Otago home-based programme to increase adherence in VIOP. This community-based feasibility study was carried out in north-west England and recruited 49 community-dwelling VIOP into a three-arm RCT with a control group receiving usual care and social visits, an experimental group receiving a home safety programme and another group receiving a home safety and home exercise programme. The social visits for the group comprised three social visits and two telephone calls and were designed to control for social contact that may reduce falls and influence lifestyle or QoL variables. 11

Falls primary outcome data were collected continuously over 6 months. Secondary outcomes on self-reported and instrumented physical activity and adherence were collected at baseline and at 3 and 6 months for home exercise and at 6 months for the home safety programme. All interventions were delivered by an OT. A total of 43 out of 49 participants (88%) completed the trial and 6-month follow-up. At follow-up, 100% reported partially or completely adhering to home safety recommendations but evidence for adherence to home exercise was equivocal. Whereas self-reported physical activity increased, instrumented monitoring [activPAL™ (PAL Technologies, Glasgow, UK)] showed a decrease in walking activity and there were no statistically significant differences in falls between the groups. However, the authors highlight that the study was not powered to detect a difference.

The VIP2UK authors suggest that further research is needed to explore the differences between the subjective and objective measures of activity and adherence issues. Whereas home safety modifications required a ‘one-off’ input from participants, ongoing adherence to the physical exercise programme (the Otago programme) might have been particularly challenging for the VIOP, some of whom lived alone. Despite the use of mentors, extra telephone calls and audio exercise clips and embedding exercises into daily living, it might have required a lot of effort to follow instructions via large print or audio and continue to be motivated to complete home exercises. The authors conclude that perhaps weekly home visits from an exercise trainer may improve adherence for VIOP or that group exercise with tailored support dependent on VI may lead to better progression in relation to strength and balance and, relatedly, to falls.

Finally, Gleeson et al. 7 highlight three small studies demonstrating that multimodal exercise and tai chi may improve physical functioning in VIOP51–53 in controlled environments when physical and verbal guidance is provided. These were carried out in residential settings and cannot be generalised to community-dwelling adults, who are likely to encounter more environmental hazards and who are more mobile. Gleeson et al. ’s randomised control trial7 with 120 community-dwelling people aged ≥ 50 years investigated the impact of 12 weeks of Alexander Technique lessons on balance and mobility with VIs compared with usual care. There was no impact on the primary outcome: the Short Physical Performance Battery score. However, there were greater intervention effects in the subgroup of multiple fallers, who are at an increased risk of injury compared with non-multiple fallers. Although the study was not powered for falls within the intervention group, trends towards a lower rate of falls and injurious falls were noted. 7 The Alexander Technique uses verbal feedback and manual guidance to teach awareness of previously unnoticed tension and, the authors suggest, may be a suitable intervention for people with VI as it does not require vision or the performance of regular exercises.

Falls Management Exercise and the VIOLET study

With an ageing population, the number of VIOP is predicted to rise. 54 Increasing incidences of key underlying causes such as macular degeneration and diabetes suggest that, without intervention, the number of people with VI in the UK could dramatically increase over the next 25 years. 54 This has implications for the need for falls prevention programmes that are tailored to VIOP. As outlined in The literature search and Falls and visual impairment, falls and FoF in later life are common, falls can be recurring and costs of care for those who fall are a disproportionate drain on local inpatient and adult social care budgets. 55,56 In the UK, falls are a major cause of disability;57 in Scotland, in 2010, three-quarters of people registered as blind or partially sighted (25,609; 74.2%) were aged ≥ 65 years and about one-third (11,158; 32.3%) had additional disabilities. Of the latter, over one-fifth (2476; 22.2%) were also deaf. 58

Our feasibility study draws on the learning from a number of previous studies, including the Falls Management Exercise (FaME)59 programme, the VIP trial49 and the recently completed VIP2UK pilot. 11 A recent Royal College of Physicians report60 demonstrated that 54% of falls exercise services have trained Postural Stability Instructors (PSIs) delivering the FaME programme in groups. A 6-month programme of FaME exercises has also been shown to significantly increase habitual physical activity (i.e. by 15 minutes of moderate to vigorous activity per day) in older people recruited through primary care even at 12 months post intervention, as well as reducing falls. 61 The FaME programme also has significant reach across the UK with a pool of qualified instructors.

A systematic review of falls prevention exercise62 suggested that at least a total of 40 hours of challenging balance and strength work is necessary to reduce falls. In the current study, 12 weeks (with a weekly group session and twice-weekly 1-hour home exercises) equates to 36 hours. We have therefore taken a pragmatic approach to what is achievable when considering roll-outs through NHS- or council-led services using the FaME approach. Indeed, both a scoping report for the Service Development Office27 and the Royal College of Physicians Audit60 suggested that, on average, falls services across the UK should deliver groups or sessions once a week for 8–12 weeks.

There are potential risks and benefits to introducing exercises to this population. As noted in Falls and visual impairment, FoF in VIOP may exacerbate existing gait and balance difficulties, further increasing the risk of falls. Recent research has identified that older people are at increased risk of falling following intensive endurance exercise bouts63 owing to exercise-induced alterations in respiration and muscular fatigue. It was also noted that VI could increase exercise-induced changes to postural control and consequently the risk of falling.

The feasibility study will provide an opportunity for VIOP to contribute to the adaptation and design of an acceptable community-based group exercise programme that could be incorporated into a general health promotion exercise programme. This study will enable VIOP to collaborate with researchers and instructors, and to draw on their expertise and experience to adapt a commonly adopted exercise programme (that is known to reduce falls risk in high-risk frequent fallers59 and low-risk older adults61) to their specific VI needs. This in turn may have a positive impact on gait and balance, increase confidence and lessen FoF. This study will add to an emerging body of work that is using FoF, as assessed by a widely validated cross-cultural tool [Falls Efficacy Scale – International (FES-I)], to address the gap in knowledge of how to successfully manage FoF.

Research aims

The rationale for the study is to provide an opportunity for VIOP to contribute to the adaptation of a group-based falls prevention programme that is prevalent in falls services across the UK. This should facilitate an acceptable, feasible and appealing intervention that will improve uptake and adherence to an intervention known to be effective. The aim of the study is to conduct a mixed-methods feasibility pilot study to inform the design and conduct of a future definitive multicentre RCT and economic evaluation of an adapted group-based exercise programme to prevent falls and reduce FoF among VIOP.

The research questions were:

1. Can an existing exercise programme be adapted for VIOP and successfully delivered in the community?

2. Is it feasible to conduct a RCT of this adapted exercise intervention and what would be the features of a future definitive trial?

Study objectives

The specific objectives were to:

1. explore participants’ ability to act as lay partners in a study to develop a condition-appropriate intervention and what methods enabled them to contribute to research as active partners

2. assess recruitment of VIOP and their willingness to be randomised

3. choose candidate outcome measures for a future RCT from FoF (confidence), activity avoidance, well-being/QoL, anxiety, depression, loneliness and number of falls

4. test the trial methodology and provide outcome data to inform sample size calculations for a definitive RCT

5. explore the capacity to deliver the adapted exercise programme

6. examine delivery of (fidelity) and compliance with the exercise intervention

7. explore participants’ reasons for participating in the exercise programme and acceptability of the exercise programme and trial procedures

8. develop a manualised intervention protocol and training package for a definitive RCT

9. assess the feasibility of collecting service use data for an economic evaluation of the intervention in a future RCT.

Prior to the completion of the feasibility study, the plan was to use the findings of this work to seek funding to conduct a definitive multicentre RCT to evaluate the effectiveness and cost-effectiveness of an adapted group-based exercise programme to prevent falls in VIOP by decreasing their FoF.

Introduction

A number of issues have been identified in the literature (see Chapter 1) regarding risk of falling and FoF in those both with and without VI. Despite evidence of the effectiveness of exercise programmes, poor adherence and compliance remain problems and have been found to reduce the effectiveness of any exercise programme. In a RCT carried out in New Zealand with VIOP,49 home safety modifications delivered by an OT resulted in a significant reduction in risk of falling, but the group receiving a home exercise programme (which had already been shown to be highly effective in reducing falls in older people without VI) showed a non-significant increase in risk of falls. On a planned subanalysis, the authors found that adherence to the home exercise component was poor and that those who did adhere to the home exercise programme did see a reduction in falls. 49 A UK study has sought the views of VIOP to help increase adherence to a home exercise programme. 28

In another, smaller, exercise study in people with VI aged ≥ 50 years,7 there was a trend for fewer falls overall, but the rate of falls was not significantly different between those who received 12 lessons on the Alexander Technique and those who did not. Levels of adherence to this programme were high as an instructor visited participants weekly at home. A recent systematic review of exercise in the prevention of falls in VIOP has highlighted the lack of studies in community-dwelling older people, particularly group-based as opposed to home-based exercise. 64

In the UK, two main programmes of exercise are offered as part of falls prevention services: the Otago home exercise programme49,65 and the FaME group exercise programme delivered by PSIs. 59 A Royal College of Physicians report66 showed that 54% of falls exercise services in the UK have trained PSIs delivering the FaME programme. FaME exercise classes are balance-specific, individually tailored and include backward chaining, functional floor exercises and targeted training for dynamic balance, strength, endurance, flexibility, functional skills, gait and ‘righting’ or ‘correcting’ skills to avoid a fall. A full description of the exercise programming and progression has been published. 67

Given the prevalence of both falls and FoF in VIOP, there has been relatively little research into adapting existing falls prevention interventions for this population. This formed the aim of this phase of the VIOLET study.

Aim

The aim of this phase was to recruit and run VIOP stakeholder groups to act as consultants on adapting an existing evidence-based falls prevention programme (FaME) for this population. The findings from this phase were to inform the adaptation of the FaME programme for the subsequent feasibility study.

This chapter reports on the findings of four stakeholder focus groups with VIOP. It explores their main concerns surrounding falls and details their recommendations on adapting a falls prevention programme.

Method

Stakeholder panels were conducted in the two geographical locations of the intervention sites: Newcastle upon Tyne and Glasgow. They took the form of focus groups facilitated by research staff. Two focus groups were completed in each area: the first covered views on falls and falls prevention; the second covered proposed tools to measure the outcomes.

Focus groups were used in this study to explore experiences, views and concerns, to clarify why particular views were held and to facilitate discussion among focus group participants. 68 Ethics approval was gained from both universities’ ethics committees. All participants gave recorded verbal consent.

Results

Across both areas, a total of 14 VIOP took part in the four focus groups. The maximum attendance was nine, the minimum five. All but one person at each site attended both focus groups. Participants had a range of VIs including macular degeneration, VIs following stroke and complete blindness. Some had been visually impaired for their whole life and some had started losing their sight more recently. Two people had guide dogs.

Adaptations made to the feasibility randomised controlled trial design

All stakeholder requests were incorporated into the planning and delivery of the classes with the exception of smaller class size and a choice of classes each week. Owing to study resource constraints, class sizes were planned at 10 and offered only once per week. However, this feedback will inform future interventions.

The request for instructor VI training was met. This training was expertly managed by Visibility in Glasgow and incorporated the use of blindfolds and adapted spectacles. All actors from both locations were included. Both exercise instructors and the research team simulated being both a person with VI and a guide and learned invaluable communication and guiding techniques. This training also covered the need for weekly discussions with participants about any changing needs regarding support, aids, position in the class (lighting/hearing) and the importance of social contact and peer support within the group.

An induction session, for introducing the venue and getting to know everyone, formed part of the first session. Participants were encouraged to bring along any carer/family member to the class and extra support personnel were available to help individuals who needed more personalised instruction in the group. The proposed home exercise content was also modified in accordance with the stakeholder panel’s overall feedback. This included adapting the 60-minute FaME home exercise element to graded exercise components of 10 minutes, 15 minutes and 20 minutes. The home exercise booklet used by PSIs71 was adapted to include a series of prompts, particularly focusing on incorporating some of the exercises into everyday activities, and split into smaller ‘chunks’ of exercise that lasted approximately 10 minutes and progressed in duration and intensity/difficulty over time (to a maximum 20-minute session). Audio clips were made available for those who wanted their home exercise programme to be verbal. These could be played on either a compact disc (CD) player or an MP3 player. A digital versatile disc (DVD) was also available to use.

In relation to VIOP’s views on completing the related assessment tools, participants were asked about their preferred mode of assessment tool delivery, such as large print or being read aloud. The preferred solution to the fall diary was implemented in the form of a weekly telephone call between a member of the research team and the participant. This protocol captured falls and ‘near misses’ and allowed a reporting of the financial costs of any such event.

Probably the single biggest impact on the conduct of the feasibility RCT has been a shift in emphasis to interpersonal communication. In trials, the mediating agent between triallists and participants is often the written word. In the case of the VIOLET study we endeavoured to make both the initial and ongoing collection of data based on interpersonal communication. Although this is more burdensome for the research team, we believe it gave the feasibility study a much higher chance of maximising retention and also provided a means of monitoring the impact of the intervention on participants. It also allowed us to address some of the issues that would otherwise be hard to address via written communication. For instance, given the feedback from the stakeholder groups that some of the QoL questions might be upsetting, we were now in a position to highlight and discuss this, both before and after administration of the outcome measures.

Limitations

The participants lived in cities and so would not necessarily represent the experiences of people living in rural areas, who might have different experiences of the accessibility of group classes or external environmental concerns (street furniture, lighting). Participants were recruited from low-vision charities and so were already engaged with external organisations and were less likely to be isolated. Although there was a range of ages, a mix of males and females, people with a history of falls and without, and a range of VIs and histories of sight loss, there was an absence of ethnic minorities and only one person had dual sensory loss. This limits the findings.

One of the focus group facilitators (DS) was the author of the FaME programme, delivers training to specialist exercise instructors and has extensive experience in presenting the exercise programme in a motivating manner. This may have influenced the participants’ enthusiasm for the programme and willingness to try floor-based exercises.

Summary of recommendations from the stakeholder group

These recommendations related to the content of the FaME programme, the way in which it was delivered, the logistics of attending the programme and the design of the subsequent feasibility study. The main recommendation was that personal choice and individual adaptation should be maximised in both the conduct and measurement of the intervention. Specific recommendations centred on venue familiarity, group dynamics and training of facilitators on VI. Additional recommendations were made regarding individual goal setting and choice regarding the methods by which study materials were received and administered.

It was the adaptation of the environment to the needs of a visually impaired person rather than the adaptation of the actual exercises that received most attention from the stakeholder group. Each point was considered in the design and delivery of the subsequent intervention. Recommendations were as follows.

Recommendations for adapting FaME:

• a fully accessible, familiar venue or the option to get to know an unfamiliar venue before sessions

• aid getting to and from the venue (e.g. by taxi)

• a maximum group size of eight persons

• no music

• individual adaptation regarding lighting and glare

• the option to bring another person (or guide dog)

• a social element (e.g. tea and a chat) after exercise sessions

• run more than one class (flexibility regarding class time) at the beginning for group bonding

• home exercises – reduce the length of home exercise sessions to 10–20 minutes; provide a variety of exercises; provide prompts (e.g. in large print, on DVD or audio); integrate exercises with daily living activities

• train PSIs on (1) the impact of VI, (2) adaptations (aids) that can be used for specific VIs and (3) communication (verbal clarity)

• more than one instructor or additional person to help VIOP

• provide the PSI with information regarding individual participants and time to fully understand the nature of the VI, how it may affect the individual and how it may vary across timeframes

• tailor content to the individual (e.g. floor work).

Recommendations for outcome measures and trial procedures:

• the possibility to choose how outcome measures are received and administered (e-mail, post, verbal)

• reduce the content of the falls dairy

• remove the resources/expenses form so that the researcher administers this

• incorporate a weekly telephone call to capture data concerning the nature of slips/trips/falls and the impact of these (to ensure capture of ‘near misses’)

• define ‘slip’, ‘trip’ and ‘fall’

• patient-centred outcome measures (personal goals)

• use a tool to capture the impact of VI on an individual and their activities.

Conclusion

Participants expressed very individual preferences and needs regarding the attendance of group exercise to prevent falls. They voiced a strong desire for instructors to be specifically trained in how to tailor instruction and delivery to each individual’s very different functional or visual limitations, which may change week to week. There was an equally strong preference for alternatives and choice in methods of engaging with home exercise and completion of outcome measures. These stakeholders gave clear and useful guidance on ways to adapt an existing falls prevention exercise intervention to maximise both uptake and adherence for VIOP. No matter how good an intervention is, if participants cannot or do not want to engage with it, it is effectively useless. Actively involving stakeholders in the modification of this intervention maximised the possibility that the intervention would be useful for the VIOP population.

Visually impaired older people stakeholder involvement in the adaptation of the FaME programme produced the first health promotion and falls prevention exercise intervention protocol developed collaboratively with older people for community-living VIOP. 46 As detailed in Appendix 1, the intervention protocol manual will enable accurate replication of the feasibility study intervention in a larger multisite study in the future, and can be used to ensure that adaptations of the original FaME intervention for VIOP can be replicated in practice in falls prevention programmes.

Introduction

This chapter describes the methods employed for the research process and conduct of the feasibility study, including identification, screening, consent and recruitment strategy, inclusion/exclusion criteria, flow of participants through the study, design and delivery of the intervention and the outcome measures used for data collection.

The overall aim of this phase was to conduct a feasibility study in order to inform the design and conduct of a future definitive multicentre RCT of an adapted community-based group exercise programme to prevent falls and reduce FoF in VIOP. This was in order to facilitate the development of a feasible and acceptable intervention to improve uptake of, compliance with and adherence to an intervention known to be effective.

The research questions were:

1. Can an existing exercise programme be adapted for VIOP and delivered successfully in the community?

2. Is it feasible to conduct a RCT of this adapted exercise intervention and what are the features of a future definitive trial?

The objectives are presented in Chapter 1.

Study design

This study adapted an existing, effective group-based health promotion intervention for older people (FaME) to reduce falls, with VIOP as experts. Because of the lack of relevant information for a full RCT on this topic, a randomised mixed-methods feasibility study was designed to inform the design and conduct of a future definitive multicentre RCT on an adapted version of FaME. The feasibility study addressed issues relating to the generalisability and transferability of the proposed intervention and investigated the viability of a future definitive RCT.

This was a two-centre randomised pilot trial of an adapted exercise programme for VIOP versus no intervention with embedded qualitative evaluation. An economic evaluation was also carried out (see Chapter 6). Older people with VI were recruited from two study sites (Newcastle and Glasgow) and randomised into one of two groups:

• group 1 – 12-week exercise programme (1-hour session per week)

• group 2 – usual activities.

Assessment of the feasibility to progress to a full trial

The criteria to judge the feasibility of progressing to a full trial at 6-month follow-up were as follows:

1. Fifty per cent or more of VIOP eligible for the study were willing to be recruited into the feasibility study.

2. Seventy per cent or more of the participants in the intervention arm have completed 9 to 12 group sessions in the exercise programme (compliance).

3. Data on key outcomes was collected at 6-month follow-up for ≥ 70% of those recruited.

4. Less than 10% of serious adverse events (SAEs) were deemed to be caused by the intervention.

Study population

The study population were drawn from community-living VIOP in Newcastle and Glasgow.

Ethics and regulatory issues

The conduct of this study was in accordance with the recommendations for physicians involved in research on human subjects adopted by the 18th World Medical Assembly, Helsinki, 1964, and later revisions.

Ethics approval from the Newcastle and North Tyneside Research Ethics Committee (REC) was sought prior to the commencement of the intervention, as was research and development approval. The REC reference was 15/NE/0057 and the UNN reference was RE-HLS-13-140707-53bb0a7806e37.

Local approvals were sought before recruitment commenced at each site. Newcastle Clinical Trials Unit received a written copy of local approval documentation before initiating each centre and accepting participants into the study. The study was registered as Current Controlled Trials ISRCTN16949845.

Information sheets were provided to all eligible subjects and written informed consent obtained prior to any study procedures. If subjects were unable to sign the consent form, verbal consent was witnessed by a third party, who signed the witness section of the consent form.

Confidentiality

Personal data were regarded as strictly confidential. To preserve anonymity, any data leaving the site identified participants by their initials and a unique study identification code only. The study complied with the Data Protection Act 1998. 72 All study records and investigator site files were kept in offices with restricted access.

Identification, screening, recruitment and consent of participants

The Ophthalmology Department at the Royal Victoria Infirmary (RVI) in Newcastle sees an average of 3138 patients per month aged ≥ 65 years, with approximately 45 patients per month aged ≥ 65 years attending the low-vision clinic. NSBP has approximately 1500 members. NHS Greater Glasgow & Clyde saw approximately 500 patients per month in the general eye clinic, with 20–25 VIOP attending the Glasgow Caledonian University low-vision clinic per month. Visibility ran both a patient support service and an information line and across both they had contact with around 1500 people per year. These organisations identified potential participants and passed on expressions of interest to the research team.

It is known from previous studies that recruitment and adherence in frail older people can be difficult, although data for recruitment and retention rates in VIOP were relatively unknown. In the VIP2UK study,11 only 51% of eligible participants and 10% of those who were originally screened agreed to take part. However, the inclusion/exclusion criteria meant that the study excluded participants who would have been eligible for the VIOLET study.

Visually impaired older people who expressed an interest in participating in the VIOLET study received further information. This information was available in large-print and audio formats. These materials had been designed in consultation with the stakeholder panel and checked to ensure that they met RNIB standards in terms of suitability for those with VI.

The researcher then invited eligible VIOP to participate and answered any questions raised. Signed consent to participate in the feasibility study was sought only when the researcher had ensured that the participant had accessed and understood the information. Once consent had been gained, a letter was sent to the participant’s GP explaining the inclusion and exclusion criteria for the study. This allowed the GP to make a judgement regarding the medical fitness of the participant to take part in the study based on these criteria. Once a participant’s GP confirmation of eligibility had been received, the participant was randomised.

Identification of potential participants

The primary recruiting source at the Newcastle site was NSBP. The organisation raised awareness of the study in its newsletter, which is sent to all members. In addition, members of the research team visited member meetings to discuss the study and how to receive further information. When recruitment began, a dedicated member of NSBP rang the members aged ≥ 60 years to discuss the study and gain permission to forward their contact details to the research team.

Identification of participants from the RVI low-vision clinic proved problematic. The physical space of the clinic area was unsuitable for the research team to approach potential participants. The responsibility for raising awareness of the study was then transferred wholly to the medical and paramedical staff in both the eye clinics and low-vision clinics. Although the managers of these departments and the lead consultant were aware of the study, very few of the staff referred potential participants to the researchers. In an effort to increase recruitment from the NHS site, a minor amendment was made to allow Eye Clinic Liaison Officers (ECLOs) to approach potential participants. They were able to identify potential participants who had recently been seen in their service and, with their permission, forwarded expressions of interest to the research team. These potential participants were then contacted by the research team and assessed for eligibility in the same manner as participants identified through the low-vision clinic and NSBP.

The primary recruiting source at the Glasgow site was Visibility. It raised awareness of the study in its newsletter to members, with expressions of interest being forwarded to the study team. As recruitment via this mode was felt to be slow, Visibility contacted potential participants directly. Expressions of interest were again forwarded to the research staff.

It had been planned to recruit participants from the low-vision clinics held in Glasgow Caledonian University. Unfortunately, recruitment coincided with the summer period when the clinics in the academic setting were not running. Therefore, no recruitment took place from this source.

Screening procedures

Visually impaired older people who expressed a continued interest in participating in the study were screened for eligibility by a researcher over the telephone or at a site mutually convenient to the researcher and the potential participant. Screening was based on the inclusion/exclusion criteria. Eligibility was confirmed by the GP.

To facilitate the first running of the first cohorts in both Newcastle and Glasgow, a staggered start was employed at both sites. This enabled those for whom GP confirmation of eligibility had not been returned promptly to start at any time within the first 3 weeks and continue to complete the 12 sessions. More than 12 weeks of sessions were offered so that each participant was able to attend 12 sessions even with a delayed start.

Consent procedures

Once screening for eligibility had been completed, the researcher invited the VIOP to participate and answered any questions they raised. Following receipt of information about the study, participants were given reasonable time (a minimum of 24 hours) to decide whether or not they wanted to participate. As previously stated, signed or recorded verbal consent to participate in the feasibility study (as appropriate for each VIOP participant) was sought only when the researcher had ensured that the participant had accessed and understood the information provided. If the VIOP was unable to sign the consent form, verbal consent was witnessed by a third party, who signed the witness section of the consent form.

The original signed consent form was retained in the investigator site file, a copy added to the clinical notes (for participants recruited from the NHS) and a copy provided to the participant. The participant specifically consented to their GP being contacted regarding their medical fitness and being informed of their participation in the study.

The right to refuse to participate without giving reasons was respected.

The information sheet and consent form for the study was available in English only. It was decided to exclude participants who were unable to understand simple instructions in English as they would not be able to understand the movement instructions given during the exercise sessions if randomised to the intervention arm of the study.

Randomisation

A computer-based system allocated participants to the intervention and control groups in a 1 : 1 ratio using permuted random blocks of variable length and stratification by centre. This was administered centrally via Newcastle Clinical Trials Unit using a secure web-based system. The principal investigator (PI) at the site, or an individual with delegate authority, accessed the web-based system. Random allocation was determined after a participant screening identification number and initials were entered into the web-based system; participants could then be informed.

Study intervention

The proposed intervention was adapted from the group-based FaME programme, which is widely used as a health-promoting activity in the UK with community-dwelling older people.

For the intervention arm, local qualified FaME PSIs working on each site were trained by a research team member (DS) on adapting the exercises for people with VI following recommendations from the stakeholder panels. The PSIs also received VI training from Visibility.

The exercise programme (the intervention) ran weekly over 12 weeks, with each session lasting up to 1 hour. Two exercise groups were held at each site. In order to maximise recruitment, a further exercise group was held at the Newcastle site. All groups were held in a community venue. In Glasgow, the exercise groups were run at Visibility because this venue was known to the participants. In Newcastle, classes were run at the Trinity Centre. These were city centre venues and required participants to travel into the city centre.

None of the groups had more than 10 participants. The exercises consisted of balance-specific, individually tailored and targeted training for dynamic balance, strength, endurance, flexibility, gait and functional skills and ‘righting’ or ‘correcting’ skills to avoid a fall, as well as backward chaining, that is, retraining of the ability to get down to and up from the floor. The exercises also included functional floor exercises and adapted tai chi exercises. The intervention increased in difficulty over the 12 weeks. Resistance bands and mats were used throughout the study. The details of the methods of delivery, group size, support and timing were decided by the stakeholder panels. A full description of the original exercise programming and progression has been published and an adapted manualised intervention is included as Appendix 1, as this was an important objective of the study.

Prior to commencing the exercise sessions, all participants completed a health screening tool that is normally administered by the exercise instructors prior to delivery of the FaME programme.

The first session was used as an induction to familiarise VIOP with the venue and instructor. Any time remaining after this induction was used for a short exercise session. All following sessions lasted 1 hour and comprised:

• a 10- to 15-minute warm-up focusing on circulation exercises

• targeted full-range joint mobilisation and light stretches

• a standing cardiovascular endurance section lasting 5 minutes in week 2 and increasing to 10 minutes by week 12

• a balance retraining section of 10 minutes starting with static exercises (such as one-leg stands) and progressing to highly challenging dynamic exercises (such as heel walking and backwards walking)

• a 10-minute seated resistance component using progressively more challenging resistance bands

• a 10-minute cool-down including developmental stretches for both the upper and lower limb muscles and one form of adapted tai chi that progressed in difficulty.

The introduction of the backward chaining approach for getting down to and up from the floor was not planned for a specific week of the intervention, given that the VIOP stakeholders had expressed some concern about the ability of the target population to perform this safely. The PSIs were advised to introduce the floor-based exercises at a stage appropriate to the group and no sooner than week 4.

Participants were also advised to exercise at home for up to 2 hours per week using an adapted standardised home exercise programme that has been developed by DS and used in other related studies. Exercises were provided in large-text or audio format and consisted of specific chair and standing exercises to improve balance, bone and muscle strength and flexibility, similar to the exercises carried out in the exercise classes. The home exercises were designed to be completed in 10- to 20-minute blocks, commencing with 10-minute blocks in the first 4 weeks, increasing to 15-minute blocks for a further 4 weeks and increasing again to 20-minute blocks in the last 4 weeks. A greater number of exercises and more challenging balance exercises were included as the duration of the sessions gradually increased. This progressive approach was implemented following VIOP stakeholder advice that 30-minute blocks of home exercise would not be realistic. The exercises were to be performed 6 days per week – that is, every day excluding the day of the exercise class. All home programmes contained ‘prompts’ that linked exercises to daily tasks (e.g. performing heel raises while waiting for the kettle to boil). The prompts were included to increase compliance, following VIOP stakeholder advice. The PSIs supplied participants with the home exercise programmes and were responsible for encouraging participants to adhere to and progress through the programmes. PSIs were asked to ensure that the content of the home exercise programmes had been taught in the group sessions and to help participants with the specific exercise techniques to reduce the risk of any adverse events (AEs) from the home exercises.

Participants were provided with transport by taxi to and from the classes, if required, and invited to bring along someone for support if they wished.

The description of this adapted exercise programme is provided in the manualised intervention, which is available on the National Institute for Health Research (NIHR) website.

Those participants who were randomised to the usual activities group received no intervention. However, they were offered an equivalent exercise programme after the week 24 follow-up data collection at both sites.

Fidelity of the intervention

Postural Stability Instructors were given a flexible framework of lesson plans for the 12-week programme prior to the start of the intervention.

A standardised quality assurance checklist was used to ensure effective and safe delivery of the programme, including progression, adherence to the protocol (i.e. providing home exercise programme, checking on adherence to home exercise and recording any falls mentioned) and completion of all paperwork.

The PSIs compliance with the course content (fidelity) was assessed by videoing two sessions at each site at weeks 3 and/or 4 and weeks 9 and/or 10 respectively.

These videos were examined by a research team member to check that the main components of fitness were covered and progressed sufficiently. Their delivery was signed off by checking against the videos provided.

The quality assurance checklist can be found appended to the manual of the adapted FaME intervention (a copy of this is available from the VIOLET project page on the NIHR website).

Qualitative interviews

As this was a feasibility study, qualitative interviews were conducted to explore acceptability and applicability of the intervention, the research methods and the outcome measures. These interviews are reported in Chapter 5. Structured interviews were also conducted with a purposive sample of the exercise practitioners at two points (before training and at the end of the intervention) to explore their (changing) perspectives on the provision of the intervention over the duration of the intervention (see Chapter 5).

Outcome measures and assessment

A key aim of the feasibility study was to determine whether or not the primary and secondary outcomes for a proposed full trial could be measured for all participants. The feasibility study was not powered to detect significant differences in these measures but was able to describe observed changes between time points and their direction.

The selected candidate outcome measures were standardised assessment instruments that have been used previously in comparable studies. These measures were selected in consultation with the stakeholder panels. The main reasons for this selection were the relevance of the questions to the participants and that the questionnaires were short and, therefore, easier for a person with VI to manage. The proposed measurement tools were reviewed by the stakeholder panels at both sites to ensure that the most appropriate and most acceptable outcome measurement tools were selected and used in the pilot trial. FoF was selected as the primary outcome because, although FoF is common, debilitating and a significant predictor of a future fall particularly among older people, successful management of FoF is limited. Recent reviews show that there is a lack of high-quality research that has FoF as a primary health outcome. There are also limited health economic data about FoF interventions.

Measuring the number of falls as a primary outcome does not take into account the more complex impact of an intervention. Reducing FoF may increase participants’ confidence in their ability to walk safely and continue with everyday activities. Assessing FoF before and after the proposed intervention captures participants’ self-reported change in confidence to continue with physical and social activities. This provides an indication of their perceived QoL and whether or not the negative impacts of FoF, such as social isolation and risk of further falls, have been reduced.

Assessment/data collection

Following consent, a researcher assessed all participants, either on site or in their own home, depending on the participants’ preferences. The participants were also asked a number of questions regarding demographic data, comorbidities, current medication and socioeconomic information and to complete the baseline outcome questionnaires discussed in this chapter.

Because of their VI, assistance was required to fill in the questionnaires, either by the researcher reading out the questions and filling out the forms or by providing the questionnaires in a format accessible to the participant. Sufficient time was allowed for this stage to accommodate each individual’s preference. Once confirmation of eligibility had been received from a participant’s GP, randomisation took place. For both groups of participants, baseline data were collected no more than 4 weeks prior to the intervention start date.

For the intervention group, it was planned that the week 12 visit was conducted ± 2 weeks from the end of the intervention date unless there were prior commitments or extenuating circumstances. For the usual activity group, the week 12 visit was planned to take place 12 ± 2 weeks from the date of their baseline visit. The week 24 visit for all participants was planned to take place 12 ± 2 weeks from the date of their week 12 visit.

The questionnaires were repeated at the week 12 and week 24 visits. All participants were also telephoned each week from their baseline visit up to their week 24 follow-up visit to record any AEs (including falls) that had occurred. A schematic diagram of the intervention is provided in Figure 1. Table 1 outlines the schedule of events.

FIGURE 1.

Schematic diagram of the RCT. Adapted with permission from Adams et al. 81 This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

Adapted with permission from Adams et al. 81 This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

Assessment of compliance

Participants’ compliance with the exercise programme was assessed by reviewing attendance records kept by the PSI and the self-reported compliance to the home exercises at the end of the 12-week course. In addition, qualitative interviews, conducted by the researchers after the completion of the exercise classes, provided information on why VIOP attended or did not attend exercise classes and what barriers prevented them from attending. Eligible VIOP who declined the invitation to participate were asked the reason for declining.

Instructors were required to submit basic lesson plans for the 12-week programme prior to the start of the intervention. The PSIs’ compliance with the course content (fidelity) was assessed by a researcher attending a sample of exercise sessions. A standardised checklist was used, similar to that used by researchers in a previously published trial, and these sessions videotaped for quality assurance purposes. Following completion of the intervention, PSIs were also invited to discuss their experiences of delivering the adapted intervention (these discussions are detailed in Chapter 5).

Withdrawal of participants

Participants had the right to withdraw from the study at any time for any reason and without giving a reason. The investigator had the right to withdraw participants from the study in the event of intercurrent illness, AEs, SAEs, protocol violations or if judged to be in the participant’s best interests or for administrative or other reasons. Participants were also withdrawn if the participant’s GP felt they were medically unfit to take part or the participant failed the pre-exercise screening carried out by the exercise instructors.

There were two withdrawal options:

1. withdrawing completely (i.e. withdrawal from both the study treatment and provision of follow-up data)

2. withdrawing partially (i.e. withdrawal from study treatment, including a request to move to another treatment arm, but continuing to provide follow-up data by completing questionnaires).

Consent was sought from participants choosing option 1 to retain data collected up to the point of withdrawal. Participants were asked if they would be happy for the reason for the decision to withdraw to be recorded.

Assessment of harms/adverse event reporting

Participants were telephoned each week from consent to record any AEs that occurred up to their week 24 follow-up. All AEs were assessed for seriousness, severity, duration, causality and expectedness. This included any SAEs reported by the instructors within the exercise sessions or reported in the falls diaries given to participants (if they requested to complete these themselves).

Adverse events were recorded by treatment arm. In addition to splitting up AEs by treatment arm, we made a further distinction for those in the intervention arm by categorising them as occurring (1) from baseline up to the first exercise class of the intervention or (2) after the start of the intervention (following the first exercise class). This was to ensure that the AEs in the two trial arms could be compared over the same period of time. All participants were telephoned weekly from consent because it was not known to which arm they would be assigned until their confirmation of GP eligibility form was completed and returned and the participant could be randomised. However, this led to a potential discrepancy in time periods: for those randomised to the usual activity arm it could be ensured that there were 12 weekly telephone calls between their baseline and week 12 visits; however, for those in the intervention arm there could be a delay between their baseline visit and the start of the intervention, resulting in over 12 weekly telephone calls between their baseline and week 12 visit.

All AEs were treated appropriately. Treatment included one or more of the following: no action taken (i.e. further observation only), concomitant medication given, non-drug therapy given, or patient hospitalised/patient’s hospitalisation prolonged.

The trial oversight committee (TOC) monitored the SAEs to bring the intervention to a close should this have been deemed necessary.

Data handling and record keeping

Collected data were entered on a secure validated clinical data management system (MACRO 4.6; Rackspace Inc., Windcrest, TX, USA). Data were handled, computerised and stored in accordance with the Data Protection Act 1998. 72

All study data were retained in accordance with the latest Good Clinical Practice Directive (2005/28/EC)82 and local policy.

Statistical analysis

As this was a feasibility trial, the main analyses were descriptive in order to inform the design of a future definitive study. The main outcomes were feasibility outcomes. The numbers of eligible participants seen over the recruitment period and the resulting rates of recruitment, compliance with randomisation and data completion were presented. Data completeness of the instruments and any potential bias in the completion of follow-up data to inform the choice of instruments in a future trial were ascertained. In Chapter 4, the majority of the outcome data have been presented in simple descriptive tables presenting percentages, means and standard deviations (SDs) or five-number summaries (as appropriate) for each arm of the study. This information was used to inform the design, choice of primary outcome, necessary sample size and approach to the analysis of a future definitive trial. See Chapter 4 for a detailed presentation of findings.

There was potential for clustering effects, particularly class-based clustering in the intervention group. It is part of the function of a pilot trial to investigate and estimate the size of any such effects. As this was a feasibility study, it was hoped that the results would provide estimates of variability in key outcomes both within and between classes and trial arms, which would be used to design a future definitive trial. However, it was understood that there would be very limited information on the size of intraclass correlation coefficient (ICC) in a study with only two centres and two classes per centre and that any estimates of ICC would be very imprecise based on this pilot trial. Estimates of ICC may be obtained from larger relevant trials such as the current ProAct65 study. 61 However, since ProAct65 is a cluster randomised trial as opposed to one in which the clustering effects are due to treatment alone, their estimate of ICC may be larger than those appropriate for use in a future individually randomised trial. 83

Identification of potential participants

Recruiting for the Newcastle site, a dedicated member of NSBP rang members aged > 60 years to gain permission to forward their contact details to the research team. It is not known precisely how many potential participants were approached by NSBP but anecdotal information suggests that > 200 calls were made. From these, 35 expressions of interest were passed to the researchers. Identification of participants from the RVI low-vision clinic in Newcastle proved problematic: very few of the staff referred potential participants to the researchers. After the involvement of ECLOs, 15 potential participants were identified and their expressions of interest forwarded to the research team.

The primary source of potential participants at the Glasgow site was Visibility; no recruitment took place at the low-vision clinic (see Chapter 3 for further details). Anecdotally, > 100 direct contacts were made by Visibility staff. Expressions of interest were again forwarded to the research staff. It is not precisely known how many potential participants were contacted by Visibility; however, the research team received 48 expressions of interest.

All potential participants were contacted by the research team and assessed for eligibility in a consistent manner.

Participant flow

Figure 2 shows the screening, recruitment and follow-up of participants by trial researchers at the two centres. In total, 82 VIOP were screened for eligibility by the researchers. Of these, 16 were eventually found to be ineligible, two declined to participate, 66 consented and 64 were randomised after further loss through failure to meet eligibility criteria. The target for recruitment was 80, with a view to obtaining a minimum of 30 responses in each trial arm at the 6-month follow-up point.

FIGURE 2.

The VIOLET Consolidated Standards of Reporting Trials (CONSORT) flow diagram. Adapted with permission from Adams et al. 81 This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

Adapted with permission from Adams et al. 81 This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

The Newcastle site opened to recruitment in June 2015 whereas the Glasgow site opened in July 2015. The numbers recruited each month (combined across sites) are shown in Table 3.

The numbers screened, recruited and randomised by site are shown in Table 4.

After randomisation, 33 VIOP were allocated to the intervention arm and 31 to the usual activities arm. Of the 33 VIOP allocated to the intervention, three did not attend any classes but two of these nevertheless provided study data. During the study, one person was lost to follow-up and four people in the intervention arm withdrew completely from the study (Table 5). The remaining subjects provided data that was included in the statistical analysis.

Baseline participant characteristics

Demographic and baseline characteristics at randomisation were compared across treatment groups. Descriptive statistics are tabulated by treatment group and overall in Table 6. The distributions of demographic variables were comparable across the trial arms; the only noticeable difference was that more participants in the usual activities arm lived alone.

As would be expected in a group of older people, they reported a number of comorbidities. Summary statistics of the numbers of baseline self-reported comorbidities are shown in Table 7.

Delivery of intervention and intervention compliance

Two sets of 12-week exercise sessions were intended to be offered at both Newcastle and Glasgow, allowing two potential start dates as recruitment proceeded. However, a staggered start was used to facilitate the running of the first set of classes in both Newcastle and Glasgow. This enabled those for whom the GP confirmation of eligibility had not been returned promptly to start at any time within the first 3 weeks and continue to complete 12 sessions. A total of 14 weeks of sessions was offered so that each participant had the ability to attend 12 sessions even with a delayed start. In Newcastle, a third set of classes was provided for those who had been recruited later or whose eligibility checks took a long time to complete. In Glasgow, the class sizes were six and nine and, in Newcastle, class sizes were six, five and four.

Table 8 summarises the number of sessions attended by participants randomised to the intervention arm on the basis of class registers and how often they exercised at home (self-report) with FD. Four participants withdrew completely from the study and attended 0, 2, 3 and 11 classes prior to withdrawing (see Table 4). One further participant was withdrawn from the trial (deceased) after attending six exercise classes. Two additional participants randomised to the intervention arm did not attend any classes but continued to provide trial questionnaire data. One participant withdrew from the intervention after three classes but continued to provide trial data. In fact, 25 out of 33 participants (76%) attended nine or more classes, which was one of the feasibility criteria for a future trial. The intervention arm participants were encouraged to carry out exercises at home. It can be seen that they spent an average of 50 minutes per week doing home exercises, though there was a large variation in the amount.

Data completeness

The trial outcomes were collected at baseline and at weeks 12 and 24, with data collection planned to be within 2 weeks of these time points. Figures 3 and 4 show the time period between baseline and when each of these two assessments actually occurred. A total of 60 out of 64 participants (94%) provided data at the week 12 visit (four had withdrawn completely from the study at this point) and 59 out of 64 participants (92%) completed the week 24 visit (four had withdrawn completely from the study and one was lost to follow-up at this point). In two cases, assessments were completed outside the 2-week limit owing to other commitments or extenuating circumstances.

FIGURE 3.

Time between start of trial and week 12 assessment. Graphs by randomisation arm: (a) intervention arm and (b) usual activities arm. Starting point is baseline visit for usual activities arm and first exercise class for intervention arm (reference lines at ± 2 weeks).

FIGURE 4.

Time between start of trial and week 24 assessment. Graphs by randomisation arm: (a) intervention arm and (b) usual activities arm. Starting point is baseline visit for usual activities arm and first exercise class for intervention arm (reference lines at ± 2 weeks).

When considering the feasibility of a definitive trial it is important to look at how well alternative outcomes have been completed. Completion rates for each questionnaire at each time point are reported in Table 9 and completion rates for elements of the Phone-FITT questionnaire are reported in Table 10. It can be seen that all those participants in the intervention arm remaining in the trial at each time point completed each of the questionnaires. There were two occasions when participants in the usual activity arms only partially completed a questionnaire (these were nevertheless still usable following missing data rules) and two occasions when whole questionnaires were not completed. All elements of the Phone-FITT questionnaire were completed for all participants remaining in the trial at each time point. This suggests that all chosen scales were suitable for use in a future trial.

TABLE 9 - Completion of FES-I, WSAS, HADS, loneliness, TUG, and FRAT questionnaires by arm at baseline, week 12 and week 24

Contains missing items but remains usable with the application of particular rules for dealing with missing data.

Either completely missing or more than the acceptable number of missing items.

Questionnaire	Time point	Intervention arm (N = 33)			Usual activities arm (N = 31)
		Complete, n (%)	Partial, n (%)a	Missingb	Complete, n (%)	Partial, n (%)a	Missingb
Concerns over falling (FES-I)	Week 0	32 (97)	0 (0)	1 (3)	29 (94)	2 (6)	0 (0)
	Week 12	29 (88)	0 (0)	4 (12)	30 (97)	0 (0)	1 (3)
	Week 24	28 (85)	0 (0)	5 (15)	31 (100)	0 (0)	0 (0)
Impact of illness (WSAS)	Week 0	32 (97)	0 (0)	1 (3)	30 (97)	0 (0)	1 (3)
	Week 12	29 (88)	0 (0)	4 (12)	31 (100)	0 (0)	0 (0)
	Week 24	28 (85)	0 (0)	5 (15)	31 (100)	0 (0)	0 (0)
Anxiety and depression (HADS): anxiety component	Week 0	33 (100)	0 (0)	0 (0)	31 (100)	0 (0)	0 (0)
	Week 12	29 (88)	0 (0)	4 (12)	31 (100)	0 (0)	0 (0)
	Week 24	28 (85)	0 (0)	5 (15)	31 (100)	0 (0)	0 (0)
Anxiety and depression (HADS): depression component	Week 0	33 (100)	0 (0)	0 (0)	31 (100)	0 (0)	0 (0)
	Week 12	29 (88)	0 (0)	4 (12)	31 (100)	0 (0)	0 (0)
	Week 24	28 (85)	0 (0)	5 (15)	31 (100)	0 (0)	0 (0)
Loneliness: emotional loneliness	Week 0	33 (100)	0 (0)	0 (0)	31 (100)	0 (0)	0 (0)
	Week 12	29 (88)	0 (0)	4 (12)	31 (100)	0 (0)	0 (0)
	Week 24	28 (85)	0 (0)	5 (15)	30 (97)	0 (0)	1 (3)
Loneliness: social loneliness	Week 0	33 (100)	0 (0)	0 (0)	31 (100)	0 (0)	0 (0)
	Week 12	29 (88)	0 (0)	4 (12)	31 (100)	0 (0)	0 (0)
	Week 24	28 (85)	0 (0)	5 (15)	31 (100)	0 (0)	0 (0)
Functional test (TUG)	Week 0	33 (100)	0 (0)	0 (0)	31 (100)	0 (0)	0 (0)
	Week 12	28 (85)	0 (0)	5 (15)	31 (100)	0 (0)	0 (0)
	Week 24	28 (85)	0 (0)	5 (15)	31 (100)	0 (0)	0 (0)
FRAT	Week 0	33 (100)	0 (0)	0 (0)	31 (100)	0 (0)	0 (0)
	Week 12	29 (88)	0 (0)	4 (12)	31 (100)	0 (0)	0 (0)
	Week 24	28 (85)	0 (0)	5 (15)	31 (100)	0 (0)	0 (0)

TABLE 10 - Completion rate of Phone-FITT questionnaires by arm at baseline, week 12 and week 24

Contains missing items but remains usable with the application of particular rules for dealing with missing data.

Completely missing.

Questionnaire	Time point	Intervention arm (N = 33)			Usual activities arm (N = 31)
		Complete, n (%)	Partial, n (%)a	Missingb	Complete, n (%)	Partial, n (%)a	Missingb
Household FD	Week 0	33 (100)	0 (0)	0 (0)	31 (100)	0 (0)	0 (0)
	Week 12	29 (88)	0 (0)	4 (12)	31 (100)	0 (0)	0 (0)
	Week 24	28 (85)	0 (0)	5 (15)	31 (100)	0 (0)	0 (0)
Recreational FD	Week 0	33 (100)	0 (0)	0 (0)	31 (100)	0 (0)	0 (0)
	Week 12	29 (88)	0 (0)	4 (12)	31 (100)	0 (0)	0 (0)
	Week 24	28 (85)	0 (0)	5 (15)	31 (100)	0 (0)	0 (0)
Total FD	Week 0	33 (100)	0 (0)	0 (0)	31 (100)	0 (0)	0 (0)
	Week 12	29 (88)	0 (0)	4 (12)	31 (100)	0 (0)	0 (0)
	Week 24	28 (85)	0 (0)	5 (15)	31 (100)	0 (0)	0 (0)
Household FDI	Week 0	33 (100)	0 (0)	0 (0)	31 (100)	0 (0)	0 (0)
	Week 12	29 (88)	0 (0)	4 (12)	31 (100)	0 (0)	0 (0)
	Week 24	28 (85)	0 (0)	5 (15)	31 (100)	0 (0)	0 (0)
Recreational FDI	Week 0	33 (100)	0 (0)	0 (0)	31 (100)	0 (0)	0 (0)
	Week 12	29 (88)	0 (0)	4 (12)	31 (100)	0 (0)	0 (0)
	Week 24	28 (85)	0 (0)	5 (15)	31 (100)	0 (0)	0 (0)
Total FDI	Week 0	33 (100)	0 (0)	0 (0)	31 (100)	0 (0)	0 (0)
	Week 12	29 (88)	0 (0)	4 (12)	31 (100)	0 (0)	0 (0)
	Week 24	28 (85)	0 (0)	5 (15)	31 (100)	0 (0)	0 (0)

The number of weekly telephone calls is shown in Table 11. Trial participants should have been telephoned each week during the 24-week study period to collect information on falls, near misses and SAEs. If the participant was in the intervention arm, they were also asked about home exercises. However, some participants received more than 24 telephone calls because they were randomised to the intervention arm but had to wait a number of weeks before a programme of exercise classes became available. These participants were still telephoned each week to collect information before the first class and then for the 24 weeks thereafter. Sometimes the classes were offered over 14 weeks (meaning a total of 26 telephone calls up to the week 24 assessment) as extra classes were available to allow participants to catch up if they missed one or two classes.

TABLE 11 - Completion rate and trial period of weekly telephone calls by trial arm

LQ, lower quartile; UQ, upper quartile.

One participant attended two exercise classes, with five telephone calls occurring before the first class, and then withdrew.

			Number of weekly telephone calls
Arm	Period of trial	n	Minimum	LQ	Median	UQ	Maximum
Intervention arm	Baseline to first class	15	1	2	3	5	17
	First class to week 24 assessment	29a	3	22	24	24	26
	No classes attended	3	3	13.5	24	24	24
Usual activities arm	No classes attended	31	20	23	24	24	24

For participants randomised to the intervention arm, any calls that occurred before the start of the first exercise class are recorded in the ‘Baseline to first class’ row of Table 11. Any calls between the first class and week 24 (with an extra 2 weeks if classes had been offered over 14 weeks) are recorded in the ‘First class to week 24 assessment’ row. The third row of Table 11 shows summary statistics for participants randomised to the intervention arm who did not attend any classes (so there was no ‘start of intervention’). The final row shows summary statistics for the usual activities arm, which start from the baseline visit and should continue for 24 weeks.

Summary statistics on trial outcomes

Serious adverse events

Details of the 16 SAEs are shown in Table 21 and their period of occurrence in Table 22.

Falls without injury

Details of the 17 falls without injury are shown in Table 23 and their period of occurrence in Table 24.

Assessment of the feasibility to progress to a full trial

The agreed progression criteria to judge the feasibility of a definitive trial were:

1. Fifty per cent or more of VIOP eligible for the study are willing to be recruited into the feasibility study.

   All but two of those VIOP found to be eligible by the researchers after screening against the inclusion/exclusion criteria were willing to be recruited into the study (66/68 = 97%); although a further two participants were later found to be ineligible, this criterion is judged to have been met. The exact number of potential participants approached by third-sector organisations was not available, but it is known that the number screened by the researchers is much lower than the number of potential participants initially contacted. We therefore feel that meeting this criterion does not reflect the difficulties encountered in recruiting participants to the study.

2. Seventy per cent or more of participants in the intervention arm have completed nine to 12 group sessions in the exercise programme (compliance).

   Of the 33 participants randomised to the intervention arm in the trial, 25 (76%) attended between 9 and 12 exercise classes, so this criterion was met. A total of 10 participants (30%) attended all 12 classes, five (15%) attended 11 classes, five (15%) attended 10 classes and five (15%) attended nine classes. Of the eight participants not complying, three (9%) attended no classes. One of these withdrew completely from the trial; the remaining two did not. One participant (3%) attended two classes, two (6%) attended three classes before withdrawing from the intervention, one (3%) attended six classes before dying and one (3%) attended eight classes.

3. Data on key outcomes was collected at 6-month follow-up for ≥ 70% of those recruited.

   Overall, 92% of those recruited provided questionnaire data at the week 24 follow-up visit, so this criterion was met. The response rate at week 24 was 85% for all scales in the intervention arm; responses were missing for the five participants who withdrew completely from the study. The response rate at week 24 was 100% for all scales in the usual activities arm; the exception to this was the emotional loneliness questionnaire – one participant did not complete this questionnaire, so the response rate was 97% for this item.

4. Fewer than 10% of SAEs were deemed to be caused by the intervention.

   1. There were no SAEs deemed to be caused by the intervention, so this criterion is met.

Sample size for definitive trial

We had originally thought that the most appropriate primary outcome measure for a future trial would be FES-I, which measures concerns about falling. We experienced no difficulties in using this scale in the pilot study and so would recommend that it is used in future. Unfortunately, no work has been done to establish the minimum clinically important difference (MCID) for this scale, which is key to a calculation of required sample size. Therefore, indicative calculations have been done using standardised differences of 0.3, 0.4 and 0.5 SDs. A future trial of an exercise class intervention has the potential for the variation between those attending the same class to be smaller than that seen in patients in different classes or in the usual care group, where there is no group interaction. 83 The very small numbers in each class in this pilot study will not provide a reliable estimate of the ICC. Unfortunately, we were not able to obtain appropriate estimates of the ICC from other similar studies. Calculations have been done to show the sensitivity of the sample size estimates by assuming ICC values of 0.01, 0.03 and 0.05. When clustering in the exercise groups needs to be taken into account in the calculations, it is also important to specify the cluster size. After discussions with the stakeholder group, we planned to have exercise classes of no more than eight people, as they felt the visually impaired would need closer attention from the PSIs taking the classes. We found that there was a 15% dropout rate in the intervention group in this study by week 24, which would reduce the class size to seven by follow-up – this is therefore the number in the following calculations. Overall, the dropout rate was 8% across both arms; we have used this figure to work out how many participants need to be recruited to obtain the minimum numbers at follow-up. Note that when there is clustering in only one arm, it is more efficient to have unequal allocation to each arm, marginally increasing the sample size for clustering in the intervention arm. 83

The key parameters necessary to calculate the sample size of a future definitive trial are as follows:

1. type 1 error = 5%

2. power = 80% or 90%

3. MCID = 0.3, 0.4 or 0.5 SD

4. estimate of ICC = 0.01, 0.03 or 0.05

5. cluster size in intervention arm = 7

6. attrition rate = 8%.

Table 27 shows the necessary number of participants to be recruited for each combination of trial parameters.

Summary of key points

• The pilot trial was a crucial element of the feasibility study and identified several important issues for the planning of a future definitive trial.

• The initial screening of potential participants was done by third-sector and NHS partners. Unfortunately, we were unable to collect data on the total numbers contacted at this stage.

• Most expressions of interest came via the third sector, but staff/volunteers found this to be more time-consuming than anticipated and a future trial would need to consider how this could be supported.

• The initial method of identifying potential participants by the research team from the Newcastle low-vision clinic was impractical. When the recruitiment source was changed to ECLOs, a small number of VIOP expressed an interest in the trial. The value of recruitment from NHS clinics in a future trial still needs to be considered.

• The estimate of the proportion of those who met the eligibility criteria was biased upwards: a number of VIOP who had initially expressed an interest told the researchers before the eligibility criteria could be formally addressed that they were unable to take part.

• Asking GPs to confirm that VIOP were eligible to take part in the study did not lead to a reduction in study numbers: only one participant was excluded by their GP.

• A total of 64 VIOP were randomised to the intervention or usual care activities trial arms. We had aimed to recruit 80 VIOP. Recruitment was more difficult than anticipated, but this was mitigated by more complete data collection at follow-up.

• Of the 33 participants in the intervention arm, three did not attend any exercise classes. The median number of classes attended was 10 (out of 12).

• VIOP were encouraged to practise exercises for 120 minutes per week in addition to the exercise classes; the median achieved was 50 minutes.

• We aimed to collect outcome data from a total of 60 participants at week 24. We collected outcome data from 59 participants. A total of 95% of trial participants provided data at the week 12 visit and 92% at the week 24 follow-up visit.

• There were very low levels of missing data at baseline, week 12 and week 24 in both trial arms.

• There was little or no evidence that exercise and attitudinal outcomes differed between trial arms at follow-up, but this must be interpreted with considerable caution given the small sample size. There were no SAEs deemed to be caused by the intervention.

Background

A focused literature review was carried out encompassing falls prevention, VI and evaluation by older people of community-based exercise programmes. Few studies had been carried out with a specific focus on VIOP and falls prevention. There were a number of studies that had focused on older people’s views on falls prevention programmes. 84–91 Specific areas that had been linked to VI were physical activity92 and depression. 93 One review article explored the impact of particular VIs on falls13 and two examined the impact of falls on older people in general. 94,95 No articles were identified that specifically focused on the views of PSIs.

Three seminal papers89–91 set out the motivators for and barriers to older people’s engagement in, and continuation with, community falls prevention programmes. These centred on being interested in and enjoying the exercise process and were linked to being previously active, reduced levels of comorbidities and personal confidence levels. Participants expected to gain both physical and psychological benefits that would help maintain independence. Older people were also more likely to attend sessions if they had received a personal invitation from a health-care professional and gained approval from family and friends.

Barriers to engagement centred on the perception of falls risk. Many older people denied having a falls risk and felt that falls prevention programmes were not beneficial to them. 89 The perceived threat of falling and its potential consequences were not motivating factors. 91 Some older people felt that the advice given was ‘common sense’ and not relevant to them and that programmes were for the more disabled (feelings linked to their denial of a falls risk). 89,90 The studies also suggested that some older people felt that they were too ill or in too much pain or that exercising does more harm than good. 89

Some older people thought the engagement with falls prevention programmes would be a threat to their personal identity, eroding their personal autonomy. 90 These findings were corroborated in a study95 that found that older people were careful to present themselves as not the type to fall, constructing an identity of someone who was physically competent. The study participants perceived falling as negative and discrediting, leading to stigmatisation. This was further explored in a study94 that found that those living with falls spoke of losing confidence, independence and their social identity. These identity and autonomy issues were also highlighted in a qualitative synthesis on falls across the life course,96 in which the majority of the studies focused on older people’s perception of falling in later life.

The projection of a capable and competent person with a denial of their falls risk was also found in a study88 designed to promote uptake of and adherence to exercise-based falls prevention programmes. This study also explored the views of the physiotherapists who were involved in the delivery of the exercise-based falls prevention classes. Uptake was found to be positively influenced by a feeling of maintenance of independence and adherence through an approach that enhanced self-efficacy and self-management. In contrast, the physiotherapists were found to be paternalistic, suggesting that the older people were frail and vulnerable, and this conflicted with the older participants’ projected self-image. The findings suggested that falls prevention programmes and those delivering them should focus on self-management and autonomy and avoid being controlling and overprotective.

A further study84 described what motivated older adults to continue with community group-based exercises to prevent falls. The findings again linked motivation to staying independent and staying healthy. Success was linked to personal confidence, functional capability and the ability to stay in control. Poor balance and pain were highlighted as factors that limited motivation.

A qualitative study92 investigating physical activity and older people with sight loss found that the participants engaged with and enjoyed physical activity when it improved physical, mental and emotional health. Emotional and mental health benefits included meeting others and experiencing a feeling of camaraderie. Health benefits were tangible and described as weight loss and improved cardiovascular function. It was also noted that participants had a competitive edge, comparing their own progress with that of others. The participants felt that their ability to exercise was affected more by other comorbidities than by their VI, and it was suggested this may be a consequence of resilience acquired through developing coping strategies to manage their specific VIs. Other research focusing on the views of VIOP suggests that barriers to the uptake of exercise programmes may be perceived risk, stigma, lack of awareness among health professionals, lack of appropriate supporting materials and FoF. 97,98

Several studies86,87,89,96 have discussed older people’s lack of understanding and knowledge of the falls prevention programmes offered as a barrier to uptake. The conclusions drawn were the necessity to present the positive benefits of exercise programmes, better inform individuals of the benefits of specific exercises for falls prevention and ensure that information in whatever form is non-patronising. 90,95 Lack of engagement was also linked to a fatalistic acceptance that falls and declining health are inevitable in older age. 92,96,97 One paper13 reviewed the literature on visual loss and falls. It discussed the multifactorial nature of falls risk in older people and how visual input was required for coordination, planning movement and maintaining balance. The work focused on the common types of VI and how each may increase falls risk. For example, age-related macular degeneration (AMD) affects the central and reading vision, leaving the peripheral vision mostly unaffected, and this has an impact on the ability to navigate, with two-thirds of those affected having visual motor and balance deficits. Older women with AMD were shown to have more impaired balance, slower visual reaction times and poorer vision than age-matched control subjects.

In conclusion, the body of work reviewed identified facilitators of and barriers to community-based exercise programmes. There is very little research on falls and VI in later life. The majority of work focused on older people in general. The literature also suggests that specific types of VI may impact on physical activity in different ways and that older people do not consider their VI to be a major barrier to exercise.

Methodology

A qualitative methodology and methods enabled the researchers to gain insight into the views and feelings of those delivering the adapted exercise programme (the PSIs) and those who took part in the exercise programme (the VIOP).

Data collection

Data were collected from both groups (PSI and VIOP) through one-to-one interviews facilitated by the use of topic guides (see Appendix 2). The PSI interviews were structured and focused and the VIOP interviews were semistructured. All interviews were conducted at a mutually convenient time and place and were audio recorded and transcribed.

Data analysis

Data analysis involved six steps: (1) organisations of data, (2) reading and memoing, (3) describing, (4) classifying, (5) interpretation and (6) representation. 98 An initial data analysis clinic was held with the interviewers, project manager and one of the PIs (L de J, DC, CB or DS). All transcripts were read and initial categories were identified, discussed and agreed by those present. These initial findings were further refined by the lead qualitative researcher (DC) and themes were produced. Themes were generated in alignment with the research objectives. As part of the data validation process the findings were presented to and discussed by the PAG.

Findings

Conclusion

The qualitative evaluation of the VIOLET study indicates a number of discussion points.

• Volunteering for the research project and the exercise programme was primarily driven by an altruistic desire to be useful and contribute to wider society. Although the exercise sessions may not have been as some had expected, the wish to be helpful, fulfil their commitment and have social interaction promoted retention. As noted within other studies96 it is likely that simply volunteering for the study is indicative of these participants having high personal confidence and being resilient problem solvers.

• The perceived relevance of the exercise intervention acted as both a facilitator of and barrier to participation. It was linked to understanding of the research and the relevance of the information being collected. Where there was a limited understanding or misunderstanding and where relevance was not felt to be strong, the participants suggested this detracted from both the research process and the exercise sessions. Had more information been given on the types of exercise and the benefits of strength and balance exercises, the participants may have been able to see the relevance of this specific exercise programme more readily. However, it may be the case that 12 weeks of sessions was not sufficient for the majority of participants to report tangible benefits.

• Participants wanted to present themselves as persons who were active, fit and not at risk of falls. There is thus a need to review falls prevention programmes to take into account the use of language and actions that directly challenge self-perception by including an education component to the programmes. Participants’ self-portrayal of not being at increased risk of falling was supported by low falls risk scores.

• Interviewees would have liked more information to have been collected about their VI and its impact on daily life; this had already been noted by the research team and added to the interview topic guide. In future studies it may be beneficial to address what appears to be an imbalance between the perceived falls risk and the impact of VI on that risk.

• Like the PSIs, the VIOP also suggested that for less physically able participants it would be beneficial for a second person to assist with the floor work element of the exercise sessions. Further consideration across the mainstream community falls prevention setting is warranted.

• The participants liked the home exercise information, although access could be improved by both offering a wider range of information technology modalities and enabling participants to progress at their own pace. This could be achieved by providing audio material that can be easily paused, revisited and generally modified to individual preference. Participants suggested formats such as DVD, DAISY, screen reader, voice synthesiser, MP3, talking book and Braille.

• In agreement with a body of research, home exercising was hard to sustain, particularly once the 12 weekly group classes were completed. Participants missed the peer support, camaraderie and overall social contact. There were also some cognitive (memory) issues and it seems that, in this setting, the home exercise component could not be relied upon to make up the total exercise time required to achieve efficacy.

• Almost all participants suggested that the social time at the end of the classes was very enjoyable and participant and PSI interviewees suggested that this could be lengthened and may provide a platform for an informal sharing and exchanging of broad health information. This should be considered in further research.

Participants reported that they did not like being referred to as ‘older people’ or considered as ‘old and frail’. It is suggested that such programmes are termed as ‘facilitating independence’ in a community-based sample and may include other aspects such as support and education in addition to an exercise component.

Methods for health economics and cost-effectiveness

The main focus of the study was on how to identify, quantify and value accurately the additional costs of delivering the exercise programme and the potential resource implications versus usual care.

The study collected data on the use of services within the health-care system, social care costs and patient and carer costs in an attempt to capture relevant components that constitute the overall costs from a health and social services perspective both with and without an exercise programme. Key items included data completeness and the costs of the control and experimental interventions.

Summary

The current feasibility study has shown that, although it is possible to collect most of the data necessary for a full cost-effectiveness analysis of the exercise intervention compared with usual care (cost of intervention per se, utility values and capability values), there were some practical issues in accessing information regarding participant self-reporting of resource use post intervention. Participant inputs on use of health and social care services and broader service use were often missing, despite telephone calls by researchers. The data collection instrument for this was the participant self-reported falls resources/expenses form (including informal caregivers’ time). This was based on a series of open-ended questions as it was initially thought by the study team that carers filling out the form would find these types of questions easier to respond to. However, the reality of this form of data capture was that it was not clear in the absence of recorded information whether data were missing or participants had not actually received any formal care. Furthermore, detail of the type of care was also lacking, rendering any estimation of costs of the use of health and personal social care resources impossible. The use of a more structured, previously piloted, data collection tool may have mitigated against some of these issues. Regardless of the design of the data collection tool, it is well documented that reliance on patients/carers as a data collection method is limited by biases in recall, nonresponse and evasiveness. 106 Furthermore, missing data are a common problem for economic evaluations that run alongside clinical trials107 and may be mitigated in a future study by either postal or telephone reminders to participants.

The HRQoL scores showed that participants in both arms of the trial at all time points were in a health state worse than death. However, capability scores were towards the high end of being capable. Perhaps this should not be surprising as the ICECAP measures potentially offer a broader assessment of QoL and well-being than measures routinely used in economic evaluation such as the EQ-5D-3L. 108 This broader assessment may allow measurement of the full effects of an intervention or treatment. Previous research has indicated that the ICECAP-O and EQ-5D-3L measures provide complementary information and are not substitutes for each other. 80 The same may be true for the EQ-5D-5L; however, this is not currently addressed in the literature.

• Costs of intervention delivery in terms of staff time were cheaper at the Newcastle site than at the Glasgow site. The Newcastle site ran three cycles of the intervention whereas the Glasgow site ran two; however, recruitment to the intervention arm was the same at both sites.

• There were no missing data for the EQ-5D-5L instrument. Compared with baseline, the utility scores improved in both groups but still showed average HRQoL as being in a state worse than death. Overall, at all three data collection points HRQoL was higher in the usual activities group than in the intervention group.

• All participants completed the ICECAP-O instrument with the exception of one who failed to complete one domain at one time point. Capability scores differed minimally across trial arms at each of the follow-up times. Furthermore, in both trial arms, there was little difference between average scores from baseline to week 24. Capability scores ranged from 77% to 83% across the two trial groups.

Patient and public involvement

There was patient and public involvement at all stages of the study from inception to completion and dissemination and various methods allowed the participants to contribute to all stages of the study. Of the stakeholder VIOP members (14 participants across the two sites), four expressed an interest in participating in the PAG and attended meetings. We also invited study participants from control and intervention groups to participate in a VIOLET study feedback session in December 2016 to share our findings, elicit their feedback and invite their opinion on next steps. Throughout the duration of the study we also had close contact with our Vision charity partners and, following the launch of the study (hosted by NSBP and with VIOP stakeholders present), we produced a press release announcing the study’s intent to work closely with VIOP to adapt a validated falls prevention programme (FaME) to their needs. This was endorsed by both Visibility – our Glasgow site vision charity partner – and the study PAG representative from the RNIB.

Additionally, the participation and membership of the PAG enabled VIOP to provide valuable input at all stages of the research regarding design, recruitment, adaptation of the intervention and dissemination. Participants also reviewed study materials and made recommendations for both content and presentation of the participant information sheet and Plain English summary.

An exploration of visually impaired older people’s ability to act as lay partners in a study to develop a condition-appropriate intervention (objective 1)

In the first phase of the study, the use of focus groups facilitated VIOP to express individual intervention preferences and needs, such as PSIs being specifically trained to accommodate differing functional and visual needs. There was also a strong preference for a range of choices in the way VIOP engage with home exercise and with the completion of outcome measures, such as the use of weekly telephone calls. The feasibility study enabled VIOP to contribute to the adaptation and design of an acceptable community-based group exercise programme that could be incorporated into a general falls prevention exercise programme.

Visually impaired older people stakeholders also offered guidance on adapting the FaME intervention to maximise uptake and adherence for VIOP. No matter how effective an intervention is, if participants cannot or do not want to engage with it, it is effectively useless. The participant qualitative interviews elicited experiences of undertaking the intervention, which identified many valuable insights into facilitators of and barriers to exercise participation, and highlighted the role of self-perception that underlies exercise participation behaviour.

Assessment of recruitment of visually impaired older people and their willingness to be randomised (objective 2)

Although the study did not recruit the target figure, we feel that the recruitment strategy was largely effective, subject to modifications and suggestions outlined. Variation in the way participants were identified across the two study sites has led to some discrepancies in the recording of eligibility for the study. For participants who had expressed interest in the study, assessment of eligibility was guided by research staff. However, the recruiting organisations did not keep a record of how many initial contacts they made to produce the expressions of interest. Anecdotally, we are aware that > 300 direct contacts (telephone or letter) were made to potential participants, which converted to 105 expressions of interest. For a future study, the recruiting organisations would need to commit to keeping an accurate count of contacts.

Overall, recruiting from third-sector organisations was successful; however, there were lessons learned. Both third-sector recruitment organisations (NSBP and Visibility) identified large numbers of members who they felt would be eligible. These did not convert to actual participants. Organisational feedback suggested that the conversion rate from initial contact to expression of interest was poor. Unfortunately, at this point, we did not collect ‘reason’ for decline, which is needed to fully evaluate the recruitment strategy from this source. In addition, one of the third-sector organisations was able to provide a dedicated member of staff for recruitment whereas the other was not. Identification of potential participants was therefore more burdensome to the third-sector organisations than expected. Furthermore, the available samples from the organisations at Newcastle and Glasgow have now been exhausted and a future study would need to recruit from other sources and sites. In future multisite studies a recruitment strategy should be discussed and agreed across recruiting organisations with additional resource and support for these organisations. It is possible that, in future, a local council register of people who have VI could also be used, as proposed by the recommendations of the VIP2UK study. 11

It proved difficult to recruit from the Newcastle NHS site – the RVI low-vision clinic. Ophthalmologist/optician staff were not always best placed to recruit from within busy and fast-paced outpatient clinics. The RVI is the regional centre of excellence, with many of those attending living outside the Newcastle area and too far away to travel to the classes provided for the intervention group. In a future study, it may be advisable to use the primary care setting with the dual purpose of GP-led assessment of health exclusions and recruitment. Previous research suggests that a ‘personal’ invitation by a health-care professional increases uptake of community exercise classes. 109

On the other hand, the ECLOs based at the RVI were committed to the study, had meetings with the study team from the outset and were able to dedicate a member of staff to identify potential participants and pass on expressions of interest. This recruitment strategy was very successful with a conversion rate of nearly 50%. That said, a lack of accurate data in the pre-screening phase makes it difficult to quantify initial contacts made by the ECLO. In future studies, more information is needed on the pre-screening phase to better articulate commitment/resource requirements. There was no recruitment from the Glasgow Caledonian University eye clinic owing to the summer vacation and we acknowledge that there is a need to take into account the ‘working time’ of all potential recruitment sites.

The impact of GP assessment was negligible as only one participant was screened out on grounds of health according to the inclusion/exclusion criteria.

Only two of those originally found to be eligible for the study declined to take part in the pilot trial. A further two subjects were lost to the study before randomisation as other exclusion criteria came to light. Therefore, we consider this objective to have been successfully achieved.

Delivery of visually impaired older people-appropriate Falls Management Exercise falls prevention intervention and production of manualised intervention (objectives 5, 6 and 8)

A key component of the VIOLET study was the successful delivery of the VIOP-adapted FaME programme by HealthWORKS in Newcastle and Visibility in Glasgow, led by PSIs. The PSIs had established experience of delivering falls prevention intervention programmes in the NHS, including FaME. For each cohort, quality assurance checks were performed, ensuring fidelity of the programmes at both sites. We recommend for future studies that participating PSIs should have an opportunity to attend two workshops, to share and discuss findings regarding successful delivery of the intervention, and that quality assurance is maintained to ensure fidelity.

Guidelines recommend 36 hours of exercise per falls prevention exercise programme; over the 12 weeks of participation, this equates to a total of 3 hours per week. However, VIOLET study participants had a mean time of 110 minutes per week: 50 minutes spent on home exercise and 60 minutes on the weekly group exercise intervention. Both the VIP and VIP2UK studies11,49 reported poor adherence to the home exercise component, although the VIP study suggested that stricter adherence was associated with fewer falls. 49 VIP2UK attempted to increase adherence to home exercise but was not successful. 11 Furthermore, some studies suggest that a 12-week intervention comprising 36 hours of exercise may not provide sufficient balance and strength work to reduce falls. 62 However, as noted earlier, we took a pragmatic approach to what is achievable in order for NHS or local authority-led falls services to be encouraged to adopt the adapted VIOP FaME approach. Research suggests that, on average, falls services across the UK deliver sessions once a week for eight to 12 weeks. 27,60 We strongly support continued development of strategies to increase adherence to home exercise for future studies so that an effective exercise dose is delivered.

Delivery also highlighted the need for an additional exercise professional to support participants with multiple comorbidities such as deafness and assist with the floor work element. In addition, a variable range of participant ability within the exercise classes led to some participants reporting not being physically challenged, which was endorsed by the instructors.

Importantly, VIOP themselves suggest that other comorbidities, such as shortness of breath, may have greater impact on their ability to exercise than their VI. The PSIs did not change the exercise components; rather, they modified their delivery style to the needs of VI in the adapted FaME programme. In their post-intervention interviews, both VIOP and PSIs felt that the majority of VIOP could be integrated into mainstream community classes, which should be stratified by functional ability to address the issues of challenge identified in the interviews by both participants and PSIs. Only those with multiple comorbidities would not be able to be integrated and this would occur in mainstream classes anyway. These individuals should continue to be offered one-to-one exercise interventions as appropriate.

Visually impaired older peoples’ recommendations for the adaptation of FaME delivery and experiences from the perspectives of the PSIs for more effective delivery in future, have been incorporated into the full manualised intervention of the adapted FaME intervention (a copy of this is available from the VIOLET project page on the NIHR website; see www.journalslibrary.nihr.ac.uk/programmes/phr/1218107/#/).

An exploration of visually impaired older people’s reasons for participating in the adapted Falls Management Exercise programme and acceptability of the exercise programme and trial procedures (objective 7)

Volunteering for the study was primarily driven by altruism. Participants expressed a desire to be helpful, fulfil their commitment and enjoy social interaction with other participants, which ultimately promoted retention. Simply volunteering for the study is indicative of these participants having high personal confidence and being resilient problem solvers. 110 Participant interviews also highlighted that interviewees would have liked more information collected regarding their VI and its impact on daily life. In future studies it may be beneficial to address what appeared to be an imbalance between the perceived falls risk and the impact of VI on this risk.

Some participants suggested that they did not understand the relevance of specific exercises. Had more information been given on the types of exercise and the benefits of strength and balance exercises, the participants may have been able to understand the relevance of this specific exercise programme. However, it may be the case that 12 weeks of exercise classes was not sufficient for the majority of participants to notice tangible benefits. Indeed, echoing related studies with VIOP28,31 many of our participants self-presented as active, fit and not at risk of falls or being a ‘faller’. The self-description of ‘not a faller’ was also supported by the falls risk scores, suggesting on the whole that these participants had a low falls risk. A future study should perhaps ensure that those who are at risk of a fall are recruited, as falls services often have the inclusion criterion of a previous fall or distinct FoF.

Self-perception as ‘not a faller’ was common and, as is often highlighted in related literature,40,41 participants also reported a dislike of being described as ‘older people’ or considered as frail. Falls prevention programmes need to emphasise facilitating independence in a community-based sample as well as other positive benefits, such as socialising and receiving useful health education, in addition to the exercise component. In this feasibility study, participants in the intervention group were slightly more afraid and more likely to avoid activity at the end of exercise classes. However, the difference was much smaller at week 24. It is possible that the classes warn people about falling, which initially makes them more likely to avoid activity, and this is well supported in the literature. 111 Conversely, research also highlights how, for some participants, there could be an increase in falls during the initial phase of increasing exercise. 109,111

Home exercising was hard to sustain, particularly once the 12 weekly group classes were completed: this concords with other research. 11,49,111 Although participants were encouraged to exercise for a further 2 hours per week, the median was reported to be 50 minutes per week. Participants missed the peer support, camaraderie and overall social contact. There were also some cognitive (memory) issues and it seems that, in this setting, the home exercise component could not be relied upon to make up the total recommended exercise time. Perhaps one consideration is the ability to attend more group sessions in a week rather than rely on home exercise. Almost all of the participants suggested that the social time at the end of the classes was very enjoyable. Participant and PSI interviewees suggested this could be lengthened and may provide a platform for an informal sharing and exchanging of broad health information. This, as a recommendation, should be considered in further research.

Similar to the PSIs, the VIOP interviewees highlighted that for less physically able participants it would be beneficial for a second person to assist with the floor work element of the exercise sessions. The participants liked the home exercise information, although access could be improved by both offering a wider range of information technology modalities and enabling participants to progress at their own pace. This could be achieved by providing audio material that can be easily paused, revisited and generally modified to individual preference. Participants suggested DVD, DAISY, screen reader, voice synthesiser, MP3, talking book and Braille formats. This area would require further attention should the study progress to a RCT.

Identification of candidate outcome measures for a future randomised controlled trial (objective 3)

The completion rates of the outcome measures were very high. We aimed to collect outcome data from a total of 60 participants at week 24; we collected outcome data from 59 participants. A total of 94% of trial participants provided data at week 12 and 92% provided data at week 24. This was probably due to the completion of the outcome measures with the researchers. Although this resulted in a high completion rate, it is noted that this was very time-consuming for the researchers and that this would need to be considered in future studies. The primary potential (FES-I) provides an assessment of falls risk. In our sample, falls risk was generally low to moderate and thus we did not recruit high-risk fallers for whom the exercise programme may have been more beneficial. This could therefore be used in subsequent studies to stratify participants both in practice and for the purpose of data analysis. The MCID of the FES-I is not known, and thus it was possible to carry out only an estimated sample size calculation based on standardised differences. The measures were acceptable to participants and not perceived to be onerous. The HADS is validated for use in the older population and assesses anxiety, a key predictor of FoF. However, the relevance of some of the QoL questions in the HADS were questionable to the older population. In general, our VIOP took longer to perform the TUG time/functional ability assessment than that reported for healthy community-dwelling subjects aged 65–84 years. 112 Future trials with this population should consider an assessment that is not so reliant on vision. For example, the short physical performance battery is a group of measures that combines the results of gait speed, chair stand and balance tests, although this also requires some vision for safety. 113 It is also suggested that additional outcome measures such as those that assess frailty may be included for this population to provide an indication of whether or not exercise is maintaining levels of resilience even if it does not lead to an improvement. This may be more appropriate and realistic for an older population with many comorbidities. This is a suggestion that may also be useful for assessment in mainstream falls prevention programmes. Further, a longer follow-up period of 12–18 months would also be recommended in future studies, in accordance with related findings. 113

Feasibility of collecting data for health economic analysis (objective 9)

There is evidence to suggest that it is feasible to collect service-use data for an economic evaluation of the intervention in a future RCT, although there were some challenges. It is possible to collect most of the data necessary for a full cost-effectiveness analysis of the exercise intervention compared with usual care (cost of intervention per se, utility values and capability values); however, participant inputs on use of health and social care services and broader service use were often missing owing to the data collection instrument (falls resources/expenses form). It was not clear, in the absence of information recorded, if data were missing or participants had not actually received any formal care. Furthermore, detail of the type of care was also lacking, rendering any estimation of costs of the use of health and personal social care resources impossible. The use of a more structured, previously piloted, data collection tool with either postal or telephone reminders to participants may have mitigated against some of these issues. This could potentially be completed more often to improve recall and accuracy. In addition, processes that allowed access to the required information from the primary medical source may have negated the issue.

The data completion rate was almost 100%. The HRQoL scores identified that some participants in both arms of the trial at all time points were in a health state worse than death. However, capability scores were towards the high end of being capable. The ICECAP-O potentially offers a broader assessment of QoL and well-being than measures routinely used in economic evaluation such as the EQ-5D-3L. This broader assessment may allow measurement of the full effects of an intervention or treatment.

Methodological implications and sample size calculation for future trial (objective 4)

Within the VIOLET feasibility study there was no assessment of VI. This was largely a pragmatic decision; however, we recommend assessment of VI in future studies. Assessment of VI would perhaps have allowed us to stratify VIOP participants, identifying those who might be able to join a mainstream class and those who would require one-to-one intervention or more intensive supervision. The inclusion criteria did not allow screening out for ‘deafness’. People who are profoundly deaf and have a VI are difficult to accommodate in a group setting. The assessment of whether or not potential participants were ‘physically able to take part in a group exercise class’ was, on the whole, carried out when face to face and thus open to individual interpretation. This was also the case when assessing a participant’s ability to walk indoors and outdoors with or without aid. We found that the majority of participants had only low to moderate concern about falling and did not report high levels of anxiety. We would therefore recommend assessment of falls risk in a future study along with assessment of VI.

A number of potential participants ‘self-reported’ an uncontrolled medical condition as a reason for exclusion. There is a need to think about what a potential participant considers to be an uncontrolled medical condition and the degree of concordance with a GP. These issues would be negated to some degree if recruitment was carried out in primary care and assessment was carried out by a GP rather than by a participant or a researcher (who might not be medically trained), although our method replicates current practice in falls services.

The PSIs also suggested stratification by functional ability and some participants discussed not being physically challenged within the exercise classes. Mixing able and less able participants has an impact on the challenge and potential effectiveness of the programme. It is recommended that stratification by functional ability is considered in a definitive trial.

If the same method of recruitment was used in a definitive trial, it would be necessary to allow for a staggered start to the exercise classes to compensate for delays in GP confirmation of eligibility and to maintain timely enrolment into the exercise sessions.

As outlined in Chapter 4, sample size calculations were performed to provide a sample size for varying combinations of trial parameters. For example, at 80% power, a MCID of 0.5 SD, an estimated ICC of 0.01, a cluster size of 7 and an attrition rate of 8%, 72 participants would be required in the intervention group and 68 in the usual activities group. However, this could rise to requiring 328 participants in the intervention group and 252 in the usual activities group if the power was 90%, the MCID was 0.3 SD and the ICC was 0.05.

Does the feasibility study indicate that a full trial should be conducted?

The criteria to determine the feasibility of progressing to a full trial were all met (see Chapter 3); that is, ≥ 50% of VIOP eligible for the study were willing to be recruited into the feasibility study; ≥ 70% of the participants in the intervention arm completed 9 out of 12 group sessions in the exercise programme (compliance); data were collected on key outcomes at the 6-month follow-up for ≥ 70% of those recruited; and, finally, < 10% of SAEs were deemed to have been caused by the intervention itself.

This suggests that this pilot study could be taken to a full study. However, recommendations from the VIOP and the PSIs and discussion within the PAG, trial management group (TMG) and TOC suggest that most VIOP could be integrated into a mainstream class. Training [Continuing Professional Development (CPD)] of PSIs to accommodate a range of VIOP in their mainstream sessions is minimal and easy to facilitate through online training, so would be affordable and have sustainable reach. Only those with multiple comorbidities (such as extreme deafness or extreme frailty) required significantly more supervision. Mainstream classes can be much larger, so, if an elderly person with VI was to join a mainstream class, an extra supervisor may be necessary and should be considered, particularly if the participant’s VI necessitates the adaptations to teaching suggested by the PSIs (e.g. touch, fuller description of the room and the people in it).

However, if a future study were to recruit those more at risk of a fall, those with more defined and limiting VI and those who are not already active in exercise sessions, our recommendations might be different. Should this progress to a future study, these inclusion/exclusion criteria should be refined and recruitment rates may differ.

Limitations

The following limitations were identified:

• The sample typically exhibited low to moderate falls risk. The programme would have benefited high-risk fallers and these should be targeted for recruitment in future studies.

• Recruitment proved difficult and we did not recruit the originally planned sample of 80 participants, though retention was better than anticipated.

• It is possible that there was bias by having the weekly telephone call, which may have acted like an intervention in itself.

• We recruited only English-speaking participants as one of the exclusion criteria for the trial was being unable to comprehend or follow simple movement instructions in English (in the event that they were randomised to the intervention arm).

• There was no assessment of VI.

• The missing data for the health economic evaluation could have been mitigated by the use of previously piloted instruments for data capture.

Conclusions and recommendations for future research

It was possible to recruit and retain participants to the VIOLET study and adapt an existing, widely used exercise intervention for falls prevention (FaME) with VIOP as stakeholders. In future multisite studies, a recruitment strategy should be discussed and agreed across recruiting organisations. It is suggested that recruitment from primary care practices and use of a local register, as proposed by the VIP2UK study,11 should be considered for a community-based intervention.

The intervention was delivered successfully at two sites. Adherence to the study was high with very low attrition rates. This is a strength of the study.

Whereas guidelines recommend 36 hours of exercise per falls prevention exercise programme over the 12 weeks of participation, VIOLET study participants had a mean time of 110 minutes per week: 50 minutes of home exercise and 60 minutes of the weekly group exercise intervention. It is possible that this would be insufficient to bring about any improvement; however, whether exercise had little impact or halted deterioration was not investigated in the VIOLET study. This is an important consideration for future research. Adherence to and duration of exercise is a key consideration in any exercise programme and this should therefore be a critical focus in any future study. Furthermore, PSIs referred to the level of challenge of the exercises for some of the participants, suggesting that there should be stratification based on ability. It is recommended that participating PSIs should have an opportunity to attend two workshops to share and discuss findings regarding successful delivery of the intervention.

Home exercising was hard to sustain, particularly once the 12 weekly group classes were completed. Thus, the home exercise component could not be relied on to make up the total recommended exercise time. The use of technology as prompts may assist in future studies, as may the option, as suggested by the stakeholder panel, for participants to attend multiple groups in a week. Participants enjoyed the social time at the end of the classes, with participant and PSI interviewees suggesting that this could be lengthened and may provide a platform for an informal sharing and exchanging of broad health information. As some participants did not understand the relevance of the exercises and the impact of exercise in maintaining resilience and halting deterioration, it is suggested that the provision of information is also an important part of the exercise intervention. This, as a recommendation, should be considered in further research.

The outcome measures were suitable to assess the study variables and had a high level of completion with very little missing data, with 94% of trial participants providing data at week 12 and 92% providing data at week 24.

Although the progression criteria were met, the following points have been suggested for a future trial.

• A future definitive trial should consider stratification of participants with low, moderate and high falls risk.

• There should be an assessment of VI.

• Recruitment sources could include primary care practices and involve the multidisciplinary team.

• Non-participants should be interviewed to explore their reasons for not participating.

• The social element of the exercise programme could be expanded to include an education component addressing the benefit of the exercises to maintain and improve levels of physical and functional ability and thus reduce falls risk.

• The development of strategies to increase physical activity and structured exercise at home should be explored to reach recommended levels of activity.

• For the economic evaluation, there should be a resource-use questionnaire that utilises more closed-ended and focused questions, and this should be administered at more time points. This may improve recall and accuracy.

• Other hard to reach groups could be targeted in future to participate in the FaME programme.

It is concluded that mainstream programmes for falls prevention could be adapted to incorporate VIOP; interviews suggested that it is the delivery style and logistics of attending venues rather than the nature and type of the exercises that is a barrier to participation in exercise programmes. Stratification according to falls risk and functional ability are recommendations of the study. The adapted FaME programme, which has previously published evidence of effectiveness, was found to be both safe and acceptable to participants.

Contributions of authors

Nicola Adams, Professor of Rehabilitation, was the Chief Investigator, took overall responsibility for the study and the writing of this report.

Dawn Skelton, Professor of Ageing and Health, as Principal Investigator for the Glasgow site of the study, contributed to the design of the study, led on the production of the adaptation of the FaME intervention manual and contributed to the report.

Cathy Bailey, Senior Research Fellow in Ageing, project managed the study and contributed to its design and the writing of the report.

Denise Howel, Senior Lecturer, Statistics, contributed to the design of the study and conduct of the trial, managed the analysis and interpretation of the quantitative and statistical results and co-drafted relevant sections of the report.

Dorothy Coe, Senior Research Assistant, qualitative research, managed the Newcastle site data, co-analysed the qualitative data and produced the qualitative findings chapter.

Rosy Lampitt, Trials Manager, and Jennifer Wilkinson, Senior Trials Manager, both of the Newcastle Clinical Trials Unit, contributed to the design of the study and managed the conduct of the trial, and Rosy Lampitt contributed to the report.

Tony Fouweather performed the statistical analyses and co-drafted the statistical results chapter.

Lex D de Jong, Research Associate and physiotherapist, research managed the Glasgow site, co-analysed the qualitative data and contributed to the report.

Sheena Gawler, Research Assistant at the Newcastle site, managed quality assurance and contributed to the intervention manual and this report.

Vincent Deary, Associate Professor of Psychology, contributed to the design of the study, led on the patient and public involvement and drafted the PPI chapter.

Joanne Gray, Senior Lecturer, Health Economics, contributed to the design of the study, conducted the health-economic analysis and drafted the health-economic analysis chapter.

Heather Waterman, Professor of Nursing and Opthalmology, contributed to the design of the study and the report.

Steve W Parry, Consultant Physician and Clinical Senior Lecturer, contributed to the design of the study and the report.

Publications

Adams N, Skelton DA, Howel D, Bailey C, Lampitt R, Fouweather T, et al. Feasibility of trial procedures for a randomized controlled trial of a community based group exercise intervention for falls prevention for visually impaired older people: the VIOLET study. BMC Geriatrics 2018;18:307.

Skelton DA, Bailey C, Howel D, Cattan M, Deary V, Coe D, et al. Visually Impaired OLder people’s Exercise programme for falls prevenTion (VIOLET): a feasibility study protocol. BMJ Open 2016;6(Suppl. 8):e011996. https://doi.org/10.1136/bmjopen-2016-011996

Data-sharing statement

All data requests should be submitted to the corresponding author for consideration. Access to anonymised data may be granted following review. Exclusive use will be retained until the publication of major outputs.

Patient data

This work uses data provided by patients and collected by the NHS as part of their care and support. Using patient data is vital to improve health and care for everyone. There is huge potential to make better use of information from people’s patient records, to understand more about disease, develop new treatments, monitor safety, and plan NHS services. Patient data should be kept safe and secure, to protect everyone’s privacy, and it’s important that there are safeguards to make sure that it is stored and used responsibly. Everyone should be able to find out about how patient data are used. #datasaveslives You can find out more about the background to this citation here: https://understandingpatientdata.org.uk/data-citation.

Disclaimers

This report presents independent research funded by the National Institute for Health Research (NIHR). The views and opinions expressed by authors in this publication are those of the authors and do not necessarily reflect those of the NHS, the NIHR, NETSCC, the PHR programme or the Department of Health and Social Care. If there are verbatim quotations included in this publication the views and opinions expressed by the interviewees are those of the interviewees and do not necessarily reflect those of the authors, those of the NHS, the NIHR, NETSCC, the PHR programme or the Department of Health and Social Care.

Graph of cumulative recruitment against calendar month in 2015

FIGURE 5.

Graph of cumulative recruitment against calendar month in 2015.

Ethics approval

The study (REC reference: 15/NE/0057) received ethics approval from the National Research Ethics Service Committee North East – Newcastle and North Tyneside 2 on 9 March 2015.

University ethics approval was obtained for the stakeholder focus groups from Northumbria University, Newcastle. Reference: RE-HLS-13-140707-53bb0a7806e37.

Protocol amendments

A minor amendment to the study was submitted to the REC in October 2015 to enable the study team in Newcastle to recruit from additional clinics within the Newcastle Eye Centre at the RVI and not just from the low-vision clinic to aid recruitment.

A substantial amendment was later approved by the REC in January 2016, introducing a usual activities arm participant exercise offer letter and usual activities arm additional GP letter. This was intended to facilitate the offer of a 12-week equivalent exercise programme to the control group at the end of the study – a REC request prior to the original study approval.

A further amendment was submitted in March 2016 regarding a number of minor changes to the protocol. The main change was the removal of reference to the qualitative interviews being transcribed onto NVivo 10 (QSR International, Warrington, UK), a qualitative data management and searching program, as the TMG decided that, as the data set was small and manageable (n = 10, five at each study site), the use of NVivo was unnecessary: interviews at the Glasgow site were transcribed by an experienced transcriber from Northumbria University. Additionally, information regarding thematic analysis was rewritten to give a more accurate picture of the analysis.