An intervention to improve outcomes of falls in dementia: the DIFRID mixed-methods feasibility study

Article history

The research reported in this issue of the journal was funded by the HTA programme as project number 13/78/02. The contractual start date was in June 2015. The draft report began editorial review in January 2019 and was accepted for publication in April 2019. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The HTA editors and publisher have tried to ensure the accuracy of the authors’ report and would like to thank the reviewers for their constructive comments on the draft document. However, they do not accept liability for damages or losses arising from material published in this report.

Declared competing interests of authors

Louise Robinson reports grants from the National Institute for Health Research (NIHR) Professor award and a NIHR Senior Investigator award outside the submitted work. Lynn Rochester reports grants from the Medical Research Council, the European Union, the Engineering and Physical Sciences Research Council, the Wellcome Trust (London, UK), Parkinson’s UK (London, UK), NIHR [Programme Development Grant, Physical Activity Interventions to Improve Outcomes for people with Rare Neurological Conditions (PARC), co-applicant (2018–20); and Health Technology Assessment, PDSAFE: a randomised controlled trial of the effectiveness of PDSAFE to prevent falls among people with Parkinson’s disease, co-applicant (2013–17)] and the Stroke Association (London, UK) outside the submitted work.

Copyright statement

© Queen’s Printer and Controller of HMSO 2019. This work was produced by Allan 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

Background and context

Recent estimates suggest that there are 850,000 people living with dementia in the UK, of whom 70% live in the community. 1 People with dementia (PWD) living in their own homes experience almost 10 times more incident falls than other older people and their falls are more likely to result in injury. 2 When injuries are sustained, PWD are less likely than cognitively intact older people to recover well. 3

Evidence shows that falls are a common reason for hospital admission in PWD4 and that most admissions for PWD with an injury are due to a fall. 5 Despite this, current UK guidelines for treatment of older people following a fall do not specifically address the needs of PWD;6 the new dementia guidelines7 recommend that falls services address the specific needs of PWD, but provide few details on how this can be achieved. The World Health Organization8 report on falls prevention in older people refers to cognitive impairment only as a risk factor for falls. There is little evidence regarding the care pathways currently experienced by PWD presenting with a fall-related injury.

For older people without dementia, there is good evidence that a multifactorial intervention by a specialist falls service will prevent further falls. 9–12 Such interventions are usually tailored to the individual and are directed at known risk factors for falls. However, their effectiveness for PWD is unclear. 13 It is possible that the lack of demonstrated efficacy is because risk factors for falls may differ in PWD or be more frequent or specific to dementia, for example wandering,14 behavioural disturbance,15 Parkinsonism,16,17 severity of cognitive impairment,16 functional impairment18 and use of neuroleptic drugs. 19,20 Nevertheless, and despite the lack of evidence, PWD are often referred directly to the local falls service. Such services are not usually tailored to meet the needs of PWD. It is possible that the referral may achieve other benefits for the PWD, such as medication review, treatment of other comorbidities or provision of aids to support activities of daily living (ADL), but it is not known if a falls service is the best setting for addressing these goals. Indeed, it is not known what goals are of most importance to PWD who fall.

In designing any kind of intervention to address the problem of fall-related injuries in PWD, it is vital that the intervention addresses outcomes of importance to PWD themselves, their informal carers (i.e. unpaid family members of friends who support the PWD, hereafter referred to as carers) and their care professionals. We accessed the Core Outcome Measures in Effectiveness Trials (COMET) database21 and found no consensus regarding suitable outcomes for fall-related injury, although there were two publications regarding interventions of relevance in this situation: the Prevention of Falls Network Europe (ProFaNE) consensus on a common outcome data set for fall injury prevention trials22 (domains include falls, injuries, psychological consequences of falling, health-related quality of life and physical activity) and the European Consensus on outcome measures for psychosocial intervention research in dementia care23 (domains include patient mood, quality of life, ADL, behaviour, carer mood and carer burden).

There is a range of ways in which improved management of fall-related injuries might reduce adverse sequelae for PWD and their carers. First, any fall in older people, whether injurious or not, is known to frequently result in fear of falling and psychological morbidity, which may lead the person to restrict their mobility, resulting in deconditioning and a cycle of further loss of mobility and frailty. 24 A successful intervention may reduce psychological morbidity and improve well-being. 25 Second, if physical recovery from the injury itself is poor, further restriction of mobility may happen and independence in ADL may decline. These restrictions may result in reduced social participation, increased burden for informal carers and increased need for formal care. Such problems lead to reduced well-being and quality of life for PWD, and substantial costs to both health and social care systems. A successful intervention may support the maintenance of physical ability and independence or reduce the degree of physical decline and loss of independence. We are not aware of any clinical trials that have specifically tried to address the management of all fall-related injuries in PWD.

Therefore, although PWD who sustain fall-related injuries currently receive a range of health interventions, a single model of care for this specific situation has not previously been described to our knowledge. Given all of the aspects of care relevant to the situation as described previously, it is apparent that a new model of care would take the form of a complex intervention. Given the frequency of falls in PWD, it is clear that this is an important area for research, although the potential demand for such an intervention is not known. There is also no current consensus on the best outcomes with which to measure the impact of such an intervention or its cost-effectiveness. This report describes the process of developing and testing the feasibility of a new intervention to help this patient group.

Research objectives

The overall aim of this study was to assess, through a series of work packages (WPs), whether or not it is possible to design a complex intervention to improve the outcome of fall-related injuries in PWD living in their own homes (Figure 1). During the course of the study, the objective was broadened to include people with a fall necessitating health-care attention and not just those with injurious falls.

FIGURE 1.

Overview of study and report structure.

Introduction

There is no consensus on how best to manage PWD who have had a fall. As part of this study, two reviews were conducted. The first focused on the effectiveness of different interventions targeted at PWD who have experienced a fall. The aim of this review was to help inform the development of the intervention to be piloted in WP4 (see Chapter 7). In addition to assessing the intervention’s effectiveness and safety, it was important to evaluate whether or not it would represent value for money. The second review, therefore, synthesised existing evidence on economic evaluations of falls prevention interventions in PWD. It was not stipulated in the economic review that the population had to have previously experienced a fall as the methods for evaluating a falls prevention intervention would be the same among those with and without a previous fall. The aim of this review was to identify the most appropriate methods and outcomes for an economic evaluation of the intervention to be developed for PWD following a fall.

Effectiveness review

The aim of this systematic review was to synthesise all existing research evidence evaluating the effectiveness of interventions intended to improve the physical and psychological well-being of PWD who had experienced a fall. The full review has been published separately. 26

Results

Selection of eligible studies

The initial search returned 1071 studies after deduplication (Figure 2). Of these, 991 were excluded by screening titles and abstracts, and the full texts of 80 were assessed for eligibility. Of these, 69 were excluded because they were not RCTs or quasi-experimental studies, they did not include PWD or the PWD did not reside in the community. A total of 11 studies remained for narrative synthesis, but four had missing data that could not be obtained. Seven studies were included in the narrative synthesis.

FIGURE 2.

Effectiveness review PRISMA flow diagram of study inclusion and exclusion. Adapted from Robalino et al. 26 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by/4.0/. The figure includes minor additions and formatting changes to the original.

Adapted from Robalino et al. 26 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by/4.0/. The figure includes minor additions and formatting changes to the original.

Characteristics of included studies

Six RCTs and one quasi-experimental study were included in this review. 34–40 In the quasi-experimental study, patients were recruited in two phases. In the first phase, all consecutive patients from two sites were recruited to the control group; in the second phase, all were recruited to the intervention group. The trials recruited mostly patients with hip fractures in hospitals and emergency departments (EDs) in high-income countries. All studies included both cognitively intact patients and patients diagnosed with dementia, except one that recruited patients with at least mild dementia (MMSE score of < 24 points). 34

Five studies evaluated multidisciplinary in-hospital post-surgical geriatric assessment, which varied in terms of the type of ward (e.g. geriatric vs. orthopaedic), mix of multidisciplinary staff and components of the intervention. 35,37,40 All studies in this group included a core team of a geriatrician, nurse, occupational therapist (OT) and physiotherapist, plus other staff, such as social workers or dietitians, as required. The interventions included different combinations of components, for example early-discharge planning, post-discharge home visits and weekly team meetings.

One study evaluated multifactorial assessment and intervention in patients presenting at an ED post fall, utilised a multidisciplinary team (MDT) similar to that used in the in-hospital geriatric assessment and followed up with risk assessments in patients’ homes. 34 Patients were then offered a variety of interventions based on the risk assessments; these interventions included home-based exercise, home hazard modification, medication review and optical correction by an optician.

The final study provided an annual dose of intravenous zoledronic acid to participants in an attempt to reduce recurrent falls and further fractures by improving bone health. 36

Critical appraisal of studies

The included studies were mostly at low risk of selection bias, and many were at high risk of performance and detection biases due to difficulties in blinding participants and/or personnel to interventions and outcomes (a common scenario with complex interventions). The risk of bias for attrition and reporting was less well reported.

Outcomes

Three RCTs34,37,38 and one quasi-experimental study40 reported different measures of mobility following the intervention, of which three studies reported limited improvement or retention of mobility in the intervention group compared with the control group. Those studies utilising multidisciplinary in-hospital postsurgical geriatric assessment reported short-term improvements in gait, but long-term improvements were either not reported or proved statistically insignificant. 37,38,40 The studies used different mobility scales that exhibited relatively little overlap in the components measured.

Three studies reported recurrent falls post intervention,34,36,37 of which one37 reported a reduction in inpatient falls in the treatment group (4%) compared with the control group (31%) (p = 0.006), although there was no difference in the rate of new fractures. A second study34 reported no difference in the number of patients with falls, the median number of falls or the median number of weeks before the second fall. The final study36 found no difference in falls for PWD, but reported a reduction in recurrent fractures at 6 months in the cognitively impaired patients.

Three studies reported on post-intervention ADL,37–39 utilising four tools that had limited overlap, with only two common items (feeding and transferring). The results were not consistent between studies.

Three studies measured length of hospital stay using multidisciplinary in-hospital post-surgical geriatric assessment, but with varying components. 35,38,39 Two studies35,39 showed a decreased length of stay for those with mild or moderate (but not severe) dementia, and the other study38 reported a significantly higher median length of stay in the intervention group.

All of the studies utilising multidisciplinary in-hospital post-surgical geriatric assessment and intervention reported on the place of discharge. 35,37–40 Three reported that PWD were more likely to return to independent living following the intervention,35,39,40 and the other two described no difference between the intervention and control groups. 37,38

Two studies reported no evidence of impact on readmission rates to hospital. 34,38

Review of approaches to economic evaluation

The aim of this review was to understand the current cost-effectiveness evidence base in the area to inform the design of a potential future economic evaluation of the DIFRID (Developing an Intervention for Fall related Injuries in Dementia) intervention should it proceed to a definitive trial. The review identified economic evaluations of fall prevention interventions in PWD to make recommendations about:

• how best to capture the resources used to provide the intervention and any changes in subsequent use of services

• appropriate outcomes that would (1) capture the benefits of the intervention, (2) be appropriate for use with PWD and (3) provide information for an economic evaluation relevant to policy-makers.

Drawing on the findings of the review, we developed and piloted data collection tools to collect information on health-care resource use in WP2 (see Chapter 3) and WP4 (see Chapter 10).

Results

Selection of eligible studies

The initial search returned 1252 reports. Eleven papers were excluded after deduplication and 124 were excluded as they were not in English. A further report was identified from an ineligible report that was a literature review concerning fall interventions. 50 Overall, six reports were deemed potentially relevant and the full papers were obtained.

Four reports were excluded after the full texts were reviewed. One paper was a critical review of another paper that was selected for full review. 51 One paper estimated the cost-effectiveness of falls prevention of a range of interventions using the results of a systematic review to populate a Markov model. 50 The population included in that review were people aged ≥ 65 years, but their exclusion criteria included ‘special populations (e.g. stroke or osteoporosis)’; therefore, we cannot assume that people with cognitive impairment were included in the model. 50 Two of the excluded papers, one protocol and one economic evaluation based on a RCT, evaluated exercise-based interventions in PWD in nursing homes [the LEDEN (Effects of a long-term exercise programme on functional ability in people with dementia living in nursing homes) study52] and patients with Alzheimer’s disease (AD) [FINALEX (FINnish ALzheimer disease EXercise trial)53]. The aims of these studies were to improve functional ability (LEDEN) and to improve physical functioning and mobility (FINALEX). Overall, although both studies collected falls as a secondary outcome, it was not the primary objective of their intervention and neither incorporated number of falls as an outcome measure in their economic evaluations. The LEDEN study protocol outlined the outcome measures for the economic evaluation as costs and changes in functional ability measured using the Alzheimer Disease Cooperative Study ADL-severe. 54 FINALEX estimated the ICER as the cost per dyad (patient with AD and their carer who was a spouse they resided with). 53

Included studies

Two papers met the inclusion criteria, both of which were protocols. 55,56 The i-FOCIS RCT aims to examine whether or not an individually tailored, cognitive impairment-specific approach to the delivery of an exercise and home-hazard-reduction programme can reduce the rate of falls in community-dwelling cognitively impaired older people. 56 The Encouraging Best Practice in Residential Aged Care (EBPRAC) programme aims to improve evidence-based clinical care for residents in aged-care homes and to enable nationally consistent application of this care. 55 Both proposed studies will be undertaken in Australia.

A PRISMA breakdown of study inclusion and exclusion is presented in Figure 3.

FIGURE 3.

Economic evaluation review PRISMA flow diagram of study inclusion and exclusion. CI, cognitive impairment.

Participant and study characteristics

The studies have different target populations. One study targets people aged ≥ 65 years living in the community with cognitive impairment (n = 360)56 and the other targets residents of residential aged-care facilities, including PWD (nine residential aged-care facilities, 670 patients and 650 staff will be invited to participate). 55

Perspective of the studies

The protocols55,56 suggest that the economic evaluations will be conducted from the perspective of (1) the health and community service provider (i-FOCIS) and (2) the societal and residential aged-care facility (EBPRAC).

Resource use data (costs)

The i-FOCIS RCT will capture information on the consumable, reusable and capital resources required to deliver the interventions as part of the trial. Data on health-care resources, including medications, will be captured via self-reported monthly calendars for 12 months. Costs will be collected from routine sources and out-of-pocket expenses for the patients and their carers will be estimated from a previous published study. 57

The EBPRAC programme will determine the resources needed to deliver the intervention by monitoring the resource use during the project implementation for 12 months. These resources will be costed using market values where possible. Fall-related health-care resource use will be collected from two participating sites. There is no information provided on where these costs will be collected from.

Outcome measures

The rate of falls will be the primary outcome for both of these studies, and both economic evaluations will incorporate this into their analysis. The primary economic outcomes are the cost per fall prevented (i-FOCIS56) and the cost per fall (EBPRAC55). The i-FOCIS RCT incorporated additional outcome measures in the economic evaluation: falls requiring medical attention, ED presentation avoided, hospital admission avoided and QALYs estimated using the EuroQol-5 Dimensions, five-level version (EQ-5D-5L).

Economic evaluation

The i-FOCIS RCT will analyse its data as a within-trial cost-effectiveness analysis. Appropriate sensitivity analyses, including deterministic and stochastic analyses, will be used to address any uncertainty in costs, effects and cost-effectiveness. The probability of the intervention being considered cost-effective at current willingness-to-pay thresholds will be presented as cost-effectiveness acceptability curves (CEACs).

The EBPRAC programme will determine whether or not there is a reduction in the cost per fall associated with the intervention by analysing the cost per fall pre and post intervention implementation. Cost per fall will be estimated by modelling the costs collected from two participating sites. Sensitivity analyses will be carried out to address any uncertainty surrounding costs and effects.

Duration of the studies and data-collection time points

The i-FOCIS RCT has a 12-month follow-up, with clinical and quality-of-life outcomes collected at baseline, 6 months and 12 months. The number of falls and health-care resource use are collected using monthly diaries. The EBPRAC programme will be a 2-year study; it also includes a review of the literature, hence it is unclear when the intervention will be implemented. However, although the study team will review falls data every 6 months, it is unclear when other outcome measures will be collected. An economic model will be used to estimate costs and effects 1, 2, 3, 4 and 5 years post intervention implementation.

Quality of the studies

As the papers suitable for inclusion were protocols, there is insufficient detail on the methodology of the economic evaluations for us to evaluate the quality of the proposed analyses using standard criteria. 49 The i-FOCIS RCT is not accounting for any longer-term costs and benefits that may be accrued after the intervention is implemented. If the intervention is effective, this could create potential bias if the follow-up is not sufficiently long for benefits and possible cost savings in subsequent care to offset the cost of the intervention. The i-FOCIS RCT is not considering the potential impact of the intervention on carers, despite their involvement in the home exercise programme by supervising practice sessions and assisting in delivering the sessions at home.

The duration of the EBPRAC programme is unclear, but if costs and outcomes are going to be estimated beyond a 1-year time frame in an economic model then discounting needs to be considered. There was no detail provided on the type of economic evaluation model being undertaken. Therefore, it is unclear if the approach provided will be sufficient to capture costs and benefits in the longer term.

Introduction

The reviews reported in Chapter 2 indicate a lack of evidence regarding the effectiveness of fall interventions for PWD and limited attention to the economic evaluation of such interventions. PWD who sustain fall-related injuries may currently receive a range of health interventions, but the current models of care for this specific situation have not previously been described and the potential demand for such an intervention is not known. In order to develop a new complex intervention for this situation, we wished to describe current usual care and assess the demand for a future intervention to ascertain the feasibility of recruitment in WP4. We planned to do this by measuring the incidence of fall-related injuries presenting via three settings: the ED, paramedic attendances and primary care consultations.

In a subgroup of people presenting with fall-related injuries, we piloted a data collection tool in the form of a diary to collect data about falls, help at home and usual care. Information on usual care was obtained by analysing the health-care services used by diary participants after a fall-related injury. We planned that this diary would also be used to refine the design of a data collection tool used in WP4 to meet the requirement to identify subsequent health and social care resource use and capture data on the number of falls (see Chapter 2).

In order to obtain further information about experiences of usual care, some participants in the diary study also took part in a qualitative interview (described in Chapter 4). This chapter describes the incidence of fall-related injuries and the findings of the diary study.

Aim

The aim of the diary study was to determine the feasibility of recruiting PWD through different settings and to pilot the data collection tool prior to the feasibility study in WP4.

Methods

Results

Diary study

Discussion

The incidence of fall-related injuries was much lower than expected. In our previous study, the total incidence of falls in people with mild to moderate dementia was 9118 per 1000 person-years. 2 A secondary analysis of the data in our previous study showed that there were 0.044 injuries per person per month among PWD with AD or VAD. Our findings that only 0.0029 cases per person per month presented to the ED and 0.0035 cases per person per month presented to primary care suggest that fewer than 15% of injuries sustained by PWD are brought to the attention of health-care practitioners. This assumes that our procedures for identifying both fall-related injuries and that a person with a fall-related injury had dementia were robust. In primary care, we are confident that the diagnosis of dementia was robust because the GPs had direct access to QOF registers. In the ED, where carers gave a history of dementia, this is likely to have been accurate. However, a carer may have been absent or the carer may not have been aware of the diagnosis. In the case of diagnosis by paramedics, the very small number of cases suggest that systems for picking up dementia were not robust. This has implications for WP4 and any future trial of an intervention to improve outcomes for PWD with fall-related injuries. If dementia is not identified at the time of presentation, it would be difficult to ensure that a referral to an appropriate intervention is made.

The limited number of diary data collected indicated that most PWD received very little health-care input following the index fall. Although 8 out of the 13 participants were offered an initial referral to a falls unit, physiotherapy or a rehabilitation unit, the diaries of the 12 participants who completed them show very little use of these services over the course of the subsequent 12 weeks. This suggests that people may have received an initial assessment but were deemed not appropriate for further input or were still waiting for the referral to be followed up. Further evidence of this was found in our qualitative study, which is discussed in Chapter 4.

Introduction

The effectiveness review provided limited guidance on the core components of an intervention to improve outcomes for PWD with fall-related injuries and highlighted uncertainty over the most appropriate outcome measures. Although the diary study provided some insight into current service use in the 12-week period following an injurious fall, low recruitment rates meant that we were unlikely to have captured the diversity of care pathways. Qualitative work was conducted with a range of stakeholders to provide additional data.

Aim

The qualitative component of WP2 aimed to develop a better understanding of current pathways and identify opportunities for intervention. Objectives were to:

• explore the range of services currently available to PWD following an injurious fall

• identify the needs of PWD and carers following an injurious fall and ascertain the extent to which these were currently met

• explore ideas for service development or intervention with a range of stakeholders

• identify outcomes of importance to PWD and their families.

We included health and social care professionals, PWD who had experienced a fall and their carers. We also aimed to identify the care needs of, and outcomes of importance to, PWD and their families.

Methods

Sampling

Participants were drawn from the three sites described in Chapter 3. Professionals were initially identified through interviews with the principal investigator of each site. We then used snowball sampling75 to identify relevant health and social care services in each area. Recruitment continued until data saturation was reached.

Professionals taking part in interviews and focus groups were asked whether or not they and their colleagues would be willing for us to observe routine practice. We selected a diverse range of services across the three sites for observation. All patients due to be seen on the agreed date(s) for observation were eligible to take part.

Patients and carers were recruited for interviews through either observation or the diary study (see Chapter 3). The process of sampling and recruitment is summarised in Figure 4.

FIGURE 4.

Summary of sampling and recruitment processes. Reproduced with permission from Bamford et al. 73 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by/4.0/. The figure includes minor additions and formatting changes to the original.

Reproduced with permission from Bamford et al. 73 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by/4.0/. The figure includes minor additions and formatting changes to the original.

Results

Qualitative findings relating to care pathways for PWD following a fall and the need for staff training have been published73,74 and are summarised below. Quotations are accompanied by a participant identifier, with role and service type given for professionals.

Proposed content and delivery of a new intervention

Through discussion of the shortcomings of existing services, and examples of good practice, participants identified key components of a new intervention. These components were grouped into three main themes, each of which included subthemes. An overview of the themes, and their implications for a new intervention, is provided in Table 5. We then provide a description of each theme, with illustrative data.

TABLE 5 - Key components of a new intervention

Theme	Subthemes	Implications for a new intervention
1. Supportive service organisation	Clear pathways	Compile and disseminate information on local care pathways Improve referrals to specialists
	Flexible service delivery	Provide more flexibility in the duration and frequency of intervention delivery Introduce proactive maintenance/follow-up
	Effective information-sharing	Involve a MDT to improve access to information
	Holistic approach	Include holistic assessment and ensure that all factors contributing to falls are addressed
2. Staff attitudes, knowledge and skills	Attitudes to dementia	Provide staff training on attitudes to dementia
	Knowledge and understanding of dementia	Provide staff training on dementia
	Practical skills for working with PWD	Provide staff training on a range of practical skills including communication, engaging PWD and embedding activities into daily routines
3. Supporting carers and their role in interventions	Assessing carer capacity and stress	Assess and address carer stress levels Explore carer capacity and willingness to take an active role
	Training and educating carers	Assess carer needs for psychosocial education Identify local services to meet carer needs

Theme 1: supportive service organisation

Interviews and observation highlighted a number of ways in which the current organisation of services could undermine or enhance the recovery of PWD following a fall. 74

Clear pathways

A recurrent theme in interviews with professionals and carers was the fragmentation of care pathways for PWD following a fall. Because local falls and dementia services were often led by different individuals, knowledge of the full range of services available and how to access them was often patchy.

Drawing on our data, we identified 21 distinct types of service potentially available to PWD following a fall. Not all service types were available in all areas but Figure 5 shows a composite set of services organised by the point in the falls trajectory at which they are typically involved. Initial services (hyperacute and acute phases) primarily focus on injury management, whereas services provided later in the trajectory (post acute and longer term) focus on rehabilitation, risk management and the prevention of future falls. Figure 5 also indicates which agencies typically provide the service and whether it is delivered in the patient’s home or elsewhere.

FIGURE 5.

Overview of services potentially available to PWD with fall-related injuries. Adapted from Wheatley et al. 74 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/. The figure includes minor additions and formatting changes to the original.

Adapted from Wheatley et al. 74 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/. The figure includes minor additions and formatting changes to the original.

An individual person with dementia with a fall-related injury will access only a limited range of services. Although the pathway might appear linear, it is possible for many of these services to be accessed in parallel and people with recurrent falls may have multiple iterations of different service configurations. Although it is difficult to identify a ‘typical’ pathway, we have provided an example of a possible care pathway for a person with dementia who fell at home and sustained a fractured neck of femur (Figure 6).

FIGURE 6.

Example pathway for a person with dementia with a fractured neck of femur. Adapted from Wheatley et al. 74 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by/4.0/. The figure includes minor additions and formatting changes to the original.

Adapted from Wheatley et al. 74 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/. The figure includes minor additions and formatting changes to the original.

A lack of knowledge of locally available services could result in PWD not receiving appropriate support; this was particularly problematic for paramedics who covered large geographical areas. Professionals emphasised the need for improved integration of existing services, for example through a single point of contact, better signposting, clear eligibility criteria and referral systems:

A standardisation of systems would be helpful. So you know, if that patient has dementia and they fall and they injure themselves, then you can ring a number to get them referred, if it’s out of hours, to a particular team who will come out in a couple of days and check that.

Professional 13, paramedic (interview)

We found one example of successful service integration among first response services. One telecare service was integrated with the ambulance and NHS 111 services, which allowed for flexibility and could reduce response times:

They are only supposed to attend non-injurious falls, so if they get there and there are signs of injury, they call the paramedics (and sometimes the paramedics refer patients to them if there is a non-injurious fall).

Field notes of urgent unplanned needs service visit, 20160728

Key services on the care pathway were thought unsuitable for PWD. The ED environment in particular was identified as problematic for PWD, despite attempts to make some EDs more dementia friendly (e.g. use of volunteers to give non-clinical attention or distraction techniques). Concerns about care provided outside the PWD’ home environment were raised, because this could lead to loss of confidence and connection with their home if admission was prolonged:

When people come in here, the quicker you can get them home the better, because the longer they’re out of their own environment the harder it is, they become very dependent here. We see it quite often where people do go home and it fails, because they’ve been here too long.

Professional 30 (joint interview with Professional 30 and Professional 31, managers, reablement service and specialist dementia unit)

Access to community support could influence length of admission. If equipment or packages of care were not available, PWD sometimes remained in rehabilitation services longer than intended or were transferred to an alternative service. A community matron model, with a dedicated person to help navigate services, was suggested to facilitate prompt discharge. This would be valued by carers, who often lacked knowledge of available services or how to access support:

[The patient’s] granddaughter says that someone said that he should have had a quad stick but that hasn’t been sorted. She doesn’t really know who is supposed to deal with that or where they’re supposed to get it from.

Field notes of domiciliary visit by physiotherapist from reablement service, 20160714

Flexible service delivery

Many professionals highlighted the poor ‘fit’ between the needs of PWD and current service organisation. The emphasis on shifts between health and social care and time-limited interventions disrupted continuity of care. Staff emphasised the need for time to enable PWD to develop rapport and trust:

When they know they’re having problems with their memory, if they know they’re not managing but they’ve got to trust you before they’re going to give you the information that you need as to what they’re struggling with. You’re not going to get that from one half an hour visit. You’ve got to build up the picture over a period of time.

Professional 62, team lead, facilitated discharge service (interview)

Lack of continuity was also challenging for professionals who relied on detailed knowledge of PWD to tailor interventions appropriately. Time-limited interventions were thought not to take account of the need for PWD to have more rehabilitation sessions, over a longer period of time, in order to achieve the same progress as older people who were cognitively intact:

You shouldn’t be able to rush people for multiple reasons. They could be afraid, they could be in pain, they could be constipated, getting down to the root of it, finding out what the triggers are. [. . .] You’ve got secondary things due to vascular changes, from strokes and things. How tired they may be. If somebody gets very tired, you can’t do a long session. You have to do lots of little sessions.

Professional 48, physiotherapist, ward-based therapy (interview)

In the context of a deteriorating condition, ongoing review was recommended by some professionals to avoid the loss of functional gains and address any emerging problems or changes:

You will only maintain your muscle strength and range of movement for a period of time before it starts to decline again if you are not doing those activities.

Professional 10, senior OT, ward-based therapy (interview)

No services providing long-term review or follow-up were identified. Carers would also have welcomed more follow-up, to address changing needs and ensure that existing equipment was still appropriate:

Dad’s got a pressure pad under his chair there and a pressure pad under his bed but he no longer has pressure sores. We’re still using them, but does he actually need them anymore? I don’t know.

Carer 8 (joint interview with Patient 8 and Carer 8)

Effective information-sharing

A need for effective systems for sharing information across service boundaries (e.g. primary and secondary care, health and social care) was identified. Joint information systems remained underdeveloped, with many services not being able to access even basic information (e.g. whether or not a formal diagnosis of dementia had been made). A lack of joint information systems led to unnecessary duplication of effort and frustration for staff:

[Professional 120, physiotherapist] comments that it would make a lot more sense if all the information was consolidated in one place for everyone to have access to. [. . .] She takes the patient’s notes to her office to fill in the referral paperwork for [short-stay rehabilitation unit], which is a 7 page document. She says that she doesn’t know why the referral is so long and complicated, since the [short-stay rehabilitation unit] doesn’t trust their data and will redo all the tests when the patient gets there [. . .] It takes her 45 minutes to complete and fax the referral.

Field notes of facilitated discharge service, 20161018

Multidisciplinary team meetings, including staff with access to different information systems, were identified as a successful way of sharing information.

Carers similarly commented on the lack of ‘joined-up working’ between different services, which could lead to conflicting advice or interventions:

[Home carers] brought a load of occupational therapists in to change some of the equipment over because they thought it wasn’t suitable, even though the hospital had provided what they thought was right. [. . .] although, that [a hoist] is very useful, the stand assist they brought originally meant Dad had to work at it, which exercised his arm muscles and his leg muscles and that’s what I would have preferred to have seen staying here, although, it was more work for us to move around and so forth. It’s easier for the carers to use that [hoist]. It’s harder for them to use the other one.

Carer 8 (joint interview with Patient 8 and Carer 8)

This is one of several examples of carers expressing their uncertainty over equipment and aids. Where there was no explicit discussion of the rationale for changing equipment, or detailed guidance on how to determine when aids were no longer needed, carers were unsure what to do for the best. In the above example, the change of equipment creates concern over whether or not the new equipment was chosen for staff convenience rather than to maximise recovery.

A holistic approach

Holistic assessment was seen as the most appropriate way of identifying and treating factors that could potentially contribute to falls. Opportunities for such assessment, however, seemed limited. Falls clinics and day hospitals routinely adopted a holistic approach, but not all PWD were referred to these services. The extent to which holistic assessment was provided in other settings was unclear. Professionals recommended that assessment of PWD with falls should include visual perception and spatial judgement, footwear and foot care, bone health and hydration. Addressing mood and anxiety was also important to rehabilitation outcomes:

I think in terms of addressing depression as well, people might have a reduced level of activity if they’re depressed. Obviously, that’s not great. If someone’s sitting for a long time, that’s going to affect their mobility. If people aren’t feeling particularly motivated, it does affect what they’re doing.

Professional 71, senior worker (focus group with specialist inpatient rehabilitation workers)

Formal tools were used most frequently to assess cognition and fear of falling. The potential value of detailed assessment in facilitating staff to tailor their approach to meet the needs of an individual person with dementia is illustrated by the following description:

We do specialist assessments in terms of what cognitive level they’re at, so that will tell us how we should be giving instructions to them, what their abilities still are. Even things like how we should be modifying the environment to best meet their needs. You can actually kind of predict what the risk is going to be in relation to falls, based on their personal cognitive level. So it’s using the strengths that they’ve still got to get the best that we can out of it.

Professional 79, OT clinical lead (focus group with mental health professionals)

Less systematic approaches were typically used in assessing other areas. Despite frequent discussion of the difficulties of assessing pain in PWD, only two paramedics described using a formal pain assessment tool (the Abbey Pain Scale77). Given the difficulties in identifying pain in some PWD, some professionals suggested that pain relief should be given routinely because it potentially:

[. . .] doesn’t do any harm and actually usually [we] can find some quite marked effects.

Professional 81, nurse consultant (focus group with mental health professionals)

The observation highlighted several occasions when pain relief was inadequate. In the following example, a person with dementia became irritated during an examination in the ED:

The doctor then helped the patient move onto the bed so that she could examine the knee fully. It was clear that the examination was painful at some points and the patient became quite cross saying ‘She’s the wickedest woman in [place]’ [. . .] The doctor took this in good part but I don’t think she really explained to the patient why she was having to move the leg and cause her pain. The patient had clearly had enough, saying ‘Well, just stop it!’.

Field notes of ED, 20160406

There may have been a number of reasons for the patient becoming un-co-operative at this point: she had not received pain relief for more than 4 hours and was potentially hungry, thirsty and tired, as she had missed lunch and her usual nap.

Theme 2: staff attitudes, knowledge and skills

The second main theme influencing the quality of existing falls services for PWD related to staff attitudes, knowledge and skills. 73 A range of approaches to staff training and development were identified, including basic training for all staff, access to specialist advice through MDT meetings, links with specialist dementia services (e.g. challenging behaviour teams) or (where available) links with the local lead for falls and dementia, and joint visits with experienced colleagues or local specialists.

Attitudes to dementia

Recognising and facilitating the rehabilitation potential of PWD require staff to be willing to invest in working with this patient group. Comments made during interviews and observation of service delivery indicated negative attitudes towards PWD among some staff. The ability of PWD, particularly those with more severe cognitive impairment, to benefit from an intervention following an injurious fall was questioned by some staff who perceived them as ‘not rehabable’ or ‘untrainable’:

And unfortunately the term ‘not rehabable’ is not uncommon, used in relation to patients presenting with cognitive issues, which could be delirium as well as patients with diagnosed or undiagnosed dementia. And is used by all professionals and grades of staff, from consultant geriatricians down.

Professional 79, OT clinical lead (focus group with mental health professionals)

Assumptions that falls were an inevitable consequence of dementia meant that detailed investigation following a fall was not always seen as relevant or necessary, because the problems were simply attributed to dementia:

I think partly because I would be thinking that the dementia or the lack of awareness is partly behind why they’re actually falling and hurting themselves, rather than falling and not falling. I’ve already identified a pathology in that group, i.e. that they’re demented, where the non-demented ones, I’m looking for another pathology.

Professional 33, GP, primary care (interview)

Negative attitudes to dementia were also evident in the extent to which staff were willing to persevere with PWD and their families who appeared reluctant to engage or were not adhering to proposed activities. Although some staff sought to address these issues by building rapport and adapting their approach, others seemed to accept reluctance to engage at face value and simply withdrew the service.

Knowledge and understanding of dementia

Many participating professionals had received no formal training in dementia. Some had attended brief courses or completed online modules, and others had learned through professional experience, from colleagues or through personal experience with dementia:

I’ve had people in the family with dementia, and I’ve done a lot of looking into how to talk to people, how to deal with people [. . .] I do think it has benefited my work, and I know when my grandma was ill a few years ago, I did a lot of research on just little things around how to talk to people, what might be missing from their lives, that sort of thing.

Professional 23, exercise class instructor, non-statutory sector (interview)

Many professionals expressed a desire for further training in working with PWD. Staff working for a mental health trust particularly highlighted the need for professionals to better understand ‘challenging behaviour’ as a barrier to rehabilitation.

Practical skills in working with people with dementia

In addition to greater understanding of dementia and the potential impacts on individual PWD, staff required a range of skills to work effectively with this patient group. These have been described elsewhere73 and are briefly summarised in Table 6. Specific skills identified by participants and through the observation included communication, use of observation, engaging PWD, tailoring interventions to build on existing interests and activities and embedding interventions into daily life.

TABLE 6 - Skills required for effective intervention with PWD

Skill	Example
Communication	[Professional 111] does tests on the arm that is damaged from the stroke, including range of movement, grip etc. There’s a misunderstanding about pain at this point. She says to [the patient], ‘Let me know if it’s sore’, but I’m not sure he understands. She just kind of says it and then goes straight in, without giving him time to process. It seems clear that he is in pain, to the point where his granddaughter says, ‘Oh, granddad, say if it hurts’. Then [Professional 111] stops and rephrases the thing about the pain, and after that he is quite good about saying where the pain isField notes of domiciliary visit by physiotherapist from reablement service, 20160714
Observation and demonstration	Because a lot of our assessment is skilled observation. I think it’s quite specific to this specialism, because you can’t always rely on the person’s account to be accurate, and we’ve got lots of patients, especially in the early stages, that have got really good verbal skills and they can mask the level of impairment that they’ve actually got. But with an OT assessment they can’t mask functioning. So we will get them to show us everythingProfessional 79, OT clinical lead (focus group with mental health professionals)
Engaging PWD	[Professional 51] says ‘I wanted to test you, [Patient 1]’. She gets a large ball from the cupboard, about the size of a beach ball. [Patient 1] becomes very animated, much more engaged than he has been. He says ‘Oh, football.’ [. . .] They stand in the middle of the gym and do some bouncing and catching of the ball, and then throwing and catching. [Professional 51] checks after each activity if he has experienced any dizzinessShe then fetches cones and a plastic football from the cupboard, which [Patient 1] greets enthusiastically ‘That’s the way forward!’Field notes of day hospital physiotherapy session, 20160504
Tailoring interventions	[Patient 1] has new trainers – last time he came, [Professional 51] had recommended that he get new shoes because his old ones had holes in. [. . .] [After the consultation] [Professional 51] mentions the shoes and says that someone other than the patient’s wife had to say it: [Patient 1] loved his old shoes even though they had holes. [Professional 51] wrote down the instruction to get new shoes so that his wife could have something to refer toField notes of day hospital physiotherapy session, 20160504
Embedding interventions	There is quite a bit of evidence now if you incorporate your strength and balance exercises into your routines it can be as effective. For me that is where we need to work on with the dementia patients. If you can put it into their habits so it becomes habitual then you perhaps don’t need to be going to a programme that you have to do specific things at specific timesProfessional 10, senior OT, ward-based therapy (interview)

Some of the quotations in Table 6 demonstrate the creativity and resourcefulness of staff in trying to engage with PWD and achieve desired outcomes. Without these practical skills, staff could struggle to engage PWD, which could reinforce the perception of some staff that PWD were unsuitable for intervention.

Many professionals recognised that PWD could have difficulty in retaining abstract and unfamiliar exercises. Making interventions understandable and appealing to the individual was therefore seen as essential. Some participants, particularly patients, talked about the importance of the intervention being fun and enjoyable. To enable activities to be tailored in this way, assessment needed to capture information about the person with dementia’s personality, likes and dislikes.

Theme 3: supporting carers and their role in interventions

Carers were recognised as potentially having a key role in facilitating rehabilitation of PWD; however, there were tensions regarding their capacity to support interventions and attitudes to positive risk.

Assessing carer capacity and stress

Because interventions were often time limited, some professionals recommended a cascade model in which relevant exercises or activities would be integrated into regular visits by a support worker or into daily routines by a carer:

I worked with somebody who had really, really bad dementia and who’d had lots of falls, and we thought he wouldn’t work with a therapist because they were scary. He didn’t understand what we were offering, so we taught his daughter all of his exercises and did all the risk assessments and that with her. Then, she did them with Dad and we just reviewed once a week, and he did really, really well.

Professional 08, manager, specialist inpatient rehabilitation (interview)

Although this is a positive example of involving a carer in rehabilitation, carers did not always have the capacity to support interventions in this way. Some staff implicitly assumed that carers would take an active role in any intervention, for example by encouraging activities between sessions and after an intervention was withdrawn. Staff understanding and recognition of carer stress varied. Some staff seemed to lack insight into the realities of caring for a person with dementia and occasionally expressed critical and even hostile views towards carers. For example, some exercise instructors expressed frustration that carers saw their service as an opportunity to have a break rather than attending the class with the person with dementia. Other staff had more insight into the challenges and stress of caring for PWD:

It is a huge demand on families as well, which I think people forget sometimes how stressful it is for the family to have someone at home where things are not working out and where they are worrying about them. They do know that they really don’t want them to go into hospital, but they are really at the end of their tether providing the support. That is something that we see very frequently here.

Professional 06, consultant geriatrician, outpatients falls service (interview)

In addition to stress, the capacity of carers to support interventions was influenced by their own health problems, work commitments and whether or not they lived with or near the person with dementia.

Training and education for carers

A number of professionals highlighted the potential for carers to undermine interventions because of a lack of understanding. Some carers were perceived as ‘overprotective’ and risk averse by both PWD and professionals. Typically, such carers tried to restrict patient mobility or daily activities because they were unfamiliar with the concept of positive risk or the dangers of deconditioning. In this context, professionals described the need for ‘mindset work’ to encourage carers to actively encourage the PWD to get up and move about. One rehabilitation ward actively involved carers in rehabilitation, for example by supervising patients while they were walking. In this setting, staff also trained carers on how to carry out certain tasks, such as transferring a patient from a bed to a commode. In contrast, other services provided little information on how to support a person with dementia following discharge:

When Dad [Patient 8] was released, I was given about a 5-minute demonstration of handling, how to transfer people to sit down onto a stand or onto the bed and moving them around and that was it. They’re released into your care.

Carer 8 (joint interview with Patient 8 and Carer 8)

Professionals suggested that carers would benefit from additional education and awareness raising on the importance of keeping moving, hydration and nutrition in reducing the risk of falls. Carers and professionals also emphasised the value of practical advice on how best to help the PWD get up after a fall.

Discussion

The findings suggest that improving outcomes after a fall depends on recognising and facilitating the rehabilitation potential of PWD. The integrative analysis of all qualitative data identified three key areas that need to be addressed by the intervention to ensure that this aim is achieved in practice.

The first of these was ensuring that services are organised in the most effective and supportive way for PWD. This included having clear and efficient care pathways for PWD following a fall; evidence from other contexts suggests that this can improve care outcomes. 78,79 Improving links between services, particularly across health and social care boundaries, could also be beneficial, although this remains challenging in the UK. 80 Flexibility in service organisation to allow care to be tailored to the individual needs of PWD was seen as particularly important. This included varying the number, frequency or duration of intervention sessions, as well as tailoring the content of the intervention to individuals. Evidence exists to support such a tailored, person-centred approach to rehabilitation following hip fracture,81 although further research is needed.

The attitudes, knowledge and skills of professionals working with PWD was another key area to be addressed. Developing a positive, rather than fatalistic, attitude towards PWD was also important; again, the presence of negative attitudes was congruent with previous research in this area. 82,83 Few health and social care professionals had received formal dementia training, a finding reported by several previous studies of staff delivering care following hip fracture. 84–86 We identified a need for staff delivering the intervention to have both theoretical and practical understandings of dementia. In particular, practical skills in communication, observation, engaging PWD and tailoring and embedding interventions are required. In addition, staff require creativity and resourcefulness in adapting their approach to suit individual PWD.

The final key area was supporting carers and their role in interventions. This included ensuring that carers had the capacity and willingness to provide any assistance required by the intervention, and finding other solutions if not. Support, training and education for carers was seen as an important factor in improving this capacity.

Introduction

Given the small number of articles included in the effectiveness review, we additionally used a realist approach89 to synthesise the current evidence regarding the management of falls in dementia, to build on the qualitative work reported in Chapter 4 and to develop a theory regarding how a new intervention might work. 90,91 The aims of the realist review were to identify the essential components of an intervention aimed at improving care for PWD with fall-related injuries and to hypothesise the mechanisms underpinning successful falls prevention and/or rehabilitation interventions for PWD.

Methods

Realist methodology involves the development of context–mechanism–outcome configurations (CMOcs), which aim to identify what works (mechanism), for whom and in what circumstances (context) to achieve a particular intervention outcome. 89

We used an iterative approach to the review, integrating data from the semistructured interviews and focus groups with professionals (described in Chapter 4) with literature searches to develop and refine our emerging theory to underpin the intervention. This was a pragmatic decision based on the intervention time scales and the availability of qualitative data, as research approvals were not yet in place for interviews and observation with PWD and carers.

The initial phase involved the development of ‘if–then’ statements from the qualitative data by the qualitative team. 92 We then refined the if–then statements, looking for data that could be interpreted as a context, a mechanism or an outcome. Initial CMOcs were presented to a group of clinician co-applicants working on the realist review and further refined using their feedback. This preliminary CMOc framework formed the basis for extracting data from the literature.

Search strategy

Multiple searches were inductively carried out. An initial comprehensive search (in November 2015) focused on dementia, falls and fall-related injuries, and interventions. Subsequent targeted searches (in March 2017) were conducted for particular areas where the research team felt insufficient information had been found. Additional papers were identified through citation chaining of included papers and relevant systematic reviews and through hand-searches. All searches were restricted to the English language. Figure 7 demonstrates the flow of studies in these multiple, iterative searches.

FIGURE 7.

Diagram of the search, screening, selection and extraction process. Reproduced with permission from Wheatley et al. 88 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/. The figure includes minor additions and formatting changes to the original.

Reproduced with permission from Wheatley et al. 88 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/. The figure includes minor additions and formatting changes to the original.

Search 1: comprehensive search

Electronic searches were conducted in MEDLINE, Cochrane Central Register of Controlled Trials, Health Management Information Consortium, EMBASE, Cumulative Index to Nursing and Allied Health Literature (CINAHL), Web of Science, Allied and Complementary Medicine Database and Physiotherapy Evidence Database (see Appendix 8 for an example of the search strategy). This search was designed around three distinct concepts: dementia, falls and fall-related injuries, and interventions. This was translated to other databases, making appropriate adjustments for individual thesaurus headings and syntax. Citation chaining was used for included papers and relevant systematic reviews to identify additional papers of interest. Trials registers were searched, but further grey literature searching was not conducted. Results from all databases were imported to EndNote.

Search 2: targeted searches

Targeted searches for CMOcs 1, 2, 4 and 6 (see Table 7) took place in MEDLINE and CINAHL on EBSCOhost (see Appendix 9 for an example targeted search strategy). The major subject heading field was used to identify studies that were focused on dementia or older people and relevant CMOcs.

Results

We identified nine CMOcs, which we grouped into three themes (Table 7). 88 Evidence relating to each of these themes is described in the following sections. At the end of this section, we provide a table indicating the intervention components suggested by the CMOcs (see Table 8).

Discussion

The findings of the realist review suggest a number of important components of interventions for fall-related injuries in PWD, as well as potential mechanisms underpinning successful interventions for this patient group. These are represented by the nine CMOcs, which were further grouped into three broad themes: (1) ensuring that the circumstances of rehabilitation are optimised for PWD, (2) compensating for the reduced ability of PWD to self-manage and (3) equipping the workforce with the necessary skills and information to care for this patient group. Drawing on the data relating to each of these themes, we suggest a number of components for inclusion in the final intervention (see Table 8).

Optimising the circumstances of rehabilitation for PWD means ensuring that basic needs, such as appropriate food, water, comfort and pain relief, are met. Given the evidence that suggests lower adherence to recommendations relating to pain relief,103 the intervention needs to consider strategies to engage health-care professionals in managing pain. This might include training for staff on recognising pain and/or ensuring that staff have the necessary permissions to prescribe or administer pain relief. Locating the intervention in a familiar environment such as the home may also facilitate engagement. Home-based rehabilitation may potentially produce good results,108–110 and pre-empting potential problems with advice (e.g. to continue to do exercises while on holiday) could help to increase adherence. If the home is not deemed suitable, specialised rehabilitation units designed for the needs of PWD may be more appropriate than general wards. 108,139 Initial structured, comprehensive assessment emerged as important to ensure that comorbidities were correctly identified and treated. However, the relative lack of emphasis on psychosocial elements of assessment in the studies examined suggests that staff training may be required to stress the importance of holistic assessment and emphasise aspects such as social support networks as part of discharge procedures. Ensuring that MDTs include experts in mental health care as well as physical health could be another potential way to address this issue. Routine assessment of cognitive function may also be necessary to ensure that dementia is correctly identified. 105

Compensating for the reduced ability of PWD to self-manage requires designing rehabilitation strategies around repetition and embedding interventions (e.g. exercises) into existing habits and routines. Indeed, patient memory and understanding may not be necessary. 114 The comprehensive initial assessments recommended in the previous theme (ensuring that the circumstances of rehabilitation are optimised for PWD) should ideally be used to tailor exercises and other treatments to daily life, including flexibility in the timing and duration of the intervention. A staff training component may be required to teach embedding and contextualising techniques. Failure of rehabilitation is a common cause of readmission,116 which the implementation of regular follow-up visits or a longer initial intervention period may help mitigate. Many studies have relied on input from carers to help compensate for the reduced ability to self-manage in PWD,113–115,122 but this may place an additional burden on carers. Therefore, the intervention should involve carers in the decision-making process and include strategies to aid with carer burden and stress, such as support and training for carers. Providing patients and carers with information about falls risks and prevention may promote their self-efficacy, reduce their anxieties and increase their skill in managing falls.

The theme of equipping the workforce with the necessary skills and information to care for PWD draws on elements from across the findings. Although the importance of information gathering for tailoring has already been identified, it may also be valuable to ensure that multiple sources are used when gathering information. Formalising collective memories of PWD and carers104 may help to ensure that care is person centred, for example by using a model similar to the TOP 5 strategy. 107,117 Again, the extra burden placed on carers should be considered. The intervention should provide staff training in dementia, including strategies to manage challenging behaviour. Staff should be challenged on their preconceived attitudes towards older adults and those with dementia to better understand their needs and goals. Training should cover the dementia-specific adaptation to practice described in the previous themes, with a particular focus on communication. In order to create continuity of care, staffing for the intervention should be consistent. Delivering the intervention through multidisciplinary, collaborative teams may also encourage information-sharing. Based on the lack of knowledge of local services identified in the qualitative study, the intervention should also provide education and pathway resources for staff to encourage referral to other services. However, the available evidence106,116,136 suggests that a simple lack of knowledge of services in staff members may not be the only barrier to the use of care pathways. This is a further area in which improving staff attitudes may be of value.

Introduction

The reviews, diary study and qualitative work reported in previous chapters suggested a range of potential components for a new intervention for fall-related injuries in PWD. The diary study (see Chapter 3) additionally provided insight into the practicalities of recruitment to such an intervention. The objectives of the final stage of intervention development were to:

• prioritise and operationalise intervention components

• consider feasible sources for recruitment and appropriate targets

• agree outcome measures to evaluate the clinical effectiveness and cost-effectiveness of the intervention

• validate the proposed intervention through qualitative work with stakeholders

• produce a logic model summarising the rationale underpinning the intervention.

The results of this development process have been published88 and are elaborated below.

Methods

The first three objectives were addressed by convening a consensus panel that participated in two meetings and completed two rounds of a Delphi survey. 140 The proposed intervention was validated through qualitative interviews and focus groups with a range of stakeholders; the logic model was then produced by members of the research team and reviewed by the Trial Oversight Committee (TOC).

Results

Logic model of final intervention

A logic model can be helpful in clarifying causal assumptions of interventions,143 and developing such a model is recommended by UK Medical Research Council (MRC) guidance. 147 We identified the resources required for the intervention, activities to be undertaken, anticipated outcomes and longer-term impacts for each of the three key stakeholder groups involved (professionals, PWD and carers) (Figure 8). The process of identifying the pathways through which we anticipated outcomes would be achieved helped to inform data collection for the process evaluation to be conducted as part of the feasibility study.

FIGURE 8.

Logic model of the DIFRID intervention. This figure is adapted from Wheatley et al. 88 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/. The figure includes minor additions and formatting changes to the original.

This figure is adapted from Wheatley et al. 88 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/. The figure includes minor additions and formatting changes to the original

Overview of the DIFRID intervention

The intervention comprised a tailored programme of activities centred on goals identified and agreed with the person with dementia and their carers. The intervention was delivered face to face to individual PWD and carers by trained rehabilitation staff [OTs, physiotherapists and rehabilitation support workers (variously called support workers, therapy assistants or assistant practitioners in different sites, hereafter referred to as ‘support workers’] at participants’ homes. Rehabilitation staff attended a half-day training programme that covered skills needed to work effectively with PWD, study procedures and all components of the intervention including assessment, goal attainment scaling (GAS) and activity-planning. The assessment included questions about participant likes and dislikes and daily routines to allow staff to develop a tailored programme of meaningful activities to achieve desired goals. The initial assessment comprised two sessions: one conducted by an OT, the other by a physiotherapist. Intervention sessions were mostly delivered by support workers, although there was scope for up to four visits from both the OT and the physiotherapist. Up to 22 intervention sessions in total were available over a 12-week period. The intervention is summarised using the TIDieR framework in Table 9. 94

Discussion

The iterative process of developing the DIFRID intervention involved mixed methods and included a wide range of stakeholders throughout. This process allowed us to integrate practical empirical data from experts and practitioners with evidence from previous studies to create a robust, theoretically informed design for a new intervention.

Furthermore, the consensus panel provided access to a wide range of expertise. This facilitated decision-making where there was no clear direction from earlier WPs (e.g. outcome measures) or where problems had been identified (e.g. recruitment). The content of the Delphi survey was informed by the initial assessment of the desirability and feasibility of the components that emerged from the initial qualitative work and realist review. The Delphi approach allowed us to gauge the extent to which different components of the intervention were supported. 142

Stakeholder feedback on the proposed intervention was generally positive, although some participants found it difficult to comment as the intervention presented was still under development.

The production of CMOcs and the logic model helped us hypothesise about how change would happen and made assumptions explicit. 89 The logic model was also used to inform the process evaluation (see Chapters 8–10).

Introduction

The process of developing the DIFRID intervention has been described in previous chapters. This chapter focuses on the methods and Chapters 8–10 focus on the key findings of a feasibility study of the intervention. The primary aim of the feasibility study was to determine whether or not to progress to a full-scale RCT of the DIFRID intervention to evaluate its efficacy and cost-effectiveness in preventing falls and improving secondary outcomes for PWD who have experienced a fall necessitating health-care attention. The specific objectives were to explore the feasibility and acceptability of:

• recruitment and retention

• implementation of the DIFRID intervention

• proposed outcome measures.

Mixed methods were used and the full protocol for the study has been published elsewhere. 148 A summary of the quantitative methods is provided in the following section, followed by a description of the embedded process evaluation in Process evaluation. The findings for each objective of the feasibility study are then reported in Chapter 8 (recruitment and retention), Chapter 9 (implementation of the DIFRID intervention) and Chapter 10 (outcome measures).

Quantitative methods: recruitment and consent

This study was designed to be was a single-arm feasibility study of the DIFRID intervention.

Quantitative methods: data collection and follow-up

Quantitative analysis

The main analysis was of feasibility outcomes. We report the numbers of eligible participants seen over the recruitment period, and the resulting rates of recruitment, retention and data completion. The majority of the outcome data are presented in descriptive tables presenting percentages, means and SDs.

Process evaluation

The quantitative data were supplemented with qualitative data from the process evaluation, which provided a more nuanced understanding and allowed us to explore if and how the intervention would need to be adapted prior to a RCT.

Criteria for progression to full trial

Stop/Go criteria were developed for progression to a definitive trial (Table 12).

In addition to the quantitative indicators, we included two indicators based primarily on the qualitative work. The first related to whether or not the intervention could be delivered with fidelity (i.e. the content, frequency, duration and quality of the intervention were delivered as set out in the intervention delivery manual). The second was an indication that the intervention was perceived as acceptable to both participants and professionals.

Intermediate outcomes were defined as amber. Whether or not to progress to a full trial was discussed by the TOC.

Ethics approval

Approval was given by Newcastle and North Tyneside 2 Research Ethics Committee (reference 17/NE/0297) and the Health Research Authority.

Introduction

This chapter describes recruitment and retention in the feasibility study and presents qualitative feedback on recruitment processes. Seven methodological issues relating to recruitment and retention for feasibility studies have been suggested,161,162 although not all of these were applicable to the present study. In this chapter, we consider the following issues:

• What factors influenced eligibility and what proportion of those approached were eligible?

• Was recruitment successful?

• Did eligible participants consent?

• Was retention in the study good?

Screening and recruitment of people with dementia

The flow of participants through the study and reasons for non-recruitment are shown in the Consolidated Standards of Reporting Trials (CONSORT) flow diagram (Figure 9).

FIGURE 9.

The CONSORT flow diagram.

A total of 113 people were screened for eligibility, of whom 29 (26%) were eligible. The most common reason for non-eligibility was not having a fall in the previous month (this mainly applied to people who were contacted via research registers). Other reasons included being unable to contact the person, the person being too unwell to participate, the person was moving into a care home, the person had died, the person lived too far away to receive the intervention, no carer was available and the person was not on the dementia QOF register. Some PWD on research registers declined before answering eligibility questions. The setting in which potential participants were screened is given in Table 13.

The majority of the potential participants were screened in the ED (54%). The exception was in Norwich, where screening was hampered because local approval processes did not allow the CTA to access patient ED records directly; instead we had to rely on a clinician approaching patients about the study while they were in the ED. This was a permanent decision that could not be resolved. As a result, it is likely that a number of potential participants were missed in the Norwich ED.

We do not have data about the potential participants in primary care or paramedic attendance who were given a summary PIS but did not return the opt-in form. Two potential participants returned forms from primary care and two returned forms from paramedic attendances. In Newcastle, 13 potential primary care participants received an opt-in form, of whom one returned the form and was screened. In North Tees, 29 potential primary care participants received a summary, of whom one returned the opt-in form, but we know that, owing to a misunderstanding, the practice sent the PIS to all PWD who had ever fallen rather than those who had fallen in the previous month. In Norwich, three potential primary care participants received the PIS, none of whom returned the opt-in form. In the East of England Ambulance Service, four summary PISs were given out and two opt-in forms were returned. We contacted the North East Ambulance Service to find out how many summary PISs were given out in Newcastle and North Tees but did not receive a reply. No opt-in forms were received via the paramedics in Newcastle and North Tees, suggesting that forms may not have been given out. Despite initial agreement to participate from telecare services in two sites, long delays with research governance in one site followed by difficulties in re-engaging with the service meant that no patients were ever approached. The second telecare service ultimately decided against participation owing to staff shortages.

Of the 29 PWD who were eligible, 12 people agreed to have their dementia status confirmed and to take part in the study (41% of eligible participants). Of those who declined, the patient declined in nine cases and the carer declined in eight cases. There were no significant differences in baseline characteristics between those who agreed and those who declined to participate in the study [enrolled mean age 81 years (SD 5.58 years), not enrolled mean age 81 years (SD 8.09 years), mean difference –0.37 years; enrolled 75% male, not enrolled 41% male; p = 0.176].

All final assessments of outcome measures were completed by the CTA for all participants (except the PWD who died shortly after recruitment). However, two participants did not complete all planned intervention visits. One participant had a fall and sustained a neck of femur fracture after visit 10 and did not have a final therapy visit. A second participant requested to stop intervention sessions after visit 10 but did agree to a final therapy visit at week 12.

Views on recruitment processes

Data on recruitment were obtained through interviews with CTAs and other professionals involved in recruitment (see Chapter 7, Data collected for the process evaluation). Comments on recruitment related to two key areas: the eligibility criteria and feasibility of recruitment processes. No specific feedback was obtained on retention because in this section we rely on data from staff involved in recruitment.

Discussion

There were uncertainties about the feasibility of achieving the target of 30 PWD from the outset of the study; nevertheless, we met the progression criterion of recruiting at least 40% of eligible PWD. If we had been able to extend the recruitment period for the study, we may have been able to reach our recruitment target.

The findings have important implications for any potential further implementation of the DIFRID intervention. Research governance processes affected recruitment in two ways: inconsistent research governance requirements (1) undermined the use of a CTA to facilitate recruitment in one ED and (2) were likely to have resulted in under-recruitment from this service. Obtaining approvals for the inclusion of telecare services was challenging: in one site, there were significant delays that eventually led to a loss of interest from the service, and, in another, we experienced considerable difficulties in identifying the department responsible for approving the work. Staff shortages and pressures of work led to one telecare service eventually declining to participate; as already discussed, these factors are also likely to have affected recruitment in the ED where we were unable to use an embedded CTA.

Although these factors were largely outside the control of the research team, the qualitative work highlighted the need for improved systems and communication between the research team and local sites and within local sites to improve co-ordination regarding recruitment.

The failure to recruit any PWD from research registers suggests that this is not a viable approach to recruitment in the light of the resources required to contact large numbers of PWD to check eligibility. Although changing the eligibility criteria to include falls in the previous 3 or 6 months was suggested, this had little impact on recruitment rates in one general practice where the requirement to have fallen in the previous month was inadvertently omitted. A more convincing argument for extending the period since the fall is to enable all acute interventions to have been completed prior to recruitment to the study.

Although there are published data suggesting that falls are common in PWD,2 we experienced recruitment difficulties despite modifying the eligibility criteria and including alternative services and approaches to recruitment. The low recruitment rates in both the diary study (see Chapter 3) and the feasibility study (see Chapter 8) suggest that PWD may not seek health-care attention following a fall. Although they may contact telecare services via an alarm system, we were unsuccessful in our attempts to include such services. We are therefore unable to conclude whether or not telecare services would represent a viable source for recruitment.

Introduction

In this chapter, we present findings relating to the feasibility and acceptability of the intervention. We focus on three key methodological issues relating to intervention delivery in feasibility research:161,162

1. intervention adherence (both the adherence of front-line staff to the intervention and the extent to which PWD engaged in the intervention and adhered to the planned activities)

2. acceptability of the intervention

3. whether or not it was possible to calculate intervention costs and duration.

Two further issues161 relate to whether or not the logistics of running a multicentre trial were assessed and whether or not all components of the protocol worked together. The extent to which we are able to address these issues is limited owing to the scale of our feasibility study and the nature of the data collected (see Chapter 7).

The findings are presented for each component of the intervention in sequence (i.e. training and supervision; assessment; MDT meetings; referrals; goal-setting and activity-planning; ongoing intervention sessions; reviews and future-planning), followed by a review of the costs of the intervention and the logistics of intervention delivery. We conclude the chapter by drawing together the findings and their implications using the NPT framework. 159

Training and supervision

The aim of the training was to ensure that staff felt equipped and confident to deliver the DIFRID intervention by improving their knowledge and understanding of dementia and discussing the intervention components. The training was developed by the research team and the physiotherapist and OT who were seconded to the team to facilitate intervention development and training. Concerns about the feasibility of delivering a day-long training programme meant that training was condensed into a single half-day session. The areas covered by the training and approximate time allocated to each are shown in Table 14. The intention had been to provide all professionals with the equivalent of tier 2 dementia training;145 however, just under 1 hour was allocated specifically to working with dementia.

Assessment

Eleven PWD received two assessments by the therapists. The assessment comprised three main sections: a generic section to be completed by the therapist (either physiotherapist or OT) making the initial visit, a physiotherapy assessment and an occupational therapy assessment. The section on goals and action planning was completed by the therapist who made the second visit. Although it was intended that both assessment visits would take place during week 1, the mean time between assessment visits was 7.3 days (SD 4.47 days) and 54% of participants took longer than 1 week to complete. This was followed by a mean of 9 days (SD 3.03 days) before the first intervention session; thus, the intervention did not start as promptly as anticipated. An overview of the assessment components and completion rates is given in Table 15. Assessment documents were completed in full with the exception of those for osteoporosis risk (36%), TUG score (82%) and lying and standing blood pressure (82%). The initial osteoporosis risk assessment [Fracture Risk Assessment (FRAX)]163 was completed by the CTA as part of the baseline outcome assessment and passed onto the therapists with the referral document. The low level of completion suggests that the scores either were not included in the referral document or were not transferred to the assessment document. A needs list was completed for 9 of the 11 participants (82%) but action planning was completed for all. Because the action planning was derived from the needs list, this suggests that the needs list may have been seen as redundant by some.

Some professionals reported that the initial assessments worked well and successfully incorporated new components, such as measuring blood pressure or the person with dementia’s ability to multitask, into the assessment:

It was good, yes. It was quite lengthy. I could follow it through fine. I managed to do the blood pressure. So, that’s good. The assessments were all ones that we are used to anyway. Apart from when you have to count and do the activity. We don’t normally count backwards and do the activity. [. . .] I think the length was fine. The patient seemed OK with it.

Professional 130, physiotherapist (interview)

Although some staff conducted additional tests not included in the assessment document for individual PWD where they seemed relevant (e.g. to get an indication of the patient’s stamina), others felt that the assessment was too long and that there was repetition between the physiotherapist and OT assessments and the baseline measures completed by the CTA. It was suggested that the assessment be combined into a single visit with both therapists to reduce burden on the PWD and carers. In teams with generic roles where a single therapist would normally cover all components of the assessment, the division between occupational therapy and physiotherapy assessment was artificial:

By the time I went out I was the third professional going and asking lots of questions. [. . .] I could sense that the wife was feeling a little bit like, ‘I feel like I’ve been asked these questions before.’. On reflection, perhaps if I’d have gone out with the physio[therapist], and we’d done the assessment together, that might have been less onerous for the carer.

Professional 150, OT (interview)

One thing that we found a little bit difficult was that, because we’re so generic here, I would go out and normally do the physio[therapist] and the OT bits, so I was having to hold myself back.

Professional 155, physiotherapist (interview)

Despite the inclusion of questions about carer assessment, capacity to support the intervention and training needs in the assessment documentation, there was little evidence that these areas were considered.

Multidisciplinary team meetings

It was intended that on completion of the assessment each patient would be discussed by a MDT, including all staff directly involved in the assessment or delivery of the intervention sessions (physiotherapist, OT and support workers) and a geriatrician. The aim of the MDT was to discuss the assessment and develop an action plan based on patient and carer goals. Intended outcomes were to agree:

• the types of activities and interventions that are most appropriate to meet this patient’s goals

• the referrals required and a named individual responsible for each

• the key worker for the participant

• the number of intervention sessions needed for the first 6 weeks and who would deliver them.

It proved difficult to identify a geriatrician to join the MDT in one site; instead, the therapists used an existing contact to discuss medical issues. However, this meant that the intended holistic review of all PWD by a MDT was not possible in this site. Owing to the difficulties in arranging meetings where members worked in different locations, only one face-to-face MDT meeting with a geriatrician was held. Other MDT meetings were conducted by telephone (sometimes without teleconference facilities). Staff with experience of both a face-to-face meeting and a telephone meeting agreed that face-to-face MDT meetings where multiple PWD were discussed were more effective:

I definitely thought face to face was better. [. . .] I think because when you are face to face, everybody could sort of add their little bit, whereas when you were on the telephone, I would have a conversation and then say pass the phone over to [Professional 141] for her little bit of conversation. It wasn’t as joined up, because obviously I couldn’t hear the conversation that [Professional 141] was having, and she could only hear my responses on the bit that I was having.

Professional 137, OT (focus group)

Not all team members involved in intervention delivery participated in MDT meetings; sometimes support workers fed back information to therapists who then met with the geriatrician:

We had our team meet up and then we had the consultant phone in. That was useful because both patients had postural hypotension, on assessment. So that was useful to have their advice on that. We only, really, had one formal MDT because that was the only medical thing we needed to talk about. Otherwise it was more ad hoc. We’ve been e-mailing feedback to each other, after interventions, and talking in the office and stuff like that, when we’re around. I hope that’s OK.

Professional 153, physiotherapist (interview)

This quotation suggests that, at this site, the MDT was seen as only being relevant when there was a ‘medical thing’ to discuss rather than being an integral part of the assessment. Following assessment and MDT meetings, a total of 16 referrals were made, suggesting that participants had significant unmet needs. Two participants were referred to a geriatrician, three were referred to their GP, one was referred to a continence adviser, one was referred to wheelchair services, two were referred to equipment services and six were referred to other services. Only one carer was referred to a carers’ centre.

Goal-setting and activity-planning

By centring the intervention on goals identified by PWD and carers (refined by the MDT if needed) we hoped to maximise engagement and motivation. At the first intervention session, goals were agreed with the participant, the GAS form was completed and the project diary was introduced. Details of the rating of goals at the outset and end of therapy are provided in Chapter 10, because this relates to their potential use as outcome measures.

Overall, 31 goals were recorded; these were reviewed and categorised by one of the therapists involved in developing the intervention. Four goals were excluded as they were either too vague or were signposting or actions rather than goals (e.g. ‘proper medical assessment’). The remaining 27 goals were grouped into four categories: outdoor activities (n = 12), self-care (n = 7), indoor household tasks (n = 5) and indoor leisure activities (n = 3). The scope of the goals varied considerably; for example, one goal was ‘To be able to go into town on the community bus and access coffee shop three times over 12 weeks’, whereas another was ‘To stand for long enough to brush own teeth and hair on a daily basis’. This suggests that goals were tailored to the specific abilities of the PWD involved. Staff sometimes struggled to set goals that followed SMART (Specific, Measurable, Attainable, Realistic and Timely) principles, suggesting that additional training or review of goals may have been useful.

The assessment and intervention documentation included space for up to three goals; this was often interpreted by staff as a requirement for each person with dementia to have exactly three goals. For PWD who identified one or two challenging goals, there was sometimes not sufficient time, energy or motivation to tackle a third goal as well. Where PWD and therapists were struggling to identify goals, there was a tendency in some MDT meetings to add ‘default’ goals that were neither grounded in the assessment nor of particular interest to the PWD, such as making a hot drink. Alternatively, therapists sometimes substituted actions for goals:

They were struggling to come up with a third goal. They have already begun the process of getting assistive device for the toilet so [Professional 130] thought that this could also be a goal.

Field notes from informal discussion with Professional 130, physiotherapist

Despite some of the challenges of goal-setting, it proved successful with some PWD to the extent that some therapists intended to integrate GAS into their normal practice:

I quite enjoyed the goal-setting. [. . .] Using the GAS score was really interesting, and it’s something that we are going to try and include with our own patients as well.

Professional 154, physiotherapist (interview)

Defining expected outcomes as part of GAS provided a series of smaller goals that enabled participants to document progress. For example, for one person with dementia with the goal of being able to ‘exit the property and walk round the driveway to the car’, the following steps were described:

It was walking outside to the patio, which he cracked. It was walking outside to the car, which he did. It was getting in and out of the car, which I did with the OT, with a little bit of equipment. He had a swivel cushion. We talked about a handle, but actually he didn’t need it. It was practising that a few times, which we did. Then he goes out with his family quite a lot in the car now. Then it was walking a little bit further. [. . .] There are lots of little goals that we’ve just kept moving along, moving along, moving along.

Professional 134, support worker (interview)

However, goal-setting was not always as successful, particularly with PWD with more severe impairment who were unable to grasp the purpose of goal-setting or to retain goals. It was suggested that the term ‘goal’ was potentially off-putting to PWD, and that framing this part of the intervention in a different way might have been more successful:

It’s usually older clientele, and they’re just not used to the word goal. So we would be, obviously, using different wording around that. What would they like to get out of it or what were they expecting? All different wording, really, to try and tease out anything, but it’s been really hard to get anything.

Professional 153, physiotherapist (interview)

Even when PWD identified goals, these were not always included. In the following quotation, the person with dementia’s goals seem more cognitive than physical, which may explain why the therapist found these goals difficult to operationalise. In the second quotation, the therapist viewed the goals as unrealistic:

[Patient 17] does not have many goals but he manages quite well currently, for example with his drinks and meals. He didn’t rate much of anything on the Compass of Life, though they used it. He said that his goals are to anticipate things better and to understand whether he’s doing the right thing.

Field notes from informal discussion with Professional 130, physiotherapist

He had very unrealistic goals about playing boules and setting up a boules club in [local area], and going [abroad], returning to his house [abroad]. He would talk a lot about that, a lot, and really, really desperately wanting to get those things set up [. . .] It was very difficult to try and get him to focus on other things, because realistically we were never going to meet any of those goals.

Professional 141, physiotherapist (focus group)

However, some therapists expressed surprise at the progress achieved by some PWD, suggesting that ambitious goals should not be ruled out:

I questioned myself whether she would achieve getting into town on the bus on her own, and she did. That amazed me; that absolutely amazed me.

Professional 154, physiotherapist (interview)

Some staff recognised that some of the difficulties with goal-setting may have stemmed from their own lack of skill or experience; more training or supervision might have helped them to find ways of engaging more successfully with PWD:

I’d probably like more training on the psychological side of motivating people that have the cognitive problems, because it can be quite a barrier. I do think we tend to say, and it might be lack of training, ‘Oh, well, they won’t do it, that’s it then’. Yes, it might be, but I think sometimes training might teach us otherwise, you know?

Professional 147, OT (interview)

A final issue relating to goal-setting concerned the timing; therapists found that some PWD had more ideas about goals once they had developed rapport with staff and had become more familiar with the process:

After the relationship developed, he could then feel comfortable about discussing other things. I think, initially, he found it more difficult to pinpoint anything in particular, because he wasn’t used to being asked to do that.

Professional 133, support worker (interview)

Ongoing intervention sessions

The intervention was tailored to each participant based on the needs identified at the initial assessment(s). On average, participants reported having 12 planned intervention visits over the 12-week follow-up period {mean 12 [SD 5], median 11 [interquartile range (IQR) 9–17] visits}. Four participants had at least one intervention session that was not delivered, despite staff recording travel time or time spent at the visit (three of these participants missed either one or two visits and one participant missed seven visits). Reasons for missed visits included ‘patient not at home’, ‘patient requested visit not take place’, ‘10 min discussion with reablement team’ and ‘no answer at door’. The average number of intervention sessions that took place was 11 [mean 11 (SD 3.9), median 10 (IQR 9–14)]. Therefore, 94% of planned visits took place. The number of sessions potentially available as specified in the manual and the number delivered by different professionals are summarised in Table 16.

The number of visits was, therefore, substantially lower than envisaged, particularly for OTs. This may have been owing to difficulties with the logistics of intervention delivery (see Factors influencing implementation of the intervention) and the fact that two participants discontinued therapy after session 10.

In each intervention session, staff were intended to review the activities the participant had undertaken since the previous visit, check the project diary and discuss any falls (forwarding details of any adverse events to the principal investigator). Documentation of this process was recorded on 77% of occasions. Having reviewed the activities, staff then considered whether or not and how to modify activities in order to progress towards the goals. The activities were to be informed by the participant’s likes and dislikes and the type of activity they were most interested in and, therefore, most likely to follow. The manual and training emphasised the importance of embedding activities into participants’ everyday lives, for example walking to the local shop (if they routinely needed items such as a newspaper or milk) or walking the dog. Documentation of the activity-planning process was available for 81% of occasions.

Observation of intervention sessions indicated varied approaches to activity-planning. We observed some PWD being given exercise sheets and being advised how many of each exercise to do and how frequently. There was little evidence of explicit discussion of how the exercises would help PWD to achieve their goals. Review of therapist notes and patient diaries highlighted the tendency of some therapists to rely heavily on exercises. In some cases, this was despite explicitly noting that the person with dementia would prefer to undertake activities and was not adhering to the exercise programme:

. . . motivated to exercises but would prefer to be doing activity rather than just doing exercises.

Notes of intervention (Patient 22)

Furthermore, over-emphasis on terms such as ‘exercise’ and ‘fitness’ could alienate PWD:

[Professional 129] tells him that she wants to do some exercises with him to improve his fitness and he is quite resistant to this, saying ‘I’m an old man. I don’t need fitness’.

Field notes of intervention session, Patient 19 and Professional 129, physiotherapist

There was also evidence in the notes that some therapists relied heavily on carers to support the intervention between visits, suggesting that ways of embedding activities were not always considered and that carer capacity was not necessarily considered:

I would say they need to have somebody that is able to facilitate it [. . .] I just think if you can’t treat at the level of intensity required for carry over, there is no point in starting what you are then not going to finish. I think if you are giving daily balance exercises and they do them once a week, you are going to make your intervention look ineffective, because it has not been done at the required intensity. Or the other way round. That would be massively increasing the amount of support worker involvement, which would be extremely expensive.

Professional 129, physiotherapist (focus group)

Although some aspects of the intervention were similar to usual practice, the key difference identified by staff was the increased time available. This allowed staff to monitor progress more closely, engage more creatively and design more tailored activity programmes, which was seen as beneficial to PWD:

I think maybe we’d seen him more than we would do normally. Normally we would give them exercises and say to the carer, ‘Can you get him to do these?’. We’d review them a couple of times and then hope that they would carry on doing them. So, it’s more that supervision, that we don’t normally offer.

Professional 130, physiotherapist (interview)

Having the time to do it, I suppose, was good, rather than being limited in my follow-up, which is what it would normally be. I felt that we did a really full job. [. . .] If I had done it in my job role, we’d have probably just had to dash in and dash out. I don’t think it would have achieved the confidence giving that perhaps she needed, and he needed.

Professional 134, support worker (interview)

In addition to enabling staff to provide more hands-on help with activities, extra time also facilitated the development of rapport and gave staff greater insight into the day-to-day lives of PWD:

You built a relationship with that particular patient/client. You actually got to know their world and the people around them as well, which you wouldn’t necessarily do on a 2-week basis. [. . .] But it just made you see their world as well on a longer term, and what it is like for them on a day-to-day basis. That is where we could see definitely from a positive point of view, the intervention is needed out there.

Professional 151, support worker (focus group)

In contrast to the assessment document, which was easy to follow and complete, the paperwork for recording intervention sessions could be difficult to understand and time consuming to use. This was exacerbated by the overlap with the project diary (in which activities were also to be recorded as an aide-memoire for the PWD and carers) and need to maintain the usual clinical notes:

Just with having the notes in the house and having the notes here, and some of the tables and the charts, it wasn’t really clear how to fill them in. I filled them in the best that I could in the way that felt appropriate to me, but whether that was what you are looking for I don’t know. It just felt like a lot of duplicating. We had to do it here and then put it on our system as well, so it was time consuming.

Professional 141, physiotherapist (focus group)

Although the project diaries could have been an effective way for intervention team members to communicate with one another, they were not consistently used as intended. Although the person delivering the intervention was supposed to update the list of activities in the diary at each visit, this was not consistently done. Some teams met informally or communicated via e-mail or telephone to discuss their work with PWD, and others suggested introducing more joint visits:

You know, maybe another review visit at a certain point. We were relying on [support worker] all the time to come back and tell us, ‘Oh, when’s our next visit?’. So it perhaps would’ve been nice to have had another – even a joint visit in the middle somewhere, or at a certain point.

Professional 154, physiotherapist (interview)

Reviews and future-planning

Two formal reviews were scheduled during the intervention: one at the approximate mid-point (6 weeks) and one at the end of the intervention (12 weeks). The purposes of these reviews were to:

• check if all the referrals from the MDT meeting had been acted on

• record the GAS scores at these time points

• discuss the participant’s progress

• consider plans for progression and ongoing support.

In addition, at the final review, the therapist repeated the TUG and explored PWD’ and carers’ views on the intervention, including whether or not they would like to be referred to any ongoing services to facilitate maintenance and progression [e.g. Staying Steady (Newcastle)/community-based balance groups]. We observed one final review, which suggested that the PWD had enjoyed the intervention and that some suggestions had been embedded:

When asked about what she liked about the intervention, [Patient 25] responded that she enjoyed the ‘chat’ and found the visits ‘uplifting’. [. . .] She said that she was likely to continue going to town on her own. [. . .] As part of the intervention, [Professional 155, support worker] had installed a whiteboard for reminders. [Patient 25] described that she had initially found this difficult to remember and required a lot of prompting from husband, therapists and her daughter. However, this had become more embedded over the course of the study and [Patient 25] and her husband said that they this was something they would also continue using.

Field notes from observation of physiotherapist and support worker intervention session

This extract highlights the work required to successfully embed new ideas; the length of the intervention and support of the person with dementia’s family were key to the successful implementation of the reminder system.

Resources to deliver the intervention

As part of the intervention documentation, staff were asked to record travelling time as well as the length of visits. However, this was not always straightforward; sometimes staff forgot, and visits were often organised sequentially, so it was difficult to separate out the travel associated with one specific patient:

Sometimes, if you had other patients and you were going straight from my patient to there, I forgot to fill in what time I set off and then what time I got back.

Professional 155, physiotherapist (interview)

Table 17 summarises the available information on time spent travelling to and from and delivering the different sessions. One participant was missing information on the time spent at their second assessment and the time spent travelling back from this assessment. Information on time was available for eight participants at their final visit.

On average, the travel time and time spent at visits were similar regardless of the type of visit. However, as expected, the most time spent at a visit was to assess the participant and tailor the intervention to their needs, with the least amount of time spent at the final visit.

Logistics of intervention delivery

The lack of dedicated posts to deliver the intervention meant that we relied on existing staff to take on extra hours. Furthermore, this approach limited the extent to which the intervention sessions could be tailored to individual PWD and carers:

[Professional 157, support worker] commented that 9:30 was not really the best time to arrange sessions for the patient but she was constrained by her own schedule – she had to arrange to see study patients on her usual day off and had been scheduled to work half a day, in the morning [. . .] She also mentioned that the hospital had changed her days several times, which meant that she had not been able to consistently offer [Patient 26] a day and time for appointments. This in turn had led to [Patient 26] declining visits on several occasions.

Field notes from observation of support worker observation session

In addition to staff availability, sessions could be disrupted by other commitments of PWD and their carers. Some staff were faced with working outside their usual geographical area. As well as increasing travelling time, staff were not necessarily aware of which support services were available in the area, limiting their ability to signpost PWD and carers to other services.

Although most staff valued the opportunity to deliver a more extensive intervention, some commented that this had in part been due to low recruitment rates and queried whether or not they would have been able to provide such an effective intervention if recruitment had been more successful:

I mean, the amount of work that [Professional 155, support worker] has put in has been brilliant, you know, we couldn’t have done it without all the input that she put in. Had we have had more patients, I don’t know how that would have affected [Professional 155]’s caseload. That would be my query, but just looking at the research, you know, ideal.

Professional 154, physiotherapist (interview)

Delays in reimbursement meant that staff in one team had to deliver the intervention within their normal working hours. This may account for the reduced number of visits and tendency of some staff to revert to usual patterns of working. For example, one person with dementia received only three intervention visits because the therapist felt that the person with dementia’s carer was successfully supporting the exercise programme provided.

Factors influencing implementation of the intervention

Understanding the likelihood of new interventions being successfully embedded into routine practice is a key component of feasibility work. NPT is a well-established theory that has been used in over 100 studies seeking to understand factors influencing implementation. 164 We used the framework of NPT to inform data collection and analysis. NPT considers both the individual work and the collective work required for successful implementation of a new intervention. It focuses on four key areas: (1) coherence – whether or not the new intervention makes sense to stakeholders and is clearly different from current practice, (2) cognitive participation – whether or not stakeholders engage with and invest in the new intervention, (3) collective action – whether or not the new intervention is adequately supported in terms of resources, skills and training, and (4) reflexive monitoring – whether or not stakeholders can determine the impacts of the new intervention and adapt it to suit the local context. 159 An overview of the key factors influencing the implementation of the DIFRID intervention within the feasibility study and recommendations for future implementation is provided in Table 18.

The NPT analysis highlights some key areas for future development. Difficulties in translating theory into practice, in relation to both the intervention and working with PWD, highlight the need to ensure that training has a practical focus, and to provide additional training or supervision to develop the skills needed for successful implementation. The findings also suggest that more attention is needed to foster investment in and engagement with the intervention. Having strong local leadership is a well-established component of implementation strategies,165 yet was not achieved in all sites. Specifically, building relationships with local geriatricians may be one way of increasing buy-in in the MDTs while also providing local leadership. A key issue was the tendency of staff to revert to established approaches, for example by focusing on exercises; similar problems with introducing new ways of working were encountered in a previous study. 58 Intervention supervision may be a key way of monitoring and addressing this issue and could link to the involvement of geriatricians if they were willing to take on this role. The limited data collected suggest that early outcomes could either challenge preconceptions about dementia (if successful) or confirm them (if staff were unable to engage PWD in goal-setting). Appropriate supervision could provide opportunities to share outcomes, problem-solve as a team and learn from one another.

Discussion

The findings of the feasibility study suggest that the DIFRID intervention is both feasible and acceptable to stakeholders.

Adherence to the initial assessment was relatively good. There were, however, some difficulties in identifying meaningful goals with or for PWD. Difficulties in goal-setting with older people have previously been reported. 166,167 This suggests that further training and review of goals by a specialist member of the research team is needed, particularly in the early stages while skills are still developing. It was clear that the goals achieved sometimes exceeded the expectations of staff; successful work with PWD could, therefore, help to challenge the negative attitudes towards dementia expressed by some professionals (see Chapter 4, Data collection and analysis).

There was evidence of poor implementation of two key aspects of the intervention: (1) MDT meetings and (2) carer assessment, support and training. The difficulties in identifying a geriatrician in one site, and limited opportunities for discussion in other sites, meant that the holistic assessment and collaborative goal-setting envisaged in the intervention was not always realised. Further consideration is needed regarding the recruitment of geriatricians to support MDT meetings, clarification of the purpose of the meetings and documentation of such meetings. Although the intervention was intended to assess and address carer needs alongside those of the PWD, the lack of explicit attention to this in the study paperwork meant that there was little evidence of staff exploring carer needs for support, education or training in any detail.

Given that the feasibility study took place in three sites, we gained some insight into the logistics of running a multicentre trial. Ongoing supervision or training is a key area for development. Concerns about the willingness of staff to commit time and effort to training led to relatively brief training sessions; this seemed insufficient to ensure that staff fully understood study procedures and provided little time for skill development.

The organisation of participating services also highlighted the need for the intervention to be flexible to fit with the local context; for example, in services in which staff have a more generic role, requiring parts of the assessment to be conducted by an OT and a physiotherapist may need further justification and discussion with local services.

Service organisation further suggests that a cluster RCT may be the most appropriate design for a future trial of the intervention. Randomising individual PWD is unlikely to be feasible because it would require individual front-line staff to alter their behaviour for some PWD but not for others. Randomising individual members of staff is also likely to be problematic because support workers typically work across all team members. Furthermore, the finding that some teams found the intervention manual sufficiently useful to share it with colleagues highlights the potential for contamination within teams.

Introduction

This chapter summarises the feasibility of the data collection tools used to capture information on the outcomes collected (outlined in Chapter 7) during the feasibility study. We focus on the extent to which the measures proved feasible and acceptable to PWD and their carers, and the CTAs responsible for the administration of these measures. We also consider the outcomes collected by the therapists during the intervention. Two methodological issues relating to intervention delivery in feasibility research are addressed: whether or not outcome assessments were complete and whether or not the outcome measures were those that were most appropriate. 161,162

Clinical trials assistant-administered outcome measures

The CTAs were responsible for collecting outcome data from PWD and carers at baseline and 12-week follow-up. There was some confusion surrounding the timing of the 12-week follow-up assessments: one CTA calculated the 12-week follow-up from the date of the CTA’s baseline assessment, whereas the other two CTAs calculated the date from the therapists’ initial assessment visit. Clearer communication and documentation of this process are required.

Outcome assessment measures were generally completed in full and all questions were generally completed with the exception of Quality of Life in Alzheimer’s Disease (QOL-AD) and EQ-5D-5L. At baseline, all PWD (n = 11) completed the EQ-5D-5L, with 82% (n = 9) of carers completing the proxy version. At 12 weeks, 91% (n = 10) of both PWD and carers completed the appropriate version of the EQ-5D-5L. There was a completed proxy version of the EQ-5D-5L questionnaire for the PWD who had not completed this measure at follow-up. All self-reported and proxy EQ-5D-5L questionnaires that were completed had no missing data for any of the domains. One CTA misunderstood that the proxy questionnaires were to be completed even if the PWD also completed their version of the questionnaire (at both time points). Summary data for the outcome measures are given in Table 19.

Therapist-completed outcome measures

As part of their initial and final visits, the lead therapist completed two outcome measures for the participant. One was a physical test of mobility and balance (TUG) and the other was GAS, which involved identifying goals, weighting the importance and difficulty of these goals, defining a range of expected outcomes and rating the person with dementia’s current abilities. The GAS was intended to be completed at the 6-week review in addition to at baseline and the 12-week follow-up.

The TUG was completed for 10 of the PWD at baseline and 9 out of the 11 PWD at follow-up. The GAS had considerable missing data, with complete information at all time points for only three PWD (Table 21).

Detailed examination of data relating to the 27 goals identified on the GAS indicated that there were marked problems with completion of ratings of the importance and difficulty of the goals identified (Table 22). Therapists appeared to improve completion of the ratings of current performance in relation to the goal at later time points, but ratings for all three time points were provided for only around three-quarters of the goals identified. Ratings at all three time points were complete for all goals identified for just over half of the PWD (54.5%) (see Table 22).

Reviewing the ratings of PWD’ abilities at the three time points provides some insight into the potential sensitivity of the use of GAS to changes resulting from the intervention. Ratings were available for at least one goal for eight PWD at all three time points, and only these data are included in the analysis (Table 23). The findings show that performance did not deteriorate for any goals over the period of the study, and that there was no improvement on a single goal. At 6 weeks, the ratings of three goals were unchanged; the performance on the remaining goals had improved (with an average increase of 2 points). At the 12-week follow-up, the most common outcome was for PWD to have maintained progress with their goals, with no change from their score at 6 weeks; however, five PWD made further gains on six goals, all but one by a single point (with one person with dementia gaining 2 points on one of their goals).

Self-completed outcome measure

Of the 11 participants, 10 (91%) completed the falls diaries. In total, 9 out of 10 (90%) had at least one fall. The median number of falls was 2 (IQR 1–6, range 0–23). A single participant had a very large number of falls (23). This very frail participant had very frequent falls before coming into the study and these continued.

Safety reporting

Adverse and serious adverse events were as expected and none were judged to be directly related to the intervention.

There were 12 adverse events reported in six participants:

1. swollen leg; drained and tired; dizziness

2. dizziness

3. facial injury caused by jewellery; buttock pain; elbow and shoulder pain

4. painful legs and hip

5. back pain

6. two instances of back pain; sickness.

There were four serious adverse events reported in four participants:

1. hospital admission with symptoms of stroke

2. hospital admission with fall and left radial head fracture

3. hospital admission with fall and neck of femur fracture

4. hospital admission with fall.

Acceptability of outcome measures

Feedback from PWD, carers and CTAs completing the outcome measures indicated that the most common concern related to the duration of the baseline and follow-up assessments:

I know that we had to do all the paperwork that we did, but I think it could be streamlined a bit, I think there was a bit of repetitiveness in the questions we were using [. . .] for the patients, it was a bit too much when you’re sat in the house. We only had, like, 90 minutes but I couldn’t do the first one in less than 2 hours because he kept getting upset and crying, it was very difficult.

Professional 145, CTA (interview)

I thought that was a bit long and drawn out, 2 hours. It was about 2 hours to start with. I didn’t think it had got to be anywhere near that long, but in the end, that went quickly. We just didn’t know what to expect. I’d got no idea, no idea at all. I just said yes thinking, ‘Well, if it doesn’t work out and he doesn’t like it, we can just stop it anyway.’.

Carer 16a (joint interview with Carer 16a, Carer 16b and Patient 16)

One carer found the question relating to Carer’s Allowance offensive and was not willing to discuss finances; however, these questions appeared acceptable to other participants. Some of the CTAs identified problems with administering the Modified Falls Efficacy Scale (MFES) and QOL-AD to PWD. The wording of the MFES was thought to be complex for PWD, and simpler phrasing was suggested:

For example, the Modified Falls Efficacy Scale, that was hard to explain to the patient that had more advanced dementia. It’s almost like a double negative. So, you’re asking them, ‘How confident are you that you won’t fall doing a certain thing,’ whereas, if you could just say, ‘Are you worried about falling when you get dressed and undressed’, that’s a much more straightforward question.

Professional 140, CTA (interview)

Although the CTAs recognised the importance of asking the questions of standardised scales verbatim, in practice they found that PWD often needed further explanation or clarification, especially when the questions were ambiguous, for example on the QOL-AD scale, which may have also affected the validity of the data:

I think there’s probably better quality-of-life questions [. . .] so physical health, that’s fair enough. Physical health, ‘How would you rate your physical health? Poor, fair, good, excellent.’. That’s pretty straightforward. Energy, that’s pretty straightforward. Mood, that’s pretty straightforward. Then, things like family. What about family? The support from your family? Whether your family live nearby? So, they’d look in and say, ‘Well, what do you mean, family?’. Do you know what I mean? ‘How do you rate your family?’. What? How proud you are of them? It’s a really difficult one to say.

Professional 140, CTA (interview)

The comments highlight the need for training for the CTAs responsible for collecting outcome data to ensure that standardised scales are being approached in the same way.

Relevance of the outcome measures to changes resulting from the intervention

The goals set focused on outdoor activities, indoor household tasks, indoor leisure activities and self-care; it is probable that achieving these goals would have an impact on quality of life as measured by QOL-AD and EQ-5D-5L. The relevance of the MFES for this study is unclear as only two PWD had significant fear of falling. One consequence of dementia is that some PWD may lose insight into their difficulties and may not be aware of their falls risk. Although we anticipated that GAS would provide a tailored and sensitive assessment of changes made during the intervention, there were difficulties in identifying appropriate goals, and the data required for scoring was often incomplete. These changes could be addressed through further training and supervision. Of more concern are the comments made by front-line staff about the timing of the goal-setting, which have implications for using GAS as an outcome measure.

Discussion

We have demonstrated that assessments were completed at all time points for 9 out of 11 PWD and 9 out of 11 carers. Furthermore, examination of completion of individual measures indicated few missing data. Only one person with dementia could not complete the TUG at baseline and follow-up, which is a good completion rate for a balance and mobility measure. Another participant could not complete the TUG at follow-up because of a fractured hip. The feasibility of using the EQ-5D-5L as a means of estimating health state utilities for PWD was demonstrated, even with those PWD who did not have capacity to consent to study participation for themselves. Removing the HUQ from the patient diary and completing it within a face-to-face interview proved successful, in terms of both potentially increasing completion rates for the (now simplified) diary and acceptability to carers. The high completion rate of the HUQ supports its use in a future definitive study. The feasibility of using the Disability Assessment for Dementia (DAD) and Zarit Burden Scales with the carer was also demonstrated with complete data.

The MFES was completed by all PWD, but in the light of the comments on the complexity of the MFES, in future it may be worth considering using the iconographical version of the Falls Efficacy Scale, which includes illustrations of common activities as verbal cues. 168 This has been validated in a number of countries and appears to have good validity and reliability. 147,168

Confusion over the timing of outcome measures and completion of proxy measures highlights the need for additional training for CTAs responsible for data collection. Our findings also suggest that training on outcome measures would ensure a consistent approach and give CTAs the opportunity to share and resolve common difficulties in administration or with question wording.

As described in Chapter 9, staff delivering the intervention required more training and supervision on the use of GAS. Although changes in scores were consistently positive, there were problems with missing data, particularly on the ratings of importance and difficulty, which are required to produce standardised scores. 151 A key advantage of GAS is that it is tailored to the priorities of individual participants; therefore, it may be more sensitive to change than standardised outcome measures.

The number of falls reported in the falls diary highlights the frequency of falling in this population. Together with the data on adverse events and health-care utilisation, the findings indicate that serious falls resulting in hospitalisation appear common (reported by 3 out of our 11 participants). That three such falls occurred during our 12-week intervention highlights the need for careful monitoring of adverse events in any future implementation and also indicates the financial and personal costs of falling for PWD and their carers.

Summary of the key findings

This report describes a series of mixed-methods approaches to answer the questions ‘Is it possible to develop a complex intervention to improve fall-related injuries in PWD living in their own homes?’, ‘What is the feasibility and acceptability of the intervention?’ and ‘Is it feasible to plan a future RCT to evaluate the efficacy of the DIFRID intervention?’. We showed that it was possible to design an intervention, although a key change was that the intervention should be delivered to all PWD living in their own homes who present with a fall requiring health-care attention and not just those who sustain an injury.

In Chapter 2, we described a systematic review assessing the previous evidence of effectiveness of interventions to improve outcomes for PWD who fall. We found gaps in the evidence base. The studies used different interventions, reported multiple different outcomes and included people with cognitive impairment as well as those diagnosed with dementia. The quality of evidence was mixed and the results across the studies conflicted even when similar interventions were utilised. Most of the study populations presented with hip fracture in hospital so interventions may not be applicable to soft tissue injuries or other types of fracture, and these studies provided no guidance about managing fall-related injuries in primary care. This suggested there was still a need for research into whether or not an effective intervention for fall-related injuries in dementia could be designed and delivered.

In Chapter 2, we also looked at how such an intervention should be evaluated. We concluded that the evaluation of a falls prevention intervention should identify and cost all of the resources required to deliver the intervention and any subsequent health and social care resource use. The outcomes that need to be considered are the number of falls and the QALYs based on responses to the EQ-5D-5L. Sensitivity analyses should be adopted to address any uncertainty.

In Chapter 3, we wished to describe current usual care and assess the demand for a future intervention for PWD who sustain a fall-related injury. We found that the incidence of fall-related injuries coming to attention in the settings of the ED, paramedic attendances and primary care consultations was much lower than expected. However, for those who did present, it was evident that usual care consisted of very little input. This suggested there was scope for improvement in the care received by such PWD. The HUQ was piloted as part of the diary study and refined for WP4 based on the data provided and feedback from the qualitative interviews.

In Chapter 4, we used qualitative methods, including interviews, focus groups and observation, to develop a better understanding of current care pathways and identify opportunities for intervention. The findings suggested that improving outcomes for PWD after a fall depends on recognising and facilitating the rehabilitation potential of PWD. The three key areas that need to be addressed are (1) ensuring that services are organised in the most effective and supportive way for PWD, (2) improving attitudes, knowledge and skills of professionals working with PWD, and (3) supporting carers and their role in the interventions.

In Chapter 5, we used a realist approach to synthesise the current evidence regarding the management of falls in dementia, further develop the key areas identified in Chapter 4 and develop theory regarding how a new intervention might work. We developed nine CMOcs, which were further grouped into three broad themes: (1) ensuring that the circumstances of rehabilitation are optimised for PWD, (2) compensating for the reduced ability of PWD to self-manage and (3) equipping the workforce with the necessary skills and information to care for this patient group.

In Chapter 6, we prioritised, operationalised and validated components of a complex intervention. We did this by convening a consensus panel that participated in two meetings and two Delphi consensus rounds. At this stage, the panel decided that the intervention should be delivered to those who had experienced a fall requiring health-care attention and not just those who had sustained an injury. The intervention that was designed was a multidisciplinary intervention to be carried out in PWD’ own homes over 12 weeks. Up to 22 intervention sessions could be undertaken but this was to be tailored according to the person with dementia’s need. The detailed methods for the intervention to be tested in the feasibility study were described in Chapter 7.

In Chapters 8–10, we described the results of the feasibility study in terms of recruitment and retention (see Chapter 8), acceptability of the intervention (see Chapter 9) and feasibility of outcome measures (see Chapter 10). We were unable to achieve the target of 30 PWD, which is not surprising given the lower than expected incidence of fall-related injuries coming to attention in the ED, paramedic attendances or primary care consultations, as we found in Chapter 3. It is likely that we would have been successful had we been able to extend the recruitment period; however, this was not possible within the funding envelope for the study. Nevertheless, we met the progression criterion of recruiting ≥ 40% of eligible PWD. The study suggested that the DIFRID intervention is both feasible and acceptable to stakeholders. A number of modifications were recommended to address some of the issues arising during the feasibility testing. These mainly centred on the need to expand training in the intervention for the staff delivering it and also clarifying the process of goal-setting. It was suggested that goal-setting could be carried out over a longer period at the start of the intervention to enable PWD to engage more in the goal-setting process. The process of measuring outcomes was largely successful. However, we did identify a need for more training for both CTAs and therapy staff, particularly in the use of the GAS. The costs associated with the intervention and subsequent health-care resource use of PWD were identified and included in the data collection tools piloted in WP4. Overall, the data collection measures were completed well. The data collection tools derived for this study can be used in a future trial evaluating this intervention and we would recommend that the HUQ is recorded by the CTA alongside the other outcome measures. The unit costs associated with each of these resources would need to be identified as part of a pilot trial.

Strengths and limitations

There were a number of strengths and limitations of the study. In Chapter 2, the systematic review followed established review methodologies including comprehensive searching for evidence and independent risk-of-bias assessment. However, the number of studies identified was small and four studies that otherwise met the inclusion criteria for this review could not be included. We were not able to perform a meta-analysis and we found significant gaps in the evidence base, especially for non-hip-fracture injuries. The studies did not show evidence of any particular adaptation of the approach, enhancement of the skills or composition of MDTs given that they were working with a population that was different from that of older people without a cognitive impairment. In addition, most of the interventions were not aimed at patients with known dementia; subgroup analysis was used to report the effects of general interventions on this group. The review of approaches to evaluate cost-effectiveness also found only a small number of studies. Not including falls recovery in the search terms means that we may have missed some potentially eligible studies. The risk of bias was not determined for the two eligible studies. This is a potential limitation of our results but in the context of this review it was not a major concern as the focus was on what sort of economic evaluation methodology to follow.

In Chapter 3, we used careful methods to quantify the number of people presenting with a fall-related injury but the incidence was much lower than expected. We believe that presentations to paramedics may have been particularly underestimated. Unfortunately, in the diary study we did not reach our target of 60 participants. We believe that the requirement for health professionals to seek permission from potential participants to share their contact details with the research team contributed to poor recruitment due to time constraints and so this was modified for WP4. Completion of the diaries was fairly successful and we modified our approach for WP4 in the light of the burden of completing health service use questions. Given the low level of use of health services by participants completing the diary study, we were able to identify a clear gap in current services that could be addressed by a new intervention.

In Chapter 4, the strengths of the study were the inclusion of multiple perspectives, in particular being able to observe staff with differing levels of experience and interest in dementia care. The direct observation of service delivery highlighted a range of issues relating to communication skills that were not emphasised during interviews. However, we experienced a number of recruitment difficulties, resulting in relatively small numbers of PWD and carers being included in both observations and qualitative interviews. Nevertheless, a range of suggestions for improving existing services and potential components of a new intervention were identified, which were presented to the consensus panel as part of WP3.

In Chapter 5, a realist approach allowed us to consider and synthesise a broader range of evidence, which was important in the light of the limited evidence identified in the effectiveness review. We used established methods of realist review. 91 The review was limited by the lack of published evidence found to support concepts that were clearly articulated in the qualitative data, such as in CMOc5 (providing ongoing support). Further research is needed in these areas. Owing to time limitations, we were not able to conduct additional targeted searches for all CMOcs and the review was completed by the time of the consensus panel meeting, so more research may have been published in these areas since we completed our review.

In Chapter 6, we used an intervention development process that was iterative and included a range of stakeholder perspectives. The development of the intervention was strengthened by including a range of professionals with expertise in falls prevention and rehabilitation. However, we did not include PPI representatives as the expected technical level of the presentations involved was not thought to be suitable for PPI involvement. We did not achieve full participation at the two meetings or in the Delphi rounds. The stakeholder interviews took place at an early stage of development of the intervention between the two meetings and it would have been useful to have time for more stakeholder interviews after development of the intervention materials. The development process enabled us to successfully develop the methods described in Chapter 7.

In Chapter 8, we showed that it was possible to recruit participants to receive the intervention, but a limitation was that because of the short time period for recruitment we did not reach our target number of participants. Nevertheless, it is a strength that we met the progression criterion of recruiting ≥ 40% of eligible patients. Retention to completion of the outcome assessments was good, although two participants did not receive the full 12 weeks of intervention sessions. In one case, this was because of a hip fracture; in the other case, this was a result of participant request. Adverse events and serious adverse events were as expected for this type of study in which frail older people can be expected to have a number of events unrelated to the study intervention. The participant who fractured their hip was having frequent falls both before and after receiving the intervention; therefore, it is not possible to be sure that their injury was directly attributable to increased activity from the intervention.

In Chapter 9, we were successful in using a process evaluation to suggest a number of ways of optimising the intervention prior to further testing or evaluation. This study comprehensively addressed the requirements of the MRC guidance on process evaluations. 169 However, a limitation is that we only had a small number of participants and carers. Nevertheless, the ways that we can optimise the intervention will be useful in further evaluations of this intervention.

In Chapter 10, we were successful in demonstrating that the outcome measures we selected could be completed by most participants and their carers. It is a strength that we supplemented data on completion rates of outcome measures with interview data with those responsible for administering or completing the measures. This highlighted that even when measures had good completion rates, the wording was sometimes complex and difficult to explain to PWD. One limitation of this aspect of the study was that the research team had no access to the notes maintained by the therapists (which included the TUG, DAD and GAS) until after the end of the intervention and completion of qualitative data collection. Opportunities for exploring the reasons for poor completion of the GAS were, therefore, missed. The findings confirm that many of the measures used in the feasibility study are suitable for use in a future trial of the intervention. The importance of additional training for CTAs and staff delivering the intervention was highlighted to ensure a consistent approach and to minimise missing data.

Implications for taking the DIFRID intervention forward

Our study used a comprehensive approach to developing a complex intervention for falls in PWD using the MRC framework. We have designed a tailored, individual approach to falls prevention in PWD and our systematic review identified that this approach has not been evaluated in previous studies. We are aware of only two trial protocols for studies that will aim to prevent falls in PWD. 56,170 Both of these protocols are aimed at the primary prevention of falls in PWD. Although the inclusion criteria do allow PWD to have had a fall before entering the study, the majority of participants in these two trials will not have experienced a fall. It is important to note that PWD who have already experienced a fall may have more severe dementia than those who are yet to experience a fall and, therefore, studies specifically in the group who have experienced a fall are needed.

We were able to meet most of the progression criteria for progress to a full trial, although the percentage of eligible participants consenting to a feasibility study was rated as amber. However, the number of people who met the eligibility criteria was lower than we were expecting at the outset of this research and a number of factors will need to be taken into consideration regarding potential recruitment rates for a full trial. We identified an important need for the CTA to have access to the details of the potential participants rather than relying on ED staff to ask if details could be shared with the CTA. In the inclusion criteria, extending the period since the fall is likely to be helpful to enable all acute interventions to have been completed prior to recruitment to the study. This is likely to be particularly useful in identifying participants from primary care. The use of research registers is unlikely to be a helpful way of identifying participants.

We identified a number of modifications to the intervention that would be useful. There were some difficulties in identifying meaningful goals with or for PWD. This suggests that further training and review of goals by a specialist member of the research team is needed, particularly in the early stages while skills are still developing. We also found that further training in working with PWD would be valued by the intervention teams. Further consideration is needed regarding the recruitment of geriatricians to support MDT meetings, clarification of the purpose of the meetings and documentation of such meetings. Although the intervention was intended to assess and address carer needs alongside those of the PWD, it is clear that more attention needs to be given to carer assessment and intervention in the materials for the intervention.

We did not have a control group in the present study and, therefore, we were unable to test procedures of randomisation and whether or not participants would be willing to be randomised. Given the modifications to the intervention that have been suggested, we recommend that it would be useful to further refine the intervention in a pilot trial before proceeding to a full trial. This would enable the procedures for randomisation to be tested as well as allowing further opportunity for refinement of the intervention.

In future work, we will apply for funding for a pilot trial. We propose a cluster randomised design. Recruitment challenges will be addressed by increasing the number of sites and lengthening the recruitment period. We will also ensure that a CTA is able to screen notes directly in each site. The time since the fall allowed in the inclusion criteria will be extended to 6–12 months. Additional training for therapists will be provided and geriatricians will be specifically recruited at each site to participate in the MDT. Additional support will be provided to carers.

Conclusions

The study has highlighted the feasibility of delivering a creative, tailored, individual approach to intervention for PWD following a fall. Although the intervention required greater investment of time than in usual practice, many staff valued the opportunity to work more closely with PWD and carers. We conclude that further research is now needed to refine this intervention through a pilot RCT.

Aims

The aims of the PPI in the study were to:

• ensure that the study was relevant to PWD and their carers

• ensure that the participant-facing materials were understandable and suitable for purpose

• assist with the dissemination of the study.

Methods

During the development of the study protocol, a focus group was held to discuss the commissioning brief and help the team develop their ideas for the protocol. This was convened by VOICE, an organisation that aims to capture the public’s vast experience, ideas, opinions and expectations of research, innovation and policy developments that affect their lives. After funding was awarded, VOICE advertised for representatives to join the PPI panel for the study. Two volunteers joined the Programme Management Group and one joined the TOC and attended these meetings regularly.

The members of the Programme Management Group met the researchers at regular intervals during the study. They reviewed all participant-facing materials and made suggestions as to how they could be improved. After WPs 1–3, they worked with the research team to produce lay summaries of the reports about each WP. These were placed on the study website (https://research.ncl.ac.uk/difrid/) and disseminated to individuals involved in the study.

Results

The initial focus group supported the importance and relevance of the research topic identified by the commissioning brief and supported the design of the protocol. However, members of the focus group did not make any suggestions that altered the overall design of the study.

The presence of PPI representatives was very helpful in designing participant-facing materials. Changes were made to the designs as a result of their suggestions.

The lay summaries were made more readable as a result of the input of the representatives.

The presence of representatives at Programme Management Group meetings and TOC meetings enabled discussion of the potential impact of the study on participants and ways of encouraging recruitment.

Discussion and conclusions

Involvement of PPI representatives was a positive aspect of the DIFRID study. Without their involvement, the participant-facing materials may not have been so engaging for participants. The team were able to balance the need to collect accurate data with the need not to place an undue burden on participants. Dissemination to members of the public via the study website was assisted by their involvement.

Reflections/critical perspective

Because of the time scale for initial development of the protocol for the funding call, there may have been limited opportunities for the focus group to contribute to the design of the protocol. A series of focus groups may have encouraged further discussion.

Only three people contributed to ongoing PPI once the study was funded and this may have placed some burden on those taking part. A larger group would be useful for future studies.

Contributions of authors

Louise M Allan (https://orcid.org/0000-0002-8912-4901) (Professor of Geriatric Medicine) was the principal investigator for the study. She contributed to drafting the report, analysed the quantitative data, contributed to the realist review, developed the manual and developed and delivered the training.

Alison Wheatley (https://orcid.org/0000-0002-6051-0286) (Research Associate) contributed to drafting the report, collected and analysed the qualitative data, contributed to the realist review, developed the logic model, developed the manual and developed and delivered the training.

Amy Smith (https://orcid.org/0000-0003-1398-9623) (Occupational Therapist) was a member of the consensus panel, oversaw the delivery of the intervention, developed the manual and developed and delivered the training.

Elizabeth Flynn (https://orcid.org/0000-0003-0646-4722) (Physiotherapist) was a member of the consensus panel, oversaw the delivery of the intervention, developed the manual and developed and delivered the training.

Tara Homer (https://orcid.org/0000-0002-6664-0671) (Research Associate, Health Economics) contributed to drafting the report, analysed the quantitative data for the health economics and was a member of the consensus panel.

Shannon Robalino (https://orcid.org/0000-0002-6359-1827) (Guest Researcher) contributed to the systematic review.

Fiona R Beyer (https://orcid.org/0000-0002-6396-3467) (Information Specialist) contributed to the systematic review and the drafting of the report.

Christopher Fox (https://orcid.org/0000-0001-9480-5704) (Professor of Old Age Psychiatry) contributed to the realist review, oversaw the study at the Norwich site and was a member of the consensus panel.

Denise Howel (https://orcid.org/0000-0002-0033-548X) (Senior Lecturer, Statistics) advised on design and statistical analysis of WP2 and WP4.

Robert Barber (https://orcid.org/0000-0001-9464-8446) (Honorary Clinical Senior Lecturer, Old Age Psychiatry) contributed to the realist review, was a member of the consensus panel and oversaw the Clinical Research Network staff at the Newcastle site.

Jim Anthony Connolly (https://orcid.org/0000-0002-3819-7772) (Consultant Physician, ED) oversaw the study in the Newcastle ED and was a member of the consensus panel.

Louise Robinson (https://orcid.org/0000-0003-0209-2503) (Professor of Primary Care and Ageing) advised on the primary care aspects of the study.

Steve Wayne Parry (https://orcid.org/0000-0002-8364-9192) (Clinical Senior Lecturer, Geriatric Medicine) contributed to the realist review and was a member of the consensus panel.

Lynn Rochester (https://orcid.org/0000-0001-5774-9272) (Professor of Human Movement Science) contributed to the realist review and was a member of the consensus panel.

Lynne Corner (https://orcid.org/0000-0002-6806-0304) (Director of Engagement) advised on PPI in the study.

Claire Bamford (https://orcid.org/0000-0003-2885-801X) (Senior Research Associate) led the qualitative work, contributed to drafting the report, collected and analysed the qualitative data, contributed to the realist review, developed the logic model and developed the manual.

Publications

Allan LM, Wheatley A, Flynn E, Smith A, Fox C, Howel D, et al. Is it feasible to deliver a complex intervention to improve the outcome of falls in people with dementia? A protocol for the DIFRID feasibility study. Pilot Feasibility Stud 2018;4:1–13.

Bamford C, Wheatley A, Shaw C, Allan LM. Equipping staff with the skills to maximise recovery of people with dementia after an injurious fall. Aging Ment Health 2018;15:1–9.

Robalino S, Nyakang’o SB, Beyer FR, Fox C, Allan LM. Effectiveness of interventions aimed at improving physical and psychological outcomes of fall-related injuries in people with dementia: a narrative systematic review. Syst Rev 2018;7:1–11.

Wheatley A, Bamford C, Shaw C, Boyles M, Fox C, Allan LM. Service organisation for people with dementia after an injurious fall: challenges and opportunities. Age Ageing 2019;48:454–8.

Wheatley A, Bamford C, Shaw C, Flynn E, Smith A, Beyer F, et al. Developing an Intervention for Fall-Related Injuries in Dementia (DIFRID): an integrated, mixed-methods approach. BMC Geriatr 2019;19:57.

Data-sharing statement

All data requests should be submitted to the corresponding author for consideration. Access to anonymised data may be granted following review.

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 HTA 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 HTA programme or the Department of Health and Social Care.

MEDLINE, EMBASE and NHS Economic Evaluation Database

Date searched: August 2016.

Recommended Read codes

Where detailed information on subtype of dementia is available, then the Read codes below can be used. The below table matches the International Classification of Diseases, Tenth Edition (ICD-10),69 codes to recognised general practice dementia Read codes.

All Read codes

Work package 2: patient and carer interviews

• Introduction.

• Your recent fall.

• Experience of <name of service>.

  • Suitability of service for people with dementia.

• Recovering from the fall.

• Improving services and developing a new approach for people with memory problems who have fallen and hurt themselves.

Work package 2: professional interviews and focus groups

• Role and service.

• Clientele.

• Referral routes into the service.

• Onward trajectory.

• ‘Fit’ of people with dementia within the service (including group activities).

• Views on perceived value of specific intervention for people with dementia.

• Views on key components of an intervention for people with dementia.

• Specific training needs in relation to people with dementia with fall-related injuries.

• Facilitators and barriers to implementing change.

• Use of outcome measures.

Work package 4: patient interviews

• Did you feel this was a good intervention for you?

  • Tell me more about that.

• What did you like about the intervention sessions?

• What did you dislike? What could have been different?

• Which aspects of the DIFRID intervention have been most useful?

• Which aspects of the DIFRID intervention have been least useful?

• How did you feel about the activities that you were asked to do? (Were they personalised enough?)

• Has the intervention made any difference to you?

• Thinking about the intervention materials such as the diary and the manual, are there any changes that you think we should make?

• Could you tell me a bit about the staff delivering the intervention?

• Can you tell me a bit about the goal you have been working towards?

• What kinds of patients do you think would benefit most from this type of intervention? Are there patients for whom it would not be useful?

• Is there anything else that we haven’t covered about the DIFRID intervention?

Work package 4: carer interviews

• What were/are your expectations about the DIFRID intervention?

• Did you feel this was a good intervention for <name>?

  • Tell me more about that.

  • What did you like about the intervention sessions?

  • What did you dislike? What could have been different?

  • How engaged did <name> seem to be in the intervention?

  • Which aspects of the DIFRID intervention have been most useful?

  • Which aspects of the DIFRID intervention have been least useful?

  • How did you feel about the goals that <name> has been working towards?

• Has the intervention made any difference to <name>?

• What about yourself, how involved have you been in the intervention?

• Thinking about the intervention materials such as the diary and the manual, are there any changes that you think we should make?

• The DIFRID intervention is delivered by physiotherapists, occupational therapists and rehabilitation assistants. From your perspective what are the advantages and disadvantages of using staff with this skill mix?

• Could you talk about your perception of the staff delivering the intervention?

• What kinds of patients do you think would benefit most from this type of intervention? Are there patients for whom it would not be useful?

• Is there anything else that we haven’t covered about the DIFRID intervention?

Work package 4: interviews with staff responsible for recruitment and assessment of outcomes

• Can you start by telling me about how the recruitment process has worked in practice?

• How interested do patients seem to be in the intervention?

• Have there been any patients who met the inclusion criteria but you felt were not appropriate for the study?

• What sense do you have of how feasible it would be to proceed to a full trial with a control group etc.?

• Can you tell me about how the assessment processes and outcome measures have worked in practice?

• What sense do you have of how useful the intervention is?

Work package 4: interviews with staff responsible for developing the intervention and training and supervising intervention delivery

• How do you think the training went?

• How has the supervision process been?

• How do you think the study/intervention is going so far?

• How confident are you in the intervention?

• What reservations do you have about the intervention?

• Which aspects of the DIFRID intervention do you feel most confident with/have been most useful?

• Which aspects of the DIFRID intervention do you find most challenging/have been least useful?

• Overall, what sense do you have of how useful the intervention is?

• In this study, the DIFRID intervention is delivered by physiotherapists, occupational therapists and rehabilitation assistants. From your perspective what are the advantages and disadvantages of using staff with this skill mix?

• What kinds of patients do you think would benefit most from this type of intervention? Are there patients for whom it would not be useful?

• Based on feedback from supervision, how interested do patients seem to be in the intervention?

• From your perspective, what are the facilitators and barriers to getting people/patients engaged in the study?

• From your perspective, what are the facilitators and barriers to implementing the intervention?

• From your perspective, what are the facilitators and barriers to evaluating the acceptability and impacts of the DIFRID intervention?

• Are there any changes we should make to the DIFRID intervention?

• Is there anything else that we haven’t covered about the DIFRID intervention?

Work package 4: interviews and focus groups with staff delivering the intervention

• What are/were your expectations about the DIFRID intervention?

• How do you think the study/intervention is going so far?

• How confident are you in the intervention?

• What reservations do you have about the intervention?

• Which aspects of the DIFRID intervention do you feel most confident with/have been most useful?

• Which aspects of the DIFRID intervention do you find most challenging/have been least useful?

• What opportunities have you had to discuss the value of the intervention with your colleagues?

• How might we modify the DIFRID intervention?

• From your perspective, what are the facilitators and barriers to implementing the intervention?

• Overall, what sense do you have of how useful the intervention is?

• In this study, the DIFRID intervention is delivered by physiotherapists, occupational therapists and rehabilitation assistants. From your perspective what are the advantages and disadvantages of using staff with this skill mix?

• What kinds of patients do you think would benefit most from this type of intervention? Are there patients for whom it would not be useful?

• Based on feedback from supervision, how interested do patients seem to be in the intervention?

• From your perspective, what are the facilitators and barriers to getting people/patients engaged in the study?

• Could you describe the process of tailoring the intervention to the individual patient?

• How helpful were different components of the intervention (e.g. training, manual, MDT meetings, supervision)?

• Are there any changes we should make to the intervention materials (e.g. the assessment form)?

• From your perspective, how well did the intervention ‘fit’ with other services?

• Do you feel you have the support you need to deliver the intervention?

• What dementia training had you previously received?

• How do you think the training went?

• How well has the supervision process gone?

• In terms of taking this work forward, how could we improve the initial training session(s)?

• How will your experience with the DIFRID intervention influence your usual practice in the future?

• Is there anything else that we haven’t covered about the DIFRID intervention?

MEDLINE, Cochrane Central Register of Controlled Trials, Health Management Information Consortium, EMBASE, Cumulative Index to Nursing and Allied Health Literature, Web of Science, Allied and Complementary Medicine Database and Physiotherapy Evidence Database

Date searched: November 2015.

The MEDLINE literature search strategy is provided here. This strategy was translated as necessary for each of the resources searched. This is one example of the initial broad search strategy. Parallel research strategies were created for other databases; the one shown in here was for MEDLINE (Ovid).

MEDLINE and Cumulative Index to Nursing and Allied Health Literature

Date searched: February and March 2017.