Improving the diagnosis and management of Lewy body dementia: the DIAMOND-Lewy research programme including pilot cluster RCT

Article history

The research reported in this issue of the journal was funded by PGfAR as project number DTC-RP-PG-0311-12001. The contractual start date was in January 2014. The final report began editorial review in May 2019 and was accepted for publication in March 2021. As the funder, the PGfAR programme agreed the research questions and study designs in advance with the investigators. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The PGfAR editors and production house have tried to ensure the accuracy of the authors’ report and would like to thank the reviewers for their constructive comments on the final report document. However, they do not accept liability for damages or losses arising from material published in this report.

Disclaimer

This report contains transcripts of interviews conducted in the course of the research and contains language that may offend some readers.

Copyright statement

Copyright © 2021 O’Brien et al. This work was produced by O’Brien et al. under the terms of a commissioning contract issued by the Secretary of State for Health and Social Care. This is an Open Access publication distributed under the terms of the Creative Commons Attribution CC BY 4.0 licence, which permits unrestricted use, distribution, reproduction and adaption in any medium and for any purpose provided that it is properly attributed. See: https://creativecommons.org/licenses/by/4.0/. For attribution the title, original author(s), the publication source – NIHR Journals Library, and the DOI of the publication must be cited.

2021 O’Brien et al.

Introduction

Lewy body dementia (LBD) is a term used to describe two closely related conditions: dementia with Lewy bodies (DLB) and Parkinson’s disease dementia (PDD). They form the second most common neurodegenerative dementia (after Alzheimer’s disease), but it is recognised that they are underdiagnosed. Diagnosis has important implications for management, as those with LBD present with a range of variable clinical features, including cognitive impairment, motor symptoms, autonomic symptoms, sleep problems and neuropsychiatric features.

We undertook this programme because little is known regarding the prevalence or diagnostic practice of LBD in the UK and, although national management guidelines [e.g. National Institute for Health and Care Excellence (NICE) 20181] do recognise the importance of both DLB and PDD and make some limited management recommendations, there are no systematic approaches to management or recognised care pathways. DIAMOND-Lewy (Improving the DIAgnosis and Management of Neurodegenerative Dementia of Lewy body type) was a comprehensive, multidisciplinary 5-year programme of work seeking to investigate and implement ways to improve both the diagnosis and management of LBD (i.e. both DLB and PDD) within the NHS.

We achieved this through a number of interlinked work packages (WPs), with a very strong patient and public involvement (PPI) core running throughout the programme. In this synopsis, we first describe our approach to PPI and how this influenced the design of the programme. In WP1, we undertook a retrospective case note study to investigate current practice in secondary care NHS services with regard to the diagnosis and management of LBD. In WP2, we undertook systematic reviews of pharmacological and non-pharmacological management of LBD and used this evidence, supplemented by a Delphi study of expert clinicians, to produce an evidence-based management toolkit. In WP3, we developed assessment toolkits for the diagnosis of LBD and, incorporating the management toolkit from WP2, piloted these assessment and management toolkits in routine NHS memory assessment and movement disorder services. In WP4, we introduced the assessment and management toolkits more broadly within a pilot cluster randomised trial. WP5 was a series of qualitative studies with three components: (1) an investigation of the barriers to and facilitators of making a diagnosis and managing LBD, (2) exploration of views on the assessment and management toolkits and (3) exploration of implementation of the assessment and management toolkits in clinical practice. Finally, in WP1 repeated (WP1R), we undertook a repeat assessment of DLB and PDD diagnostic rates following introduction of the assessment toolkits in some of the same services that had participated in WP1.

A research pathway diagram showing how these WPs linked together is provided (Figure 1). Each WP will now be discussed in turn after a description of the PPI core running throughout the programme.

FIGURE 1.

Research pathway diagram. PD, Parkinson’s disease; RCT, randomised controlled trial.

Introduction

Patient and public involvement is seen as fundamentally important to the quality and relevance of research. In the light of calls for greater transparency and consistency in reporting PPI, we have followed the GRIPP2-SF (Guidance for Reporting Involvement of Patients and the Public 2–short form) reporting guidelines, highlighting the impacts of PPI and critically reviewing our approach to PPI. 2–4 Much of this section of the report was drafted collaboratively with members of the PPI group.

Aim

The aim was to use PPI to inform all phases of the study from patient and carer perspectives.

Methods

Patient and public involvement in this programme started very early, during the development of the proposal, with one PPI member (Derek Forster) being a co-applicant. Once the programme started, we established a PPI group that met regularly throughout the programme. Potential volunteers were sought through a public meeting organised by the North East Dementias and Neurodegenerative Research Network (DeNDRoN) in July 2013 and through the existing networks of Derek Forster and the DeNDRoN PPI co-ordinator. Three PPI representatives agreed to join the group with Derek Forster, and all four members contributed throughout the programme. All members’ lives have been touched by neurodegenerative diseases, either as a patient or as a carer. In addition, members were motivated by their desire to ‘give something back’, interest in medical research and enthusiasm to help progress research into LBD. Experience of PPI varied. One member had been involved in PPI for > 10 years; however, for another this was their first PPI experience. All members received training on research through DeNDRoN. This was supplemented by an introductory session on LBD by Ian McKeith. One of the co-applicants (CB) facilitated the group and meetings were held on university premises. We followed the INVOLVE guidance on reimbursing group members for their time and expenses. 5

The first PPI meeting focused on how to achieve the desired aim. This led to agreement that the group would:

• share personal knowledge and experience of DLB, Parkinson’s disease (PD) and PDD

• comment on the practicalities and pressures that participating in the research could create for patients and carers (safeguarding)

• ensure that information was user-friendly and did not alarm potential participants

• ensure that a patient/carer focus was included in developing the management toolkit

• contribute to study newsletters.

Members agreed that the timing and duration of PPI meetings should reflect the needs of the programme. Although the group thought it was unnecessary to link individual PPI members to specific WPs, they were keen to embed PPI into the programme management structures. To achieve this, Derek Forster attended both the PPI group and the Programme Management Group (PMG) and Valerie Argent joined the Programme Steering Committee (PSC) as a lay member.

In addition to the PPI group, two PPI workshops were convened to inform the development of the management toolkit (see Work package 2: development of management toolkits for Lewy body dementia). Potential participants were identified by team members through patient and carer interviews conducted during WP5 and local contacts. One member of the PPI group (Anne Lister) also attended the PPI workshops for WP2. The content of the workshops is described in Work package 2: development of management toolkits for Lewy body dementia.

Outcomes of patient and public involvement

The PPI group met 20 times, with full attendance at nearly all meetings. The number of meetings ranged from one to five per year. Members often prepared for meetings by reading documents, and commented on documents by e-mail between meetings. The initial expectation of group members and the research team was that the group would adopt a consultation approach, with an emphasis on ensuring that documents would make sense to patients and carers, including advising on the feasibility and acceptability of the study procedures. Over time, however, the group grew in confidence and were proactive in developing and expanding their role, adopting collaborative and user-controlled approaches at different times. 6 For example, rather than commenting on the public version of the study newsletter (which was initially drafted by the research team), members took ownership of the newsletter by redesigning the layout, introducing a new structure and writing much of the content themselves.

A user-controlled approach was adopted to obtain feedback on participants’ experiences of taking part in the pilot trial. Rather than inviting a single participant to describe their experience at the PPI dissemination event in the North East, members were keen to present a broader range of experiences. This led to the development and inclusion of some open questions about study participation with the letter inviting participants to the PPI event. Overall, 28 detailed responses were received. Two members of the group (VA and AL) reviewed and collated the information and presented their findings at the PPI dissemination event. Key themes were the interpersonal skills of the interviewers and having an opportunity to talk about their situation. Everybody found the home visits convenient and > 70% of respondents said that they would be willing to take part in a similar study. These comments provided useful feedback for the researchers and highlight aspects of study design, which may maximise recruitment and retention in future studies. Furthermore, the comments highlighted the lack of support available after diagnosis, suggesting new avenues for research and intervention.

Key activities undertaken by both the PPI group and the PPI workshops in WP2 are related to the research cycle7 and are summarised in Table 1.

Having highlighted some of the successful contributions made by PPI during the programme, we also acknowledge that its impact in some areas was limited. Owing to uncertainty about which outcome measures would prove sensitive to changes resulting from the introduction of the management toolkit, the team proposed a range of measures. Reconciling the perspectives of PPI members and those of the research team was sometimes challenging, for example in relation to the number of outcome measures and the perceived burden of these on participants, and in relation to the wording of questions in the assessment toolkit, some of which were considered to be inaccessible to patients and carers. However, there was limited scope to make changes because they were validated measures.

Discussion and conclusions

Using more than one approach to PPI in the programme proved effective. While the PPI group provided continuity throughout the programme, the workshops enabled the inclusion of a broader range of perspectives, including those of people with dementia. By combining a ‘fully intertwined’ approach to PPI and the ‘one-off’ approach for the workshops, we were able to include patients and carers who did not wish to commit to ongoing input. 8 The inclusion of PPI members on the PMG and PSC ensured opportunities for communicating with the full team. The detailed understanding of the programme developed through the PPI group, facilitated the ability of the PPI members to contribute confidently and effectively to meetings.

All members of the PPI group continued to be involved throughout the programme. This continuity helped the group to develop in confidence and expertise. Although we have benefited from stable group membership throughout the programme, members recognise that this may be unusual, as the health or circumstances of PPI members may deteriorate over a 5-year period, particularly if they are directly affected by the condition.

A limitation of the PPI group was the homogeneity of members in terms of ethnicity and class, and the lack of ability to alter some validated questions that members thought contained complex and inaccessible terms. Previous studies have emphasised the challenges of ensuring diversity in PPI, in particular the inclusion of people with dementia. 9 None of the members of the PPI group had dementia (although two had PD). Although caring for a spouse with LBD clearly provides significant insight into living with the condition, this is not the same as first-hand experience. 9 Despite the efforts of the DeNDRoN PPI co-ordinator, no people with dementia were identified for the PPI group. This partly reflects the low diagnostic rates and delays in diagnosing LBD (see Work package 1: baseline study of the diagnosis and management of Lewy body dementia). The situation was exacerbated by the high demand for PPI members in Newcastle upon Tyne, a Centre of Excellence in LBD research.

Reflections

Recent publications on PPI have emphasised the need for increased understanding of effective approaches to PPI. 4,8,10 Unfortunately, we did not obtain formal feedback from workshop participants, although the attendance of several participants at the subsequent PPI dissemination event suggested an ongoing interest in the programme. The perception of the PPI group was that although not everyone in the research team initially recognised the potential value of PPI, their contributions were increasingly valued as the programme progressed. Although no formal evaluation or opportunities for reflection were planned at the outset, the process of drafting newsletters, giving presentations at the PPI event and drafting this section for the final report prompted the group to recognise their wide-ranging contributions and reflect on their experiences of this approach to PPI. Key factors thought to contribute to the success of the PPI group were as follows:

• having a simple two-page overview of the programme to help members understand the different WPs and how they were interlinked

• having regular meetings to maintain interest and understanding of the programme

• working together in a group and sharing ideas

• having a facilitator

• attending to practical issues, for example ease of access, free parking, timely reimbursement of expenses and flexibility over the timing of meetings

• having access to a computer and large screen to enable the group to produce or edit documents collaboratively during meetings

• having stability of group membership

• ensuring that meetings were business-like, but enjoyable.

Several of these factors are consistent with those identified in previous studies. 8 Three suggestions were made for facilitating PPI contributions to future projects. The first concerned the importance of all team members recognising and valuing the different perspectives offered by PPI members, even when these challenged their own views. The second related to greater clarity over the responsibilities and remit of the PPI member within the PSC and finally the group emphasised the need for the chairperson of the PSC to explicitly invite comments from the PPI member.

After 5 years, members of the PPI group are surprised at their continuing levels of enthusiasm and commitment. Reflecting on this, members attribute their loyalty to the relevance of the programme to their lives, compatibility of members and the appreciative facilitation. In terms of personal benefit, members have valued the opportunity to keep up with research, maintain mental agility and develop new skills. Although the programme has now ended, the PPI group is planning to share their experiences to wider audiences through a conference presentation and an article for lay audiences.

In summary, the inclusion of PPI throughout the programme was extremely valuable and added very positively to the design of the DIAMOND-Lewy programme, its conduct and outcome, with PPI members making key contributions that affected the design and administration of the assessment and management toolkits. PPI also played a key role in dissemination and feedback of the study to participants, with members taking ownership of content for regular study newsletters, designing suitable methods for thanking participants and playing key roles in the co-design and delivery of the programme dissemination events.

Introduction

The accurate recognition of dementia and the diagnosis of dementia subtype helps ensure appropriate management, and are central to improving patient care. LBD, comprising both DLB and PDD, is the second most common cause of neurodegenerative dementia in older people, accounting for between 15% and 20% of all cases of dementia, according to autopsy studies. 11 However, clinical and epidemiological studies suggest rates of only 5–7%. 12 This suggests that a considerable number of people with LBD go undiagnosed and so are not offered potentially beneficial management strategies. More worryingly, incorrect diagnosis risks inappropriate management with drugs such as antipsychotics, which, even at a low dose, can have serious side effects and cause mortality in those with LBD. 11

Currently, the proportion of cases of dementia diagnosed with DLB and PDD in NHS secondary care services is not known, nor has there been any systematic study of diagnostic practice or management in secondary care, although a US-based retrospective survey indicated frequent erroneous diagnosis and delays to a final correct diagnosis. 13 This is an essential first step to improving care. Therefore, this WP sought to establish the baseline for current diagnostic rates for DLB and PDD, providing a foundation for the other WPs. The WP was split into two parts: (1) WP1A examined diagnostic and management practice for DLB and took place in old age psychiatry and memory services, and (2) WP1B examined diagnostic and management practice for PDD. This latter element was undertaken in movement disorders services run by neurologists and geriatricians.

Initially, the plan had been to undertake WP1 solely in services in one geographic region of the UK (the north-east of England). However, there arose an opportunity to undertake a much more informative two-region study, whereby diagnostic rates for DLB and PDD could be ascertained in both the North East and East Anglia, two representative but clearly separate UK regions. This change also allowed some comparison of whether or not any variation in diagnostic rates and practice existed between regions.

Work package 1A: baseline study of the diagnosis and management of dementia with Lewy bodies

Work package 1B: baseline study of the diagnosis and management of Parkinson’s disease dementia

This study, also undertaken over the first 2 years of the programme, looked at diagnostic and management practice in those with PDD, with the aim of establishing baseline diagnostic rates for PDD and comparing investigations, assessment and management of those with PDD with control participants (i.e. patients with PD without dementia).

Objectives

The objectives of WP2 were to (1) complete a systematic review of evidence about the management of LBD and (2) use this to develop a clinical guideline, using a Delphi process incorporating the views of a multidisciplinary panel of independent experts combined with two facilitated PPI workshops.

Systematic reviews

At the time that these reviews were undertaken and to the best of our knowledge, no evidence-based reviews of the comprehensive management of LBD existed. As a consequence, little formal management guidance for LBD was offered in the NICE dementia guideline used at that time,18 despite the fact that this group of patients has multiple symptoms requiring treatment. Clinical guidance at the time appeared solely in the form of pharmacological reviews and expert consensus opinion statements. 14,19,20 This was an unsatisfactory situation for prescribers, health-care providers and recipients of care. We therefore conducted systematic reviews on pharmacological and non-pharmacological management of DLB.

Public–patient workshops

Two public–patient workshops were held, with the participation of 38 people with LBD and their family/carers. The first event focused on identifying best practice in LBD clinical management, based on their own experiences. The content of the workshops is shown in Figure 2. The emergent themes were developed further in the second event and refined into a set of guiding principles.

FIGURE 2.

Workshops and content.

Some of the key themes that were held to be important are listed in Table 3. The themes were structured in such a way as to help the clinical team deliver the guideline recommendations in a user-friendly format.

TABLE 3 - Management principles arising from public–patient workshops

Management principle	Summary of discussion point
1	Provide practical information at an early stage (e.g. power of attorney, attendance allowance, council tax)
2	Advantages and disadvantages to a problem list were identified: Risks that problem lists generate unwarranted concerns A long list of problems could be demoralising Reassuring as realise that these are expected symptoms not part of a new illness
3	Recognise that the carer is often the expert on the patient Recognise the danger of doctors treating what they know best
4	Main reliance on decision-making is guidance from the doctor; people want to be told what the doctor feels would be best for them
5	Carer diary can be useful to see the pattern of response to medication on good and bad days (but may not be completed when things are OK or improving)
6	Increased carer stress in DLB – need to explicitly discuss with carer on their own
7	Peer support groups may be helpful (providing people have the same condition) Where geographical distance is a problem, telephone or e-mail contact may still be helpful Carer groups are very valuable for support and practical help

From our work, a core element was the recognition that the carer is often the expert on the person with dementia and they need to feel that this is recognised. Although people want to be told what the clinician feels would be best for them, they also expressed concern about doctors treating what they know best, rather than following patient and carer preferences, or seeking help from other types of expert. Carer diaries were often mentioned as a useful way of dealing with these issues. Opinions were divided as to whether or not the patient and carer should be seen together by the clinical team, separately or offered both. Some issues, such as increased carer stress, which is high in LBD, probably need to be explicitly discussed with the carer on their own.

Although the advantages of developing and working through a problem list were recognised, it was also said that a long list of problems could be demoralising, something that is particularly relevant for LBD where there are often many symptoms, not all of which may need to be formulated as problems. It was noted as helpful that an expert experienced in LBD is often able to offer reassurance that these symptoms can be expected as part of the disorder and not part of another illness.

Carer and peer support groups are very valuable for support and practical help, providing people have similar conditions. Support groups are one source of practical information at an early stage on important general matters, including power of attorney, attendance allowance and council tax. Where geographical distance is a problem, telephone or e-mail contact may still be helpful.

A clear message throughout the PPI consultations was that although they appreciated the levels of care and concern that clinical teams offered, they often had difficulty in obtaining and understanding information about LBD. A lack of post-diagnostic support, advice and counselling was frequently mentioned, as was a shortage of easily accessible materials. As a result, the guideline that we produced contained links to what were judged to be the best-quality information sources at the time of writing, and it is hoped that these will be made freely available to those attending our NHS services and that they will, in turn, be replaced by improved versions and adapted for local use.

Introduction

Work package 3 was developed because the accurate recognition of LBD is crucial to ensuring appropriate management and, therefore, central to improving patient care in dementia. 26–28

Prior to this programme, there was no single simple-to-administer toolkit that incorporated assessment of all the symptoms needed to make a diagnosis of DLB and PDD. Therefore, WP3 aimed to develop such assessment toolkits to improve the recognition and diagnosis of both DLB and PDD.

Methods

We developed the final LBD assessment toolkits in three stages: (1) identifying validated assessments for different symptoms and signs needed for a LBD diagnosis to make a pilot instrument, (2) obtaining feedback on the acceptability and feasibility of the toolkit through interviews with clinicians and the programme PPI group, and (3) piloting the toolkits in memory and movement disorder services prior to producing a final version to be used in WP4.

Methods

We included 11 services in four NHS trusts in the North East and 12 services in four NHS trusts in East Anglia. The trial was supported by the Newcastle Clinical Trials Unit, which undertook the randomisation via a statistician blinded to other aspects of the study. One service in the North East was subsequently unable to recruit any patients and withdrew part-way through the study. This study was approved by the West Midlands – Coventry and Warwickshire Research Ethics Committee (reference 16/WM/0025).

Assessment toolkits and, for services randomised to the intervention arm, the management toolkit were introduced during an in-person site initiation visit undertaken by the research team. The site initiation visits comprised standardised presentations and handouts followed by a question and answer session. The site initiation visits included as many clinical team members as were able to attend and all those involved in the diagnosis and management of those with DLB and PDD were encouraged to use the diagnostic and management toolkits. Follow-up support and further sessions were provided as required, and the study team maintained regular contact with all services during the course of the recruitment and follow-up period.

Assessment and management toolkits were provided as paper copies, with laminated copies of the overview and symptom summary sheets for the management toolkit. Some sites requested electronic [Portable Document Format (PDF)] versions and, where these were requested, they were supplied. If services had questions regarding the assessment or management toolkits, they were free to ask members of the research team. Regular trial newsletters that contained a ‘frequently asked questions’ section were sent to services to facilitate engagement with the study. Questionnaires regarding the value and use of the management toolkit were sent to services in the intervention arm halfway through the study and at the study end. Qualitative interviews (see Work package 5: qualitative studies throughout the DIAMOND-Lewy programme) were conducted to ascertain barriers to and facilitators of the implementation of both the diagnostic and management toolkits.

In all services, patients with DLB and PDD were recruited for data collection at baseline, 3 months and 6 months. A carer/informant was recruited for all but two patients. The full study protocol, which includes all the assessments undertaken, is available on the DIAMOND-Lewy website. 36 In brief, assessments included those of cognition (MMSE and MoCA), neuropsychiatric symptoms (Neuropsychiatric Inventory and Geriatric Depression Scale), fluctuations (Cognitive Fluctuation Scale), activities of daily living (Bristol Activities of Daily Living Scale), motor symptoms [Unified Parkinson’s Disease Rating Scale Part 3 (UPDRS)] and quality of life [DEMQOL (Dementia Quality of Life) and EuroQol-5 Dimensions, five-level version (EQ-5D-5L)]. Current management and use of health care and personal/social care services were collected on a bespoke questionnaire. Global outcome scales were completed by the visiting researcher and the carer/informant. Informant assessments included quality of life of the carer (EQ-5D-5L), carer burden (Zarit Burden Interview) and the Hospital Anxiety and Depression Scale.

All baseline and 3- and 6-month assessments were undertaken by members of the Clinical Research Network DeNDRoN teams in the North East and East Anglia, who were unaware of the service allocation (to management toolkit or standard care).

Data entry was completed by those undertaking the assessments, and central checking and quality assurance followed double-entry at local sites with an error rate of < 0.1% (all errors noticed were, of course, corrected). Database integrity was therefore felt to be very good. Identical shell databases were completed in the North East and East Anglia and then merged and sent to the study statistical team.

Results

One hundred and thirty-one participants consented to take part in the study. A CONSORT (Consolidated Standards of Reporting Trials) flow diagram, indicating patient progression through the study, is shown in Figure 3. One hundred and twenty-seven participants underwent a baseline assessment, and 6-month data were available for 109 participants (86% of those who completed baseline, 83% of all participants who consented). Eighteen participants were lost to follow-up, seven of whom died (four in the intervention arm, three in the control arm). Those lost to follow-up were similar in demographic characteristics to those remaining in the study except for age, with those lost to follow-up being significantly older (median age of 83 vs. 77 years).

FIGURE 3.

DIAMOND-Lewy WP4 CONSORT flow diagram of participants’ completion rate (27 March 2019). CRF, case report form.

The recruitment target was more than achieved. We aimed to recruit 120 patients and actually recruited 131. Patient characteristics at baseline are shown in Table 4. Participants randomised to receive the intervention did not differ significantly from those randomised to the control group on any of the baseline measures except for carer-reported DEMQOL and carer anxiety symptoms as assessed on the Hospital Anxiety and Depression Scale.

Health economics

The main objective of the economic component of the pilot RCT was to rehearse the methods for a future definitive economic evaluation of the new management toolkit compared with the usual care provided for dementia. As is typical for pilot trials, the modest sample size meant that a full economic evaluation was not appropriate because estimates would be both imprecise and unreliable. Therefore, the focus of the economic component was restricted to provide a descriptive analysis of the costs and outcomes. Data on costs and outcomes for each patient sample (DLB and PDD) were examined separately, as it was expected that the different diagnoses would lead to a differential use of services and outcomes.

Discussion

This study demonstrated the feasibility of successfully recruiting a large number of participants with DLB and PDD over an 18-month period from 23 sites, and we more than achieved the recruitment target (target, n = 120; recruited, n = 131). Follow-up rates were good (85%), given the condition under study, as LBD is known to be associated with more rapid functional decline and increased mortality than other dementias. 11,40 Completion rates of assessments for both patients and carers were high.

Using results from this study to inform a power calculation (80% power, alpha 0.05) gives a required sample size of 726 for outcome based on the Neuropsychiatric Inventory {18.8 [standard deviation (SD) 19.3] in the intervention and 22.6 [SD 16.3] in the control group}. Assuming 80% completion rates, the recruitment sample would need to be 908 for the Neuropsychiatric Inventory. However, given that the toolkits are already evidence based and associated with some positive outcomes, even in this pilot study, it is not clear that a larger trial based around demonstrating efficacy of the toolkits is needed. Instead, an approach focused more on their routine implementation, through service improvement or other approaches, may be preferable.

Taking account of the cost of implementing the management toolkits, the health economic analysis indicated that the total costs for both DLB and PDD increased over the course of the study in the control arms and decreased in the intervention arms, despite a slight increase in medication costs in intervention groups. With such a modest sample size, data were insufficient to draw conclusions, although there is no indication that the intervention increased costs. Data indicated that further study on health economic consequences in a larger sample is needed.

The qualitative studies (see Work package 5: qualitative studies throughout the DIAMOND-Lewy programme) indicate that implementation and use of the management toolkit varied considerably between individual clinicians and between services. Limitations of the study are that we were not able to standardise or measure this, nor – because this was randomised at a cluster level – could we directly associate the impact of the management toolkit on patient outcomes at the individual patient level. Further study should investigate these issues, but our results strongly support the need for a large, definitive trial of the management toolkit and indicate that its introduction is not associated with increased costs.

Methods

We used ethnographic methods, including semistructured interviews, focus groups and observations of clinical practice, throughout WP5. Structured questionnaires were used to collect information on current practice, confidence in diagnosing and managing LBD, and (in WP5.3) views on the assessment and management toolkits (Table 5).

Results

Stakeholder perspectives on the acceptability of the toolkits (work package 5.2)

Clinician views

Clinicians generally held positive views towards the concept of the toolkits. Potential benefits of the assessment toolkits included increased awareness of core and suggestive features of LBD and a more consistent and systematic approach to assessment. The management toolkit successfully translated academic papers to a format and style suitable for practice:

I think the good thing about it is it’s all in one place, and it’s really clear. The instructions and the advice are really clear, which certainly, when I was training, Parkinson’s and Lewy body dementia always seemed quite complicated illnesses to look after. You felt like you might not know quite what the current guidance was. Whereas, at least with that, it’s a one-stop shop, which tells you everything that you need to know.

0701, consultant OAP

Feedback on the assessment and management toolkits covered a range of areas (for illustrative quotations see Appendix 9, Tables 27 and 28). The majority of the feedback on the management toolkit related to the version used in the pilot trial, as the format changed substantially after the feasibility study. Issues relating to the assessment toolkits (already covered in Work package 3: development of assessment toolkits for the diagnosis of Lewy body dementia) are not reiterated here, with the exception of question wording where additional comments were made during the pilot trial.

Views on layout and presentation

The assessment toolkit for movement disorder services began with questions about relevant symptoms and ended with diagnostic criteria. In contrast, the toolkit for memory services began with the diagnostic criteria, followed by specific questions. Although few comments were made about the order of the PDD toolkit, some participants thought that the order of the DLB toolkit was counterintuitive. Furthermore, non-medical staff who did not make the diagnosis could find the first page off-putting.

Although few participants explicitly commented on signposting in the assessment toolkits, several did not realise that the diagnostic criteria in the DLB toolkit were followed by specific questions to explore core and suggestive features, or that instructions on how to perform and score the five-item UPDRS were provided. The structure of the PDD assessment toolkit was widely misunderstood, with many participants thinking that they were required to perform a formal cognitive test with all patients, whereas this was relevant only if responses to the initial questions indicated potential problems.

The management toolkit was organised in three levels (Figure 4). A one-page overview provided a brief, accessible summary of the symptom areas covered and the general principles for managing LBD. The colour coding of the overview then carried through to the next level, which comprised one-page summaries of recommendations for each of the five symptom areas. The final level comprised the full reference guidelines.

FIGURE 4.

Formats of the management toolkit.

Clinicians valued the colour coding, which facilitated navigation to different levels of detail. The majority of clinicians found the symptom summaries most useful, although some relied primarily on the overview.

Format

Participants suggested a number of alternative formats to facilitate use of the toolkits, most commonly integration with existing information technology systems or paperwork. Using the local intranet to display the toolkits automatically for patients with relevant diagnostic codes was also suggested. Applications were identified as an alternative way to ensure that the toolkits were easily accessible. For the assessment toolkit, participants suggested portable summary ‘prompts’ for each section (similar to the symptom summaries).

Several clinicians requested an A5 version of the overview and/or symptom summaries from the management toolkit to make it more portable. Others recommended displaying a poster of the overview to maintain awareness of the management toolkit. Another suggestion was to embed the management toolkit within an e-learning module for trainee doctors.

Some clinicians in movement disorder services suggested that the initial questions on memory and executive function from the assessment toolkit could form a separate document to be completed by patients and carers before their consultation. Other clinicians, however, thought that this would be too burdensome for patients and carers.

The common perception that the toolkits were more relevant to doctors led to suggestions that separate versions be produced for non-medical staff. This would enable non-medical staff to focus on those aspects of the toolkits most relevant to their work and skills (e.g. questions on core and suggestive criteria in the DLB toolkit and non-pharmacological interventions in the management toolkit). Although some consultants saw the assessment toolkits as more valuable to nurses or trainees than themselves, others thought that it was a useful reminder even for experienced staff. Some participants thought that the management toolkit would be useful in other settings, notably primary care or inpatient wards.

Content

Feedback on the content of the toolkits related to five themes, some of which were more relevant to specific toolkits.

Perceived relevance of content

The inclusion of the diagnostic criteria was typically seen as a useful reminder and prompt by medical staff. A number of clinicians, however, asked why key features they associated with LBD (e.g. falls) were not included in the assessment toolkits. Some clinicians in movement disorder services queried the absence of questions on hallucinations and fluctuations, as they saw these as essential for differentiating between LBD and other dementias in PD. Although the toolkit for movement disorder services was modified after the feasibility study, questions on hallucinations, REM sleep behaviour disorder and fluctuations were not integrated into the toolkit for patients with established PD.

Some clinicians queried the omission of key areas, such as carer support, psychosocial interventions and information for patients and carers, from the management toolkit. These comments suggest that the rationale for focusing on selected areas should be included in the toolkits.

Terminology and question wording

Comments on terminology and question wording related mainly to the assessment toolkits, particularly the DLB version. Unfamiliar terminology (such as ‘kinetic tremor’, ‘spontaneous parkinsonism’ and specific types of scans) was alienating, particularly for non-medical staff.

Although some clinicians emphasised the need for standardised questions, some questions were thought to be inappropriate for patients and carers, or to lack sensitivity. Less experienced clinicians valued the questions on REM sleep behaviour disorder, as they were sometimes unsure how to explore this symptom. Few comments were made about wording in the management toolkit, although greater clarity about incontinence (i.e. whether faecal or urinary) was suggested.

Details of drug dosages

Several clinicians valued being able to use the symptom summaries to check drug dosages. For drug treatments used infrequently, some clinicians suggested providing more detailed protocols:

Clozapine is such a complicated drug. People like psychiatrists are afraid to touch it. You put it in one line saying clozapine. There is some evidence for clozapine and I think clozapine should be written out in like almost a protocol how to use it [. . .] Nobody really knows how often you are supposed to do the blood test. When do you start and what is the best maximum dose. So, unless you de-mystify it and put it in there [. . .] I think people will be worried to [use it].

1901, consultant geriatrician

The recommendations regarding melatonin were inconsistent with the clinical experience of some participants, and several reported that GPs were reluctant to prescribe melatonin. To address this issue, one clinician suggested including an information leaflet to support prescribing in primary care.

Feasibility of the five-item UPDRS in memory services

Although few comments were made about the five-item UPDRS during the feasibility study, this was the most contentious aspect of the DLB assessment toolkit in the pilot trial. Three main issues with the five-item UPDRS were highlighted: (1) the rationale for focusing on these five items, (2) uncertainty over the value of scoring each item and (3) the challenges of introducing the scale in services where non-medical staff were responsible for assessments.

The toolkits used in the feasibility study and pilot trial gave no explanation for using the five-item UPDRS. Understanding that the items in the scale had been demonstrated to be useful in distinguishing between DLB and AD26 could have increased staff buy-in to the scale. Although the current version of the toolkits (v3) includes a brief statement, it is unclear whether or not this is enough to counterbalance the negative attitudes to this scale.

Clinicians frequently questioned the need to score each item on the UPDRS on a five-point scale, preferring to record simply the presence or absence of symptoms. Common arguments against scoring related to the difficulties in scoring accurately without extensive training, the additional time needed and the perception that scoring was more relevant to research than to clinical practice.

Although non-medical staff in memory services were often confident in informally assessing for parkinsonism, most saw physical examination as outside their area of expertise and role:

Because you haven’t got a full caseload of clients with Lewy body dementia you’re doing it occasionally. So, for me, I need to be doing something all the time for it to click. I’m fine with asking the questions, speaking to the carers and speaking to the patients, but when it comes to the actual of doing some of the assessment here I lack confidence.

0904, community nurse, memory service

Consultants had varied views on whether or not it was feasible and/or appropriate to train non-medical staff to perform the five-item UPDRS. Some were keen to train their colleagues, whereas others felt that physical examination was a medical responsibility.

Feasibility of formal cognitive assessment in movement disorder services

Some clinicians thought that their ongoing contact with patients facilitated identification of emerging cognitive problems, but others argued that symptoms could easily be missed without a formal assessment. Several clinicians in movement disorder services described (or were observed) using selected items from formal cognitive tests to explore cognitive impairment. Most clinicians in movement disorder services showed little interest in changing their existing approach often because of time constraints:

I just think as a concept it is really frightening to try and commit to doing it [the assessment toolkit] in a standard clinic. I think if you are going to do it, you would have to set up a stand-alone clinic [. . .] but that’s the problem isn’t it? The sort of balancing the desirable against the practical.

2402, consultant geriatrician

Patient and carer views of the assessment toolkits

A number of patients and carers invited to take part in an interview declined because they felt unable to comment on the assessment toolkit because they had been unable to distinguish it from the rest of the consultation.

Interviews with six patients and four carers confirmed that the assessment toolkit was ‘invisible’ to them. None of the questions from the assessment toolkit stood out as being inappropriate, unclear or in any way different from the rest of the consultation. In view of the consistency of these findings, data collection ceased before achieving the target sample size. Cognitive interviews with patients and carers highlighted some issues with question wording, typically with the same questions identified as problematic by clinicians.

Implementation of the toolkits (work package 5.3)

The implementation of the toolkits is summarised in Table 7, using the framework of NPT (for illustrative quotations see Appendix 10, Tables 29 and 30). Comments relating to all of the main NPT constructs were made about both the assessment and management toolkits and by staff working in memory and movement disorder services. Lack of investment in the assessment toolkits in many movement disorder services meant that few of these clinicians had experience relating to collective action or reflexive monitoring. They were able to comment, however, on both of these areas when discussing the management toolkit.

TABLE 7 - Implementation of the assessment and management toolkits

Making sense of the toolkits (coherence)	All toolkits	Translation of technical information into more practical formats Similarity to existing practice
	Assessment toolkit	Improve consistency
	Management toolkit	Confidence in robustness of toolkit
Investing in the toolkits (cognitive participation)	All toolkits	Perceived as more appropriate for ‘others’: Medical staff Less experienced/trained staff Other services
		Lack of investment as a team
Adopting the toolkits (collective action)	All toolkits	Lack of integration with existing paperwork or information technology system
	Assessment toolkits	Selective use when LBD suspected (memory services)
		Lack of resource for implementation
	Management toolkit	Insufficient alone to change practice
Assessing impacts of the toolkits (reflexive monitoring)	All toolkits	Increased awareness of LBD
	Assessment toolkits	Increased confidence in diagnosing LBD Increased exploration and documentation of core and suggestive features (memory services)
	Management toolkit	More holistic practice (memory services) Increased confidence Use to offer suggestions to GP for drug treatment or referrals

Making sense of the toolkits (coherence)

The concept of the toolkits was relatively straightforward to grasp and clinicians valued the translation of technical knowledge into practical tools. Key potential benefits of the assessment toolkits were improving the consistency of assessment (both within and between clinicians) and increasing diagnostic accuracy.

Although many clinicians reported that their current practice was similar to the assessment and management toolkits, this did not necessarily detract from their value. Most consultant and trainee doctors were confident in the recommendations in the management toolkit because of the expertise of the study team. Non-medics, however, were not necessarily familiar with the authors and some would have valued more detail on the background and references to source papers. The need to update the toolkits was emphasised by several participants, in part prompted by changes to the diagnostic criteria for DLB11 and to the NICE guidelines for management of PD50 during the pilot trial.

Investing in the toolkits (cognitive participation)

The generally positive views towards the toolkits did not, however, necessarily translate into a commitment to implementation. Consistent with earlier findings on the acceptability of the toolkits, they were often seen as more relevant to ‘other’ clinicians, typically junior doctors and non-medical staff. In addition, some clinicians queried whether or not the toolkits were relevant to their specialty, service or geographical location. Some movement disorder clinicians viewed diagnosis and management of LBD (particularly DLB) as more appropriate for memory services. Participants in some memory services argued that the management toolkit was of limited relevance, as their focus was primarily on diagnosis.

In three memory services, the lead clinician successfully made concerted and ongoing efforts to engage staff with the toolkits. In movement disorder services, such efforts were less evident, with only two services making efforts to engage some of the staff team. The majority of services, therefore, were characterised by a striking lack of collaborative work regarding the implementation of both toolkits. In these services, it was largely up to individual members of staff to decide if and how to use the toolkits:

For me, it was because we weren’t all invested as a team together. If you are trying to work with a team it is really important to know, ‘This is what we are all going to sign up to. This is what we are all going to do. This is how we are going to work together. This is how it will or won’t work’. You just feel as a team we never had those conversations about, ‘How can this be used in practice? How are we going to implement it?’

2602, specialist PD nurse

For some non-medical staff, the lack of perceived relevance of the toolkits limited their willingness to engage with the pilot trial.

Putting the toolkits into practice (collective action)

Implementation of both toolkits varied within and between services. Although the lack of integration with existing electronic or paper documentation was identified as a significant barrier to implementation, several clinicians reported incorporating some questions from the assessment toolkits into their routine practice:

What I’ve done is, actually, I’ve taken some of those criteria that are easy and then I have prepared my own clinic pro forma. So that when I see someone I have a piece of paper and the kind of things I need to ask. And part of my standard assessment question is to cover those criteria. That’s how I use them.

0102, trainee OAP

Although views varied on the additional time required to use the assessment toolkits, most participants in memory services used the toolkits only when LBD was suspected, and often did so after a consultation or during supervision, rather than embedding the questions into routine practice. The training on the five-item UPDRS provided by the clinical research associates to staff in some services in the pilot trial was well received, but insufficient to change practice on its own. The questionnaire explored the extent to which respondents explored core and suggestive features of DLB in consultations with new patients. Although over three-quarters of respondents from memory services reported that they always conducted a cognitive test and explored fluctuation and hallucinations, fewer than half explored REM sleep behaviour disorder, parkinsonism or used diagnostic criteria (see Appendix 7, Figure 10). Use of the assessment toolkits was limited in movement disorder services, with most services continuing with their usual approach to exploring cognition, although several individuals reported incorporating the initial questions on executive function into their consultations.

The management toolkit was similarly typically used after consultations, partly because of the logistics of accessing the documents but also because of concerns over patient reactions. Some clinicians appeared reluctant to change their practice purely on the basis of the management toolkit, preferring to continue with tried and tested prescriptions (even if the evidence for these was less robust):

We had that discussion this morning about clozapine being the most helpful but we don’t tend to use that, we tend to go for quetiapine.

Unidentified female participant, memory service 07

The questionnaires also explored views on the likelihood of embedding the toolkits into routine practice. Similar views were expressed about integration and sustainability of the assessment and management toolkits, although respondents had slightly more positive views about the management toolkit becoming part of their normal practice (see Appendix 7, Figure 11). Data also indicated that more positive views towards integration of the assessment toolkit were associated with having access to the management toolkit (see Appendix 7, Figure 12) and working in memory services (see Appendix 7, Figure 13).

Assessing impacts of the toolkits (reflexive monitoring)

For many clinicians, a key outcome of both toolkits was increased awareness of LBD. Where lead consultants in memory services had encouraged implementation of the DLB assessment toolkit, they reported improved documentation of core and suggestive features by non-medical staff. Other changes resulting from use of the assessment toolkit identified by staff in memory services included more consistent documentation of negative findings and a more holistic approach to assessment. Even when non-medical staff did not perform the five-item UPDRS, the assessment toolkit was thought to have increased awareness of the range of possible signs of parkinsonism.

We tried to identify possible mechanisms through which the management toolkit could lead to improved outcomes. Qualitative data from clinicians suggested that only a small number had adopted new treatments and there was little evidence of increased management of new symptom areas. We included questions on confidence in managing common LBD symptoms in all three WP5 questionnaires administered during the programme (see Appendix 7). Comparison over time indicated that confidence in managing three symptom areas (autonomic symptoms, parkinsonism and REM sleep behaviour disorder) increased significantly in control practices during the programme (Figure 5). Clinicians who received the management toolkit showed small, but (statistically) significant, changes in confidence in managing all but two symptom areas (Figure 6). This suggests that a possible mechanism of change may be the accumulation of marginal gains in confidence over a wide range of areas. However, the results should be interpreted with caution, as the respondents at each time point are overlapping, but not identical.

FIGURE 5.

Mean confidence ratings for symptom management for respondents in control services. *p < 0.05, **p < 0.01, ***p < 0.001. T0, initial data; T1, follow-up data 1; T2, follow-up 2.

FIGURE 6.

Mean confidence ratings for symptom management for respondents in intervention services. *p < 0.05, **p < 0.01, ***p < 0.001. T0, initial data; T1, follow-up data 1; T2, follow-up 2.

An alternative mechanism may be that change is achieved via indirect routes. For example, some clinicians reported that they were now more willing to suggest possible changes of medication or referral to other specialties to the patient’s GP. This more holistic approach to management is potentially a mechanism through which change could be achieved:

It gives guidelines for things that I wouldn’t consider treating in a memory service [. . .] So, I’m not really starting medications for constipation and even things like hypotension, but they might trigger me to ask the falls clinic or somebody else to see people.

1601, consultant OAP

Views on implementation

The insights from the qualitative work in WP5.1 and WP5.2 were used to develop an implementation strategy for the pilot trial (see Appendix 11, Tables 31–34). The extent to which this was implemented during the pilot trial was limited because of the restricted time and resources available, concerns over the risk of challenging and potentially alienating professionals, availability of credible professionals to address some issues (particularly for movement disorder services) and lack of skills within the research team to develop local implementation strategies.

During interviews and focus groups, we explored participant views on implementation in the pilot trial and alternative ways to support toolkit use. The interviews and focus groups were often the first opportunity for clinicians to reflect on if and how they were using the toolkits. Participants thought that implementation had been influenced by local leadership and tailored education sessions. In a small number of sites, a lead professional adopted the role of a local champion, typically acting as a focal point for the intervention, engaging the wider team and maintaining momentum (e.g. by adding the toolkits to the agenda of regular meetings). Some participants suggested that the lack of a lead clinician had been a factor contributing to poor implementation:

What I would say in retrospect, is I’m not sure that we’ve maximised the value of this within the service. But to do that, we would have actually needed a more sustained input, I think, from the team. And somebody, one or two of us, actually, really taking a role in terms of getting hold of this and running with it. You, actually, need somebody in the team who is going to take a real lead to try and sort of keep it on people’s minds. Find out what people’s experience was, address any issues, advocate for its value, which, to be fair, I don’t think there was somebody who took on that role.

1209, consultant OAP

The above quotation also mentions the need for sustained input from the research team. Tailored educational sessions were delivered by the study team to some sites (usually at their request). Feedback on the sessions indicated that they were valued by practitioners and often provided new insights into the assessment or management of LBD:

We had your colleague who came along and did the training on a Wednesday morning – which was excellent, I learned loads from that [. . .] [He] was much better at assessing for increased tone and did a far greater range of movement than I would have been routinely doing.

1101, consultant OAP

Although the site initiation visit provided an overview of the study and toolkits, several participants would have preferred a more interactive approach in which they could have reviewed the toolkits in detail with a knowledgeable clinician from the study team. This would have allowed queries to be raised and discussed and could explicitly have addressed the misperception among many movement disorder clinicians that the assessment toolkit required them to perform a cognitive assessment on all patients. A final strategy suggested by several participants was to use team meetings or peer learning groups to identify divergences between their current practice and the assessment or management toolkits as a way of highlighting scope for change.

Discussion

The detailed qualitative work provided new insights into the complex and inter-related factors influencing the diagnosis and management of LBD. The inherent complexities of the disease itself required input from different specialties, leading to potential fragmentation of services. Other important factors identified were a workforce with variable levels of training and confidence in LBD, and persisting negative attitudes among some clinicians towards diagnosis and disclosure of dementia in patients with pre-existing PD.

Observation confirmed the lack of routine assessment of core and suggestive features of LBD in memory services. In movement disorder services, resources were a key factor in determining whether or not cognitive function was formally assessed. The assessment toolkits potentially address these issues by including standardised questions for memory services and brief screening questions for movement disorder services to determine whether or not formal cognitive assessment is required. The management toolkit could potentially facilitate management of the complex and multiple symptoms in LBD. Other barriers to diagnosis and management were not amenable to change through the toolkits and some acted as barriers to their implementation.

Most clinicians welcomed the concept of the toolkits to facilitate the diagnosis and management of LBD. Key potential benefits were increased awareness of LBD, more consistent and comprehensive assessment and evidence-based management. The assessment toolkits were thought to be particularly beneficial to junior doctors and non-medical staff.

Qualitative data highlighted aspects of the assessment toolkits requiring further revision. It is now clear that training non-medical staff to administer the five-item UPDRS26 is neither feasible nor acceptable. However, with nurse-led models of memory assessment services becoming more common, alternative ways to conduct a simple neurological screen for non-medical staff are needed. Other issues related to question wording, signposting and the rationale underpinning the diagnostic criteria and choice of symptoms.

The positive feedback on the presentation of the management toolkit suggests that applying the same design principles to the assessment toolkits may make them more appealing and easier to use in practice. Ideally, the toolkits would be available in a variety of formats to suit personal preferences.

Implementation of the assessment toolkits in the pilot trial varied within and between sites. In many movement disorder services, key barriers related to cognitive participation or willingness to invest in the assessment toolkits. This resulted in limited use of the assessment toolkits (which is likely to explain the unchanged diagnostic rates of PDD observed in WP1R, see Work package 1 repeated: re-assessment of diagnostic rates for Lewy body dementia after introduction of the assessment toolkits for diagnosis). In three memory services in the North East, the lead clinician successfully encouraged widespread adoption throughout the team. Although this was attempted in two further memory services (both in East Anglia), uptake in these teams remained patchy. In other memory services, there was little collective work to support implementation and it seemed largely up to individual members of staff to decide if and how to implement the toolkits.

Uptake of the management toolkit varied, with some participants reporting that their practice was already consistent with the toolkit. Barriers to implementation were that few LBD patients were seen and that (memory) services focused on diagnosis with little involvement in management. Identifying possible mechanisms through which the management toolkit could lead to improved outcomes was challenging. Few participants reported changing their prescribing habits or managing additional symptom areas. However, the use of the toolkit appeared to raise awareness of the range of symptom areas affected in LBD and to increase confidence in managing common symptoms, and this may have facilitated a more holistic approach to care. In any future implementation study, we recommend the collection of data on prescriptions and referrals, and content analysis of letters to GPs, as it is possible that some benefits accrue from changes made by GPs, rather than directly by the specialist clinicians using the management toolkit.

Methods

We used exactly the same methods as in WP1 (see Work package 1: baseline study of the diagnosis and management of Lewy body dementia) to ensure comparability of findings. For DLB, case notes of patients in four of the same memory services in three different NHS trusts from the two regions were examined. For PDD, case notes were reviewed from three movement disorder services in three NHS trusts in the two regions.

Results

Of 2058 referrals to the memory services over an 18-month period, 1279 received a diagnosis of dementia, of whom 6.2% received a diagnosis of DLB.

This rate, found after introduction of the assessment toolkits, was significantly higher than the 4.6% found at baseline before the introduction of the toolkits (χ² = 5.3; p = 0.021). On further examination, diagnostic practice in East Anglia was unchanged, whereas rates had significantly increased in the North East (9.7% vs. 5.6% at baseline, χ² = 14; p = 0.000019).

A total of 3405 referrals to movement disorder services were identified, of whom 1968 received a diagnosis of PD. Of these, 8.2% received a diagnosis of PDD. This diagnostic rate, after introduction of the assessment and diagnostic toolkits, was not significantly different from the 9.7% found at baseline (χ² = 2.8; p = 0.09). For PDD, there was a difference between regions, with the North East showing a significant decrease in PDD rates (7.8% vs. 10.5%; p = 0.006); however, East Anglia showed a significant increase (13.3% vs. 8.3% at baseline; p = 0.046).

Discussion

The introduction of assessment and diagnostic toolkits for DLB and PDD was associated with a significant increase in diagnostic rates for DLB, but not PDD. However, there were significant differences between services and regions, with some services increasing their diagnostic rates and others not. Differences between regions were also noted for WP1 in terms of baseline rates for DLB diagnosis, although it is not entirely clear what factors drive these differences. There was diversity both within and between North East and East Anglia in terms of service organisation. Both sites included memory services where initial assessments of ‘straightforward’ patients were conducted by non-medical staff. Access to scans in the North East appeared to be better. In East Anglia, issues with scans reflected both the significant travelling required by some patients in rural areas and the informal limits described by some participants on the ‘appropriate’ number of scans. The North East has a long track record of research in DLB and many clinicians working in the North East have been exposed to DLB, either during training and/or through their involvement in research studies. This was reflected in a number of comments made by participants in the qualitative work on the North East being particularly sensitised to DLB. There were, however, clinicians in East Anglia with a long-standing interest in DLB who had similarly been involved in training medical staff over a period of years. Interest in DLB was therefore by no means limited to the North East.

The strengths of this study include the inclusion of the same services for which we ascertained baseline diagnostic rates in WP1, and use of exactly the same methodology for determining cases and diagnosis. Limitations include that, as this was not a controlled study, we cannot definitely conclude that the introduction of the assessment toolkits was causally related to the change in diagnostic rates. Other factors may have been involved, including generally heighted awareness of DLB over time and/or service or other changes. Finally, unlike WP1, in WP1R we did not undertake a repeat of the more detailed case note study and so we could not determine whether or not the factors associated with the diagnosis of DLB and PDD found at baseline had been moderated in any way. We did not collect data on the additional time taken to administer the toolkits and so cost-effectiveness of the toolkits for increasing diagnostic rates cannot be determined.

The qualitative studies (WP5) provide some insights as to why diagnostic rates for PDD (and DLB in East Anglia) were unchanged, but further research is required to better understand whether the unchanged diagnostic rates were linked to variation in implementation or whether additional factors were involved. Further exploration of ways to tackle the barriers to diagnosis identified in WP5 and to improve implementation of the assessment toolkit, especially in sites with low diagnostic rates at baseline, is required.

Implications for practice and any lessons learned

The findings of the programme have a number of important implications for clinical practice:

• In the regions we studied, DLB and PDD appeared to be underdiagnosed compared with expected rates, with variability in diagnostic rates between services. This suggests that improvements may be needed in the way in which clinicians assess people for symptoms and make diagnoses.

• We found that the reluctance of some clinicians to make a formal diagnosis of dementia could be a significant factor contributing to the underdiagnosis of LBD (particularly PDD). This implies that negative attitudes to disclosure may need to be challenged, possibly through supervision, appraisal or local audits to examine the diagnostic rate of PDD.

• In eliciting symptoms, observation suggested that leading questions were not uncommon and that some symptom areas (e.g. sleep) were often explored only superficially during memory assessments. Existing clinical practices, such as supervision, may provide an opportunity to address both of these areas. An iterative process of discussing and sharing ways of asking questions may enable staff to critically evaluate their existing approach and develop more consistent ways of approaching certain symptoms or issues. The concerns expressed over whether or not non-medical staff had received adequate training for their role in assessment imply that additional on-the-job training on dementia subtypes may improve diagnostic accuracy.

• There may be benefits for more training to enhance awareness in secondary care clinical teams of how to make a diagnosis, and of why, including raising awareness of the opportunities for evidence-based management.

• Using a structured method, like the assessment toolkit, for diagnosis may help increase diagnostic rates.

• The complexities of managing LBD may require input from a range of specialties, which has the potential to lead to fragmentation of care pathways. Successful strategies for joint working included telephone multidisciplinary team meetings for complex patients, informal consultation arrangements across specialties where clinicians were co-located and adapting a memory service pathway to enable patients diagnosed with LBD by a geriatrician or neurologist to access support without further assessment. These initiatives highlight the potential value of developing tailored, local strategies to address barriers to cross-specialty working.

Recommendations for research

Findings from the programme highlight a number of areas for future research:

• Further study to ascertain whether or not unchanged diagnostic rates, despite access to the toolkits, reflect poor implementation of the assessment toolkits or additional factors.

• Research into how best to co-ordinate multispecialty input to patients with LBD is needed to streamline management and facilitate a holistic approach.

• Further work is needed to better understand how assessment toolkits for PDD can be integrated into practice to improve diagnostic rates. Qualitative studies indicate that there remain important barriers and negative attitudes to diagnosis and management, and further work is needed to see how these are best addressed.

• The evidence base informing the management of LBD is limited, especially for non-pharmacological interventions. More therapeutic studies are needed, especially well-designed RCTs for both cognitive and non-cognitive symptoms.

• Future research should investigate the specific domains contained within the management toolkit and the extent to which global improvements are due to specific symptom improvement, as opposed to a number of marginal gains in several areas.

• Our successful pilot demonstrates that a larger, more comprehensive trial of introducing the management toolkit could be undertaken, but it would need to include a minimum of 410 (and up to 908) people, depending on the primary outcome.

Contributions of authors

John T O’Brien (https://orcid.org/0000-0002-0837-5080) was chief investigator of the programme, lead for WPs 1A, 1R and 4A, generated the first draft of this report, commented and made suggested revisions and subsequently approved the final version of the report.

John-Paul Taylor (https://orcid.org/0000-0001-7958-6558) was co-lead for WP2, generated the first draft of this report, commented and made suggested revisions and subsequently approved the final version of the report.

Alan Thomas (https://orcid.org/0000-0002-6667-9533) was lead for WP3, generated the first draft of this report, commented and made suggested revisions and subsequently approved the final version of the report.

Claire Bamford (https://orcid.org/0000-0003-2885-801X) was co-lead for WP5, led and supported the PPI work, generated the first draft of this report, commented and made suggested revisions and subsequently approved the final version of the report.

Luke Vale (https://orcid.org/0000-0001-8574-8429) led the health economic analysis, generated the first draft of this report, commented and made suggested revisions and subsequently approved the final version of the report.

Sarah Hill (https://orcid.org/0000-0002-5408-2473) undertook the health economic analysis, generated the first draft of this report, commented and made suggested revisions and subsequently approved the final version of the report.

Louise Allan (https://orcid.org/0000-0002-8912-4901) advised on various aspects of study design and strategy, commented and made suggested revisions and subsequently approved the final version of the report.

Tracy Finch (https://orcid.org/0000-0001-8647-735X) was co-lead for WP5, commented and made suggested revisions and subsequently approved the final version of the report.

Richard McNally (https://orcid.org/0000-0001-6685-6467) led statistical aspects of the study, commented and made suggested revisions and subsequently approved the final version of the report.

Louise Hayes (https://orcid.org/0000-0001-6685-6467) supported statistical aspects of the study, generated the first draft of this report, commented and made suggested revisions and subsequently approved the final version of the report.

Ajenthan Surendranathan (https://orcid.org/0000-0003-3809-1545) undertook data collection and analysis for WPs 1A, 1B, 1R, 4A and 4B, commented and made suggested revisions and subsequently approved the final version of the report.

Joseph Kane (https://orcid.org/0000-0002-8479-9977) undertook data collection and analysis for WPs 1A, 1B, 1R, 4A and 4B, commented and made suggested revisions and subsequently approved the final version of the report.

Alexandros E Chrysos (https://orcid.org/0000-0002-4568-4202) undertook the health economic analysis, commented and made suggested revisions and subsequently approved the final version of the report.

Allison Bentley (https://orcid.org/0000-0001-9673-580X) undertook data collection and analysis for WPs 1A, 1B, 1R, 4A and 4B, commented and made suggested revisions and subsequently approved the final version of the report.

Sally Barker (https://orcid.org/0000-0002-2981-3748) undertook data collection and analysis for WPs 1A, 1B, 1R, 4A and 4B, commented and made suggested revisions and subsequently approved the final version of the report.

James Mason (https://orcid.org/0000-0001-9210-4082) advised on various aspects of study design and strategy, commented and made suggested revisions and subsequently approved the final version of the report.

David Burn (https://orcid.org/0000-0001-7658-1209) was lead for WPs 1B and 4B, commented and made suggested revisions and subsequently approved the final version of the report.

Ian McKeith (https://orcid.org/0000-0002-9250-0568) was co-lead for WP2, led and supported the PPI work, commented and made suggested revisions and subsequently approved the final version of the report.

Publications

Stinton C, Bamford C, Cambridge V, Lafortune L, Mak E, Mason J, et al. A Systematic Review of Management Strategies for Lewy Body Dementia (Dementia with Lewy Bodies and Parkinson’s Disease Dementia). PROSPERO 2014:CRD42014007180. URL: www.crd.york.ac.uk/PROSPERO/display_record.php?RecordID=7180 (accessed April 2021).

Stinton C, McKeith I, Taylor JP, Lafortune L, Mioshi E, Mak E, et al. Pharmacological management of Lewy body dementia: a systematic review and meta-analysis. Am J Psychiatry 2015;172:731–42. https://doi.org/10.1176/appi.ajp.2015.14121582

Thomas AJ, Taylor JP, McKeith I, Bamford C, Burn D, Allan L, O’Brien J. Development of assessment toolkits for improving the diagnosis of the Lewy body dementias: feasibility study within the DIAMOND Lewy study. Int J Geriatr Psychiatry 2017;32:1280–304. https://doi.org/10.1002/gps.4609

Connors MH, Quinto L, McKeith I, Brodaty H, Allan L, Bamford C, et al. Non-pharmacological interventions for Lewy body dementia: a systematic review. Psychol Med 2018;48:1749–58. https://doi.org/10.1017/S0033291717003257

Kane JPM, Surendranathan A, Bentley A, Barker SAH, Taylor JP, Thomas AJ, et al. Clinical prevalence of Lewy body dementia. Alzheimers Res Ther 2018;10:19. https://doi.org/10.1186/s13195-018-0350-6

Thomas AJ, Taylor JP, McKeith I, Bamford C, Burn D, Allan L, O’Brien J. Revision of assessment toolkits for improving the diagnosis of Lewy body dementia: the DIAMOND Lewy study. Int J Geriatr Psychiatry 2018;33:1293–304. https://doi.org/10.1002/gps.4948

Surendranathan A, Kane JP, Bentley A, Barker SA, Taylor JP, Thomas AJ, et al. Clinical diagnosis of Lewy body dementia. BJPsych Open 2020;6:1–8.

Taylor JP, McKeith IG, Burn DJ, Boeve B, Weintraub D, Bamford C, et al. New evidence on the management of Lewy body dementia. Lancet Neurol 2020;19:157–69.

O’Brien JT, McKeith IG, Thomas AJ, Bamford C, Vale L, Hill S, et al. Introduction of a management toolkit for Lewy body dementia: a pilot cluster randomised trial. Mov Disord 2021;36:143–51.

Surendranathan A, Kane J, Bentley A, Barker S, McNally R, Bamford C, et al. Introduction of an assessment toolkit associated with increased rate of DLB diagnosis. Alzheimer Res Ther 2021;13.

Data-sharing statement

We shall make data available to the scientific community with few restrictions as feasible, while retaining exclusive use until the publication of major outputs. Anonymised data can be obtained by contacting the corresponding author.

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, CCF, NETSCC, PGfAR 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 PGfAR programme or the Department of Health and Social Care.

Comparison of health-care service utilisation and costs between Lewy body and non-Lewy body dementia patients: the DIAMOND-Lewy study

PROSPERO: international prospective register of systematic reviews

Citation

Stinton C, Bamford C, Cambridge V, Lafortune L, Mak E, Mason J, et al. A Systematic Review of Management Strategies for Lewy Body Dementia (Dementia with Lewy Bodies and Parkinson’s Disease Dementia). PROSPERO 2014:CRD42014007180. URL: www.crd.york.ac.uk/PROSPERO/display_record.php?RecordID=7180.

Review question(s)

What are the benefits, harms and costs of management strategies for Lewy body dementia delivered to (1) patients and (2) carers?

How, when and where should management strategies for Lewy body dementia be implemented? What views do patients and carers have of management strategies for Lewy body dementia?

Searches

We will search the following electronic bibliographic databases: AgeInfo, Allied and Complementary Medicine, Applied Social Sciences Index and Abstracts (ASSIA), BiblioMap, British Nursing Index (BNI), Campbell Collaboration, CenterWatch, ClinicalTrials.gov, Cochrane Central Register of Controlled Trials (CENTRAL), Cochrane Database of Abstracts of Reviews of Effects (DARE), Database of Systematic Reviews, Cochrane Dementia and Cognitive Improvement Group Specialised Register, Cumulative Index to Nursing and Allied Health Literature (CINAHL), Database of Promoting Health Effectiveness Reviews (DoPHER), EMBASE, Educational Resources Information Centre (ERIC), Health Services Research Projects in Progress, Health Technology Assessment database, IFPMA Clinical Trials Portal, Lilacs: Latin American and Caribbean Health Sciences database, MEDLINE, metaRegister of Controlled Trials, NHS Economic Evaluation Database (NHS EED), NHS Evidence, PROSPERO, PsycInfo, Scopus, Social Care Online, Trials Register of Promoting Health Interventions, UK Clinical Research Network Portfolio database, UMIN Clinical Trials Registry, WHO International Clinical Trials Registry Platform.

The search strategy will include terms relating to Lewy body dementia.

We will also search the grey literature (i.e. data reported outside peer-reviewed publications), reference lists of studies included in the review and seek expert input from within the extended review team.

There will be no restrictions on language or publication period.

Where the function is available, we will set up auto-alerts in databases to capture any new data arising after we have conducted our searches.

Types of study to be included

There are no restrictions on the types of study designs for this review, but opinion papers will be excluded.

Condition or domain being studied

Lewy body dementia (LBD) is the second most common cause of neurodegenerative dementia after Alzheimer’s, accounting for around 15% of cases. It is characterised by psychiatric symptoms (e.g. hallucinations), impairments and fluctuations in cognition, and parkinsonism. LBD comprises two related disorders: dementia with Lewy bodies (DLB) and Parkinson’s disease dementia (PDD). DLB and PDD are distinguished on the basis of the temporal sequence of the onset of motor symptoms; a diagnosis of PDD is applied when motor symptoms occur at least one year before dementia and a diagnosis of DLB when dementia precedes motor symptoms or they arise within one year of one another.

Participants/population

Patients: We will include study participants who have received a formal diagnosis of LBD on the basis of operationalised criteria or clinical judgement.

Carers: We will include participants who provide care (formal or informal) to people who have received a diagnosis of LBD.

Intervention(s), exposure(s)

Any management strategy, or combination of strategies, that are intended to address the needs (medical, social, economic, etc.) of people who have LBD or their carers. These will include, but not be limited to, pharmacotherapy, psychotherapy, physiotherapy, education, speech and language therapy, support services.

Comparator(s)/control

We will include studies with and without comparison/control groups.

Context

We will place no restriction on the settings in which managements strategies are delivered.

Outcome(s)

Data extraction, (selection and coding)

Risk of bias (quality) assessment

The assessment of quality of studies will be conducted independently by two reviewers using the Quality Assessment Tool for Quantitative Studies (Effective Public Health Practice Project, 1998), the NICE Methodology Checklist: Qualitative Studies (NICE, 2009), and the Drummond Checklist for Economic Evaluations (Drummond & Jefferson, 1996). Any disagreement between the reviewers will be resolved through discussion and, if necessary, by consulting a third reviewer. Where reviewers identify significant lack of quality, these studies will be excluded.

Strategy for data synthesis

Quantitative and qualitative data will be treated as separate streams in the initial phase of the review. Qualitative evidence will subsequently be used to aid in the interpretation of quantitative evidence.

We will distinguish between published and unpublished literature and assess the impact that publication status has on the results of the review using sensitivity analysis and/or descriptively.

Analysis of subgroups or subsets

If sufficient data are available, we will consider data for participants who have DLB and those who have PDD separately.

If sufficient data are available, we will conduct subgroup analyses based on dementia severity.

Dissemination plans

The findings will be disseminated to NHS professionals, NHS commissioners, Social Care, and patient/public organisations. Scientific findings will be published in academic journals and presented at conferences.

Contact details for further information

Chris Stinton

Organisational affiliation of the review

University of Cambridge, University of Newcastle, University of Durham.

Review team

Dr Chris Stinton, University of Cambridge.

Ms Claire Bamford, University of Newcastle.

Dr Victoria Cambridge, University of Cambridge.

Dr Louise Lafortune, University of Cambridge.

Mr Elijah Mak, University of Cambridge.

Professor James Mason, University of Durham.

Professor Ian McKeith, University of Newcastle.

Dr John-Paul Taylor, University of Newcastle.

Professor Alan Thomas, University of Newcastle.

Professor John O’Brien, University of Cambridge.

Anticipated or actual start date

16 September 2013.

Anticipated completion date

30 April 2014.

Funding sources/sponsors

NIHR Programme Grants for Applied Research (Project ref. DTC-RP-PG0311-12001) – Improving the diagnosis and management of neurodegenerative dementia of Lewy body type in the NHS (DIAMOND-Lewy).

Conflicts of interest

None known.

Language

English.

Country

England.

Subject index terms status

Subject indexing assigned by CRD.

Subject index terms

Humans; Lewy Body Disease.

Any other information

This review is being conducted as part of a programme of research to improve the care of people who have LBD within routine NHS secondary care.

Pharmacological management strategies21

Non-pharmacological management strategies24

Introduction

The main objective of the economic component of the pilot cluster RCT was to rehearse the methods for a definitive economic evaluation of the new management toolkit for DLB and PDD compared with usual care. For a definitive study, a within-trial economic evaluation is anticipated. Such a study would seek to estimate costs and outcomes for each trial participant over the duration of the trial follow-up. These data would then be used to estimate cost-effectiveness. As is typical for pilot trials, because of the modest sample size available for each randomised patient group, a full economic evaluation was not conducted. The small sample size would not provide reliable conclusions on the cost-effectiveness of the management toolkit. Therefore, the focus of the economic component deviated slightly from protocol and the analysis aimed rather to provide a descriptive analysis of the costs and outcomes. Data on costs and outcomes for each patient sample (DLB and PDD) were examined separately, as it was expected that different diagnoses would lead to a differential use of services and outcomes. Suggestions for reducing the number of missing data will be made and methods that would be employed in a full evaluation of a definitive trial will also be discussed.

Methods

Intervention delivery costs

The delivery of the intervention incurred costs from the production of the management toolkit and training staff to use it. The cost of producing the toolkit included materials and staff time. Unit costs for each element of production were obtained from invoices provided by the trial support team and service costs listed on the Newcastle University print services website. 60 The service costs stated by print services were assumed to include labour time.

The cost of training staff to use the toolkit was estimated from data collected on staff attending training sessions and approximate duration of the sessions, in addition to staff time required to develop training sessions. Unit costs for staff were obtained from publicly available pay scales for university staff and the Unit Costs of Health and Social Care 2018. 58

Results

Discussion

Robust conclusions cannot be drawn on the impact of the intervention on costs and quality of life, given the small sample size of the trial. In a definitive trial, the impact on cost would be more reliably assessed in conjunction with the impact on quality of life. The impact of the intervention on quality of life, indicated by this pilot study, is small and appears to differ between patient samples. As with costs, the impact of the intervention on QALYs cannot be reliably assessed from this sample.

Given the disease area being studied in this trial, a substantial number of missing data might be expected because of difficulties in collecting data from people with dementia. Within a definitive trial, precision of results could be improved through the use of statistical methods to replace missing data with appropriate values where some data collection was unsuccessful. Numerous methods exist for handling missing data. The choice between these methods depends on the nature and pattern of missing data.

A substantial number of quality-of-life data were missing from responses to the EQ-5D-5L. If participants are unable to complete the questionnaire themselves, as a sensitivity analysis, proxy utilities from EQ-5D-5L responses completed by a carer on behalf of the patient participant could be used. Statistical tests for correlation between proxy and self-reported EQ-5D-5L can be conducted to assess whether or not the proxy values are sufficiently similar to the self-reported values (where both are available). If the tests indicate that proxy and self-report EQ-5D-5L utilities are not strongly correlated, using proxy values would introduce bias to the outcome. In a base-case analysis, imputation may be more reliable if there is sufficient data for this to be a reliable option.

Data collection

In January 2015, the first questionnaire was sent by post to staff working in most participating services, along with a prepaid envelope. One reminder was sent, including a copy of the questionnaire and a prepaid envelope, after approximately 3 weeks. As one NHS trust was unwilling to release staff details, we had to rely on local contacts to distribute the questionnaires. Although some attempts were made to send reminders within this trust, the process was hampered by a lack of information on individual participants and limited local resources to support distribution of the questionnaires.

The intention was to distribute the second questionnaire (T1) shortly after completion of all site initiation visits for the pilot trial. As there was insufficient capacity within the team to send out questionnaires to each site as approvals were obtained, distribution was delayed until all approvals were in place. Unfortunately, approvals for the qualitative work were not necessarily processed in parallel with approval for the pilot trial and obtaining approval from one trust was a protracted process. As a result, by the time T1 questionnaires were sent out (in January 2017), some sites had been using the toolkits for some time. The final questionnaire (T2) was sent out after the end of the pilot trial, as planned (in October 2018). The T1 and T2 questionnaires were administered using SurveyMonkey^® (Palo Alto, CA, USA), with one electronic and one paper reminder.

Questionnaires completed on paper were entered into SPSS version 16 (IBM, New York, NY, USA). Questionnaires completed electronically were transferred to SPSS for analysis. All questionnaires were identified by a unique identifier.

Analysis

The questionnaires were analysed using simple descriptive statistics. Analysis of variance was used to explore the impact of potential explanatory variables. The reasons for selecting potential plausible explanatory variables are summarised in Table 20. It is evident that these potential explanatory variables are not independent, with region and role, in particular, being confounded.

There were difficulties in examining changes over time, as, although the respondents to the three questionnaires overlapped, they were not identical and only seven individuals completed questionnaires at all three time points. For the analysis of confidence in symptom management (for which information was available at all time points), we maximised the number of observations for analysis by constructing a file containing people who responded to the T0 questionnaire and either the T1 and/or T2 questionnaires (where respondents completed both questionnaires, the T2 response took precedence).

Response rates

Response rates to each questionnaire are summarised in Table 21. Undelivered questionnaires are excluded from calculation of the response rate. Overall response rates were similar for each of the questionnaires (ranging from 37% to 48%). Data from SurveyMonkey indicated that, overall, nearly two-thirds of questionnaires delivered electronically (62%) were unopened, highlighting the need to provide copies of the questionnaire by post as well as e-mail.

Data were available to compare response rates to each questionnaire by region and specialty. There was no evidence of differential response rates by region. Response rates were significantly higher for movement disorder than for memory services at T0 (p < 0.01) and T2 (p < 0.001), but not at T1 (Table 22). Response rates were examined by study arm for T1 and T2 questionnaires. There was no difference at baseline, but a trend for a better response rate from the control arm at T2 (54.5% vs. 31.8%; p < 0.05). This is likely to reflect the unequal distribution of larger teams between study arms.

Characteristics of respondents

Respondent characteristics at each time point are shown in Table 23. Respondents had a range of roles, although the majority were either consultants or nurses. To facilitate subsequent analyses, role was recoded into doctors (including trainees) and allied health professionals (including nurses).

At T0, information was collected on experience of working with patients with dementia and/or PD. This varied from < 1 to 40 years, with an average of 13 years’ experience with dementia (median 12 years) and just under 10 years’ experience of working with PD patients (median 9 years). The T1 and T2 questionnaires asked more specifically about the training in LBD received by respondents. The response categories were chosen on the basis of comments made by interview participants that emphasised the value of exposure to LBD through working with experts in the field. Data were combined to explore training received by each unique participant to T1 and/or T2 questionnaires (n = 80). A series of binary variables was constructed and we explored whether or not training varied systematically by professional background (doctors vs. others), specialty (movement disorder vs. memory services), region (East Anglia vs. North East) or study arm (intervention vs. control). Overall, around one-third of respondents (33.8%) had received only basic training, a similar proportion (32.5%) reported self-directed study (through reading or attendance at conferences) and nearly half (45.0%) had worked with a team with expertise in LBD (the totals do not sum to 100, as some respondents reported both self-directed study and working with an expert team).

No differences in training were found by specialty or study arm. Respondents in East Anglia were more likely to report only basic training than those in the North East (43.2% vs. 22.2%; p < 0.05). Although working with an expert team was not associated with professional background, doctors were more likely to report self-directed study and less likely to have received only basic training (Table 24). The findings suggest that those completing and returning the questionnaires may have considerable knowledge of and expertise in LBD, and highlight the need to provide additional training on LBD for non-medical staff.

Confidence in diagnosing selected conditions (T0 only)

A seven-point scale was used to assess confidence in diagnosing and managing a range of conditions commonly seen in memory and/or movement disorder services. A not applicable column was included, as some staff were not responsible for making diagnoses or for managing certain conditions (the number of respondents therefore varies for different conditions). Figure 7 presents the mean level of confidence (where 1 = low and 7 = high confidence).

FIGURE 7.

Confidence in diagnosing selected conditions by specialty at T0 (n = 98–120). *p < 0.01, **p < 0.001. AD, Alzheimer’s disease; FTD, frontotemporal dementia; HD, Huntingdon’s disease; MS, multiple sclerosis; PSP, progressive supranuclear palsy; VaD, vascular dementia.

With the exception of FTD, clinicians were, generally, less confident in diagnosing DLB and PDD than other types of dementia. There were significant differences between specialties in confidence in diagnosing six conditions (see Figure 7). Greater confidence in diagnosing predominantly physical conditions would be anticipated for staff in movement disorder services (i.e. multiple sclerosis, progressive supranuclear palsy, PD and stroke, all p < 0.001). Staff in these services also rated their confidence higher in diagnosing both PDD and DLB than those in memory services (p < 0.001 and p < 0.01, respectively).

Confidence in managing selected conditions (T0 only)

Parallel analyses were conducted for confidence in managing the same set of conditions (Figure 8). Significant differences between specialties were generally as would be anticipated. Professionals in movement disorder services were more confident in managing more physical conditions (multiple sclerosis, p < 0.01; progressive supranuclear palsy, PD and stroke, all p < 0.001), whereas those in memory services were more confident in managing some dementias, including frontotemporal dementia, mixed dementia, Alzheimer’s disease and MCI (all p < 0.001). Staff in both specialties reported similar levels of confidence in managing DLB and PDD. Examination of confidence in managing symptoms common in these conditions suggests a possible explanation for this.

FIGURE 8.

Confidence in managing selected conditions by specialty at T0 (n = 130–150). *p < 0.01, **p < 0.001. AD, Alzheimer’s disease; FTD, frontotemporal dementia; HD, Huntingdon’s disease; MS, multiple sclerosis; PSP, progressive supranuclear palsy; VaD, vascular dementia.

Confidence in managing symptoms common in Lewy body dementia (all time points)

The questionnaire also included a section on confidence in managing common symptoms in LBD (again rated on a seven-point scale). Figure 9 shows that both specialties have areas of expertise, but neither are confident in managing the full range of symptoms associated with LBD. The pattern of responses underpins the dual pathway experienced by many people with LBD.

FIGURE 9.

Confidence in managing common LBD symptoms by specialty at T0 (n = 136–150). *p < 0.01, **p < 0.001. RBD, rapid eye movement sleep behaviour disorder.

As this set of questions was included in all three questionnaires, responses at T0 were compared with those at either T1 or T2. Radar plots were used to illustrate changes in confidence in managing common LBD symptoms between the beginning and end of the programme for respondents in the control and intervention arms (see Figures 5 and 6). The plots indicated that confidence in symptom management reported by staff in intervention services later in the programme was significantly higher for all but three symptom areas than at the beginning of the programme. For these figures, we have included changes significant at a p-value < 0.05, as these suggest a trend towards greater confidence in some areas. Although, as previously emphasised, interpretation is hindered by the lack of matched samples, these findings suggest that including questionnaire data may help in understanding the mechanisms through which changes occur, particularly if larger samples are available with more consistent respondents at both time points. The findings lend some support to the hypothesis that the positive outcomes achieved in the pilot trial may result from the accumulation of marginal gains.

Approach to new assessments (T1 and T2 only)

The questionnaires administered during the pilot trial asked respondents how frequently they explored key LBD symptoms, completed formal cognitive assessments and used formal diagnostic criteria. These items related to the content of the assessment toolkits. As the focus of the toolkits varied for clinicians in movement disorder and memory services, the findings are presented by specialty. Analysis of variance indicated no significant changes between responses to T1 and T2 questionnaires relating to new assessments. Figure 10 therefore presents data from only T1, which indicated that staff working in movement disorder services more consistently reported exploring parkinsonism and REM sleep behaviour disorder, whereas those in memory services were more likely to report using formal cognitive tests and applying diagnostic criteria.

FIGURE 10.

Percentage of T1 respondents reporting always covering specific areas in new consultations by specialty. **p < 0.001.

These results suggest that, despite the availability of the assessment toolkits, staff working in memory services were not routinely exploring all key features of DLB in new patient assessments. With respect to diagnosing PDD, it would not be anticipated that formal cognitive testing would be conducted at every new consultation, as clinicians often emphasised that there was insufficient time and that it was not always appropriate to look at cognition in the first appointment. In the absence of data on practice prior to the introduction of the assessment toolkits, it is difficult to interpret these results.

Integration of the toolkits into routine practice (T1 and T2 only)

Questions related to NPT explored the perceived value of the toolkits, whether or not participants believed that the toolkits were a legitimate part of their role, ease of integrating the toolkits into their work and whether or not they thought that the toolkits would become part of their normal practice. Parallel questions were asked about the assessment and management toolkits, with responses on a seven-point scale (where higher scores indicate greater agreement with each statement). Responses from T1 have been coded into three groups (agree, neutral and disagree) in Figure 11. Views on the assessment and management toolkits were generally similar, although respondents had slightly more positive views about the management toolkit becoming part of their normal practice. Only one person disagreed with any of the items, and this was on the item relating to whether or not the assessment toolkit was likely to become part of their normal practice.

FIGURE 11.

Views on integration of the management and assessment toolkits into routine practice at T1. AT, assessment; MT, management.

Further analyses explored whether or not views on integration of the assessment toolkits were related to role, specialty and study arm or changed between T1 and T2. There were trends for views on implementation to become less positive over time and for allied health professionals to express less positive views on whether or not the assessment toolkit was a legitimate part of their role (both p < 0.05). The mean values on the seven-point scale for each NPT item by study arm and specialty are shown in Figures 12 and 13.

FIGURE 12.

Mean scores on NPT questions relating to the assessment toolkit by study arm (T1). *p < 0.01, **p < 0.001.

FIGURE 13.

Mean scores on NPT questions relating to assessment toolkit by specialty. *p < 0.01, **p < 0.001.

Respondents in the intervention arm generally held more positive views towards implementation of the assessment toolkits than those in the control arm. Although rarely explicitly mentioned during interviews, the availability of the management toolkit may have created more positive views towards accurate diagnosis of LBD and, therefore, towards the assessment toolkit. Respondents in memory services had more positive views towards implementation of the assessment toolkits than their colleagues in movement disorder services (see Figure 13). These findings are consistent with the views expressed in qualitative interviews when professionals in movement disorder services often expressed concerns over the feasibility of implementation of the assessment toolkits.

Parallel analyses were conducted for the NPT questions about the integration of the management toolkit. There were no significant differences in views expressed according to region, role or specialty (study arm was not relevant, as only those in the intervention arm received the management toolkit).

Overall views on toolkits (T2 only)

On the final questionnaire, we asked if and how the toolkits had an impact on the respondents’ work. These questions were included in only the intervention questionnaire (n = 34). Six respondents had not used the assessment toolkits (17.6%), half reported that it had changed their approach to new or follow-up assessments (50.0%), five were unsure (14.7%) and the remaining six felt that it had not changed their approach (17.6%).

Three respondents did not answer the question relating to the impact of the management toolkit and nine reported that they had not used it. The remaining respondents had most commonly used the management toolkit for reference (19/22, 86.4%), with smaller numbers using it to explore new treatment options (9/22, 40.9%) or with trainees (8/22, 36.4%). (As respondents could use the toolkit in more than one way, the percentages do not sum to 100.)

Comments on open questions (T1 and T2 only)

Overall, just under half (45.5%) of respondents made at least one comment on either the T1 or the T2 questionnaire. Analysis of the comments indicated that they were consistent with the analysis of qualitative interviews and focus groups. As some questionnaire respondents were based in services where it had not been possible to conduct any observation or interviews, this provided useful triangulation of the qualitative analysis from new perspectives.

Summary

Response rates to the questionnaires were acceptable and staff with a variety of roles took part. The large number of questionnaire invitations left unopened when delivered electronically suggests that both electronic and paper copies of questionnaires should be sent to maximise response rates. The findings from the T0 questionnaire confirm that confidence in diagnosing and managing DLB and PDD is lower than for most other subtypes of dementia. This supports the need for assessment and management toolkits. Confidence in managing common LBD symptoms varied according to specialty and underpinned the dual pathway, with referrals to both movement disorder and memory services, experienced by many people with LBD. Consistent, small but (statistically) significant changes in confidence in managing these common symptoms were found between the first questionnaire and those administered during or after the pilot trial for respondents based in services that received the management toolkit, lending support to the hypothesis that the potential mechanism for change is through the accumulation of marginal gains. Caution is, however, needed in interpreting differences over time, as the respondents at each time point are overlapping, but not identical.