Study design
This feasibility study is designed as a two-arm, parallel group randomised controlled trial of community singing for people with dementia and their carers, with a waiting-list control group. The study will aim to recruit 80 dyads each consisting of a person with dementia and their carer. About half of the sample, selected at random, will be invited to attend a singing group, whilst the other half will be asked to wait for three months before being invited to attend a group. Outcome data will be collected at baseline (time 0), three months after baseline (time 1), and six months after baseline (time 2).
Setting
The singing groups created for the study will take place in community settings in Nottinghamshire, UK, and data collection will take place during visits to participants at home. The venues selected will be chosen for their suitability for holding singing group sessions; this means well-lit, attractive spaces with good acoustics that are large enough to comfortably accommodate up to 50 people. Further considerations include the provision of accessible facilities and a kitchen for refreshments, adequate car parking and a reliable and frequent public transport service nearby. We will focus on two geographical areas and will aim to recruit half the intended target of 80 dyads in each of these areas. Recruitment will be optimised by selecting venues in geographical areas which are believed to have the highest levels of need in terms of provision of recreational singing for people with dementia.
Inclusion and exclusion criteria
To be included in the study, participants with dementia must
- be aged 18 or over, and diagnosed with dementia, which is at a mild or moderate level
- have a carer who spends at least two hours per week with them,
- be able to give informed consent at the start of the study,
- be able to speak and understand English[1], and
- be willing in principle to join a singing group and attend weekly.
Participants with dementia will be excluded if
- they lack capacity to give informed consent,
- they are already attending a singing group or have done so in the past 6 weeks (other than religious services),
- they have a significant hearing impairment,
- they are already participating in any other interventional study, or
- they have a history of severe mental illness or drug/alcohol addiction.
Carers can be included if they are
- able to speak and understand English
- willing in principle to attend the group weekly, and
- able to give informed consent.
Relationships between people with dementia and their carers may be that of spouses/partners, parent/child or other family members, or friends. In certain cases, professional caregivers may be part of the study if no family member or friend is available to accompany a person with dementia to the sessions and if the same professional caregiver is available each week. The nature of the relationship between the dyads does not matter, but it is important that the same carer attends the sessions throughout the study.
Intervention
The intervention in this study is attendance at weekly singing group sessions for 10 weeks. Each session will include approximately 60 minutes of warm-up activities and singing, with about 30 minutes for refreshments and socialising before or after the session. The sessions will be run jointly by two professional community musicians who have extensive experience of running similar singing groups together. The format of the sessions includes a full physical and vocal warm-up, followed by singing familiar songs accompanied by guitar. Themed songbooks will be provided for participants; each week a different songbook will be selected by the facilitators, so each session will have a different theme. Themes might include songs about the weather, from a particular era, or relating to the current season (e.g. Christmas, summer). Singing group participants will be encouraged to choose their favourite songs from the books for the whole group to sing. The facilitators’ approach will emphasise that the aim of the group is to sing for fun and enjoyment, and that the participants’ musical skills or singing voices are not being assessed or judged.
The content of each session will be fully documented, and three out of every ten sessions will be video recorded in order to check that the intervention is delivered consistently. Attendance at singing sessions will be monitored and recorded. Participants will be free to miss individual sessions or withdraw from the singing group entirely as they wish, and they do not need to give a reason.
Since each singing group can accommodate up to 20 dyads, we propose to run four groups in total in two different geographical areas. The first two groups will be set up for those allocated to the intervention group; after the three month waiting period, we will start another two groups for those in the waiting-list control group. There is provision for those in the intervention group to continue to attend the singing groups after the initial 10-week period, if they wish to. This is for ethical reasons and to facilitate embedding the groups into the local community, as discussed later in the “Ethics” section. Continued attendance by participants in the intervention group will not compromise the primary objectives of recruitment, retention and acceptability, and its effects on the six-month data collection are considered an acceptable trade-off.
Control condition
Participants who are allocated to the waiting-list control group will receive care as usual during the three months that they will be required to wait before beginning to attend a singing group. They will be requested to abstain from attending other organised group singing activities during this time, with the exception of religious services.
Endpoints
Since PRESIDE is a feasibility study, the endpoint is the attainment of the feasibility criteria at 6 months. The proposed criteria for proceeding to a full trial are 70% recruitment; 70% retention of participants over 3 months; and acceptability of the control condition demonstrated by 70% of the waiting-list participants starting singing groups after being on the waiting list. The acceptability of the control condition is considered to be the most important of these three criteria since this ensures the heterogeneity of the sample. The other feasibility criteria are considered negotiable.
An additional endpoint for PRESIDE is the measurement of quality of life, activities of daily living, mood, musical engagement and cognition of people with dementia, and the measurement of quality of life, mood, caring relationship and experience of caring for the carers of the people with dementia. Each dyad’s use of health social care services and resources will also be documented. Additionally, qualitative data will be collected using interviews and participant observation, in order to better understand the experiences of dyads who attend the singing groups.
Participant timeline in study
Figure 1 shows the timeline of participants in the study, while Figure 2 gives the schedule of assessment and enrolment for the study. Since we need to recruit a sufficient number of participating dyads before the singing groups can begin, there will be a delay between recruitment and beginning the singing group or waiting list control for participants who are recruited early. We will not collect baseline data or randomise participants until the singing groups are just about to start.
Sample size
As PRESIDE is a feasibility study with two arms, it is suggested 40 dyads per arm will provide enough information to inform future, definitive RCT design and implementation [19]. Therefore the recruitment target is 80 patient-carer dyads in total in the current study.
Allocation and blinding
Participants will be recruited in pairs consisting of a person with dementia and a carer. They will be randomised in pairs to either the intervention group or the waiting-list control group. Randomisation will be stratified by sex and Clinical Dementia Rating (CDR). Randomisation will be carried out using a web application created by the trial programmer at the University of Nottingham Stroke Trials Unit, using web and database servers provided by University of Nottingham Information Services (IS). The researcher (BD) will use this web application to randomise each participant. The system will hide information relating to the intervention groups from other users until explicitly granted access by the trial programmer, for example when a statistician requires unblinded data.
Dyads will be randomly allocated to intervention or waiting-list control groups by the web-based system. BD will collect baseline data prior to randomisation, and so will be unaware of allocation until the initial data collection is complete. Thereafter, both the researcher (BD) and the participants will be aware of the group allocation, along with those delivering the sessions. Due to the nature of the intervention and control groups, it is not feasible to blind participants to group allocation; they will know whether they are attending singing group sessions. The research team will be split into two groups, which will define their access to data and to group allocation. The data management group will be responsible for the data collection and running of the study, and will not be blinded to group allocation once baseline data collection has taken place. The data analysis group will be responsible for conducting the analysis, and will remain blinded to group allocation for the duration of the study. The data analysis group will not have any contact with study participants.
Follow-up data collection will be undertaken by researchers who are blinded to group allocation. However, since the participants will know their own group allocation there is a possibility that they will reveal this during the follow-up data collection. Unblinding of outcome assessors which occurs in this way will be recorded. All researchers undertaking data collection will have been fully trained in the administration of these instruments.
Recruitment
We will recruit for the study through various channels in order to maximise the chances of reaching as many people with dementia as possible. We will recruit through NHS Memory Assessment Services by publicising the study to people who have recently been diagnosed with dementia. We also have agreement from the Alzheimer’s Society to advertise the study to people who are currently on the waiting list in the area for Singing for the Brain sessions. We will harness the support of local community organisations, charities and the local press, and will also use targeted social media advertising. Recruitment will also be undertaken using the online platform “Join Dementia Research”.
Data collection
Quantitative data
Outcome measures will be administered to participants during home visits arranged at times convenient to them. Each visit is expected to take up to two hours. Baseline data collection will be conducted by BD, and follow-up data collection by blinded researchers. All researchers undertaking data collection will have been fully trained in the administration of these instruments.
The following outcome measures have been selected for use in the study. These are summarised in Table 1, which also shows whether each measure corresponds to the person with dementia, the carer, or both, as well as who completes the measure (the person with dementia, or the carer, either self- or proxy-rating). All outcome measures will be administered at each of the three time points. Additionally, demographic data will be collected from people with dementia (age, sex, diagnosis, date of diagnosis, ethnicity, education, musical/singing experience) and from carers (age, sex, ethnicity, education, musical/singing experience) at baseline. Measures will be completed by the person with dementia themselves and/or via a proxy rating by the carer. The carer will complete outcome measures about their own mood, quality of life and experience of caring. Some of these measures will be administered by interview and some will be self-completed by the carer.
A large number of outcome measures are included in this study, in order to assess which of them might be most suitable for future use in a full trial. We recognise that data collection may be time-consuming and tiring for participants, and consequently missing data and incomplete measures are expected in some cases. The order and format of the administration of outcome measures is given in Table 2. This order has been chosen so that instruments measuring similar domains are separated, to maximise the range of information included in incomplete data sets. The interview with the carer will take place first, and subsequently the person with dementia will be interviewed while the carer does the self-completed outcome measures.
Quality of Life – Alzheimer’s Disease (QoL-AD) Scale
This 13-item measure asks respondents to rate various aspects of the person with dementia’s life as poor, fair, good or excellent [20]. It will be completed by participants with dementia, and a proxy rating will also be provided by the carer. Each item is scored on a scale from 1 to 4, so the measure yields a total score out of 52, where a higher score indicates a higher quality of life.
EQ-5D-5L
This instrument comprises five questions about health-related quality of life, and a visual analogue scale [21–23]. It will be completed by participants with dementia and carers. The results will be used to calculate Quality-Adjusted Life Years (QALYs) for all participants. The first five questions are scored on a five-point scale and generate a five-digit health state for each participant (e.g. a score of 12345 would indicate no problems with mobility, slight problems with self-care, moderate problems doing usual activities, severe pain or discomfort, and extreme anxiety or depression). The visual analogue scale is the final item of the EQ-5D-5L and asks respondents to rate their current health as a point on a line. This gives a score out of 100, where a higher score indicates a higher perceived level of overall health.
Dementia Quality of Life (DEMQOL)
This scale asks the person with dementia to rate their quality of life by responding to 29 questions about their feelings, memory and everyday life [24]. The respondent selects “a lot”, “quite a bit”, “a little” or “not at alL” in response to each question. Each question is scored on a scale from 1 to 4, giving a total score out of 116 where a higher score indicates a higher quality of life.
Bristol Activities of Daily Living Scale (BADLS)
The 20-item scale presents the respondent with various descriptions of performance at activities such as cooking, hobbies and personal care, and asks them to rate the person with dementia’s current level of performance [25]. It will be completed as a proxy rating by the carer. Each items is scored on a scale of 0 to 3, and gives a total score out of 60, where a higher score means a higher level of impairment.
Cornell Scale for Depression in Dementia (CSDD)
The 19-item scale assesses symptoms of depression in people with dementia and their severity [26]. It is completed through interviews conducted with the person with dementia and the carer separately, with the interviewer assigning a score based on the severity of symptoms within the past week. For each indicator of depression, absent symptoms score 0, mild or intermittent symptoms score 1, and severe symptoms score 2. The total score is out of 38, where scores above 10 indicate probable major depression, and scores above 18 indicate definite major depression.
Geriatric Depression Scale (GDS-15)
The 15-item version of this scale presents the respondent with “yes/no” questions about how they have been feeling over the past two weeks [27]. It will be completed by people with dementia and carers. Each answer indicative of depression scores one point. A score of 5 or higher is suggestive of a depressive disorder, while a score of 10 or higher is indicative of a probable depressive disorder.
Musical Engagement Questionnaire (MuseEQ)
This scale has 35 items which ask the participant how they have used and engaged with music over the past month [28]. It will be completed by people with dementia. The scale consists of statements such as “I play a song over and over if I like it” and the respondent indicates the extent to which they agree with the statement on a five-point scale, scored from 1 to 5. The total score is out of 175, and a higher score indicates a higher level of musical engagement.
ADAS-Cog
The measure assesses a number of facets of cognitive performance including memory, verbal skills, motor skills, and orientation [29, 30]. It will be completed by the person with dementia. The tests include memorising a list of words, naming familiar objects as they are presented, copying drawn shapes, and following verbal instructions. The interviewer also rates the participant on their verbal fluency, comprehension and ability to concentrate. The total possible score is 70, with higher scores indicating higher levels of cognitive impairment.
Clinical Dementia Rating (CDR)
This instruments gathers data through semi-structured interviews conducted with the carer and the person with dementia [31]. The interviewer uses this data to rate the severity and stage of dementia across six different domains: memory, orientation, judgment/problem solving, community affairs, and home/hobbies. The rating of each of these domains is used to generate an overall CDR score which indicates the level of impairment present: 0 (none), 0.5 (questionable), 1 (mild), 2 (moderate) or 3 (severe).
Quality of the Carer-Patient Relationship (QCPR) Scale
The 14-item measures invites the respondent to state their level of agreement with questions about their relationship with the person for whom they care [32]. It will be completed by carers. The responses are scored from 1 to 5, and the total score is out of 70, with a higher score indicating higher perceived relationship quality.
Carers of Older People in Europe (COPE) Scale
The 15-item scale asks the carer about their experiences of caregiving and its consequences for their physical and mental health [33]. The responses are scored from 1 (never) to 4 (always). The 15 questions are presented together, but when total scores are calculated they are divided into three subscales which are treated separately in analysis: negative impact of caregiving, positive value of caregiving and quality of support. The precedent for the use of these subscales in analysis is set in one of the paper describing the original development of COPE [34].
Short Sense of Competence Questionnaire (SSCQ)
The SSCQ is derived from the Zarit Burden Inventory and evaluates the carer’s experience of caring across seven items [35]. The carer indicates how strongly they agree with the statements in the scale and each item is scored from 1 to 5. A higher score indicates a higher perceived sense of competence.
Client Service Receipt Inventory (CSRI)
The CSRI is a tool to measure consumption of services with a view to estimating their costs for economic analysis [36]. It will be administered in a form adapted to the typical activities of people with dementia and their carers, in order to measure the use of health and social carer resources by each participant. This instrument will be completed by the carer.
Table 1: Summary of outcome measures
|
Instrument name
|
Domain measured
|
Participant
|
Respondent
|
|
Person with dementia
|
Carer
|
Person with dementia
|
Carer (proxy)
|
Carer (self)
|
|
QoL-AD
|
Quality of life
|
x
|
|
x
|
x
|
|
|
EQ-5D-5L
|
Health-related quality of life
|
x
|
x
|
x
|
|
x
|
|
DEMQOL
|
Quality of life
|
x
|
|
x
|
|
|
|
BADLS
|
Activities of daily living
|
x
|
|
|
x
|
|
|
CSDD
|
Depression
|
x
|
|
x
|
x
|
|
|
GDS
|
Depression
|
x
|
x
|
x
|
|
x
|
|
MuseEQ
|
Musical engagement
|
x
|
|
x
|
|
|
|
ADAS-Cog
|
Cognition
|
x
|
|
x
|
|
|
|
CDR
|
Dementia severity
|
x
|
|
x
|
x
|
|
|
QCPR
|
Caring relationship
|
|
x
|
|
|
x
|
|
COPE
|
Experiences/consequences of caring
|
|
x
|
|
|
x
|
|
SSCQ
|
Experiences of caring
|
|
x
|
|
|
x
|
|
CSRI
|
Use of health/social care
|
x
|
x
|
|
x
|
x
|
Table 2: Order and format of outcome measure administration
Person with dementia (by interview)
|
Carer
|
By interview
|
Self-complete
|
1. QoL-AD
2. GDS
3. CDR
4. MuseEQ
5. ADAS-Cog
6. EQ-5D-5L
7. CSDD
8. DEMQOL
|
1. QoL-AD proxy
2. GDS
3. CDR
4. QCPR
5. CSRI
|
1. COPE
2. SSQC
3. EQ-5D-5L
4. BADLS
|
Qualitative data
Previous research into group singing for people with dementia has shown that quantitative measures may not capture the benefits perceived by participants [14, 37]. Qualitative data concerning participants’ experience of the singing groups will be collected during the study using participant observation. BD will conduct structured in-person observations of 3-4 opportunistically selected dyads in each singing group, including both intervention and control group participants. Each dyad will be the focus of observations three times: in one early session, in one session towards the middle of the group’s course, and in one later session. Detailed notes on the observations will be made after the sessions.
We will also collect qualitative data from a sub-sample of participants through semi-structured interviews. We will aim to interview 10 dyads in total, 12.5% of the 80-dyad intended sample. Dyads will be selected for interview opportunistically and those who do not wish to be interviewed will be free to decline. We will interview at least one dyad from each singing group, and will aim to spread the sample evenly between the groups. Interviews will last 30-45 minutes, and each dyad will be interviewed together where possible. The interview guide will focus on understanding participants’ experience of taking part in the group. Interviews will be audio recorded if possible; if participants decline to be recorded, we will seek their permission to make written notes during the interviews.
Data management
Each participant will be assigned a trial identity code number, for use on Case Report Forms (CRFs) and in the database. The trial documents and database will also use participants’ initials, sex and date of birth. Data will be recorded for each participant using paper CRFs. The data from these forms will be entered into the trial database and stored electronically. Access to the database will be password protected and restricted to trial personnel, who will each have a unique login code and PIN. The paper CRFs will be stored in a locked cupboard on University of Nottingham premises. Consent forms (which contain participants’ real names) will be stored separately in a different locked cupboard. Participant’s names and contact information will be securely stored on a separate part of the database system, not associated with their trial data.
Any video or audio recordings taken during data collection will be transferred daily from the recording device to password-protected encrypted drives on University of Nottingham computers. They will then be deleted from the recording device.
Data will be stored securely for seven years after the last person has been recruited. All data will be analysed on University of Nottingham computers and backed up to the University of Nottingham servers. Only members of the research team authorised by the Data Custodian will have access to the research data.
Data coding and analysis
Statistical analysis of quantitative data
Quantitative data analysis will be conducted by the study statistician (BG), using the software package Stata 16. The data analysis will be mainly descriptive [38]. All measures will be summarised with mean and standard deviation (S.D.) for normally distributed variables, median and inter-quartile range (IQR) for skewed variables, and frequency (%) for response of categorical variables, by intervention arm across follow-up time. Analysis of Co-variance (ANCOVA) will be performed to quantify treatment effect size and its precision by means of regression modelling with baseline measures as covariate, and to derive the change from baseline score for each arm. Due to the non-independence of patient-carer dyad measures and with data being measured repeatedly, multivariate, multilevel modelling will be performed to account for the association between same kind of outcome measure of patient-carer data and repeated data. We will calculate recruitment rate, retention rate and attrition rate, exploring patterns of missing data and drop out from the intervention group and the waiting list control group. All results will be used to inform future definitive trial design.
Analysis of qualitative data
Recordings/notes from interviews and observational notes will be transcribed as necessary and analysed using NVivo 12. The technique for analysis will be the “general inductive approach” in which the coding is largely guided by evaluation objectives [39]. The evaluation objectives in this case will be the feasibility of the study design and the experiences of all participants who attend a singing group.
BD will undertake the first stage of the analysis and will assign first-order codes to the data. Through discussion with the rest of the research team, second order codes will then be developed. In order to enhance the validity of the coding process, a sample of transcripts and notes will be coded by another member of the research team, and any disagreements between coders will be discussed.
Data monitoring
A trial Data Management and Ethics Committee (DMEC) will be formed, consisting of members who are not part of the University of Nottingham. All the members of the DMEC will have experience of running clinical trials, and the group will include a statistician, an academic and a clinician. The Chief Investigators (JS & MO) will be the data custodians.
Adverse events
The occurrence of an adverse event as a result of participation in this study is not expected due to the nature of the intervention, which is a non-clinical intervention based on social music-making and has no known side effects. However, in order to provide further assurance of the safety of the intervention, data on Severe Adverse Events (SAEs) will be collected for people with dementia and their carers during follow-up data collection at 3 and 6 months. Additionally, attendance at the group will be monitored by the singing group facilitators and the research associate, so any SAE which occurs among the group members will be recorded. These include hospitalisation, extension of hospitalisation, other serious medical events and death. Such SAEs might conceivably be related to participation in the intervention if they were associated with falls, emotional arousal or depression.
[1] Participants with aphasia or other speech problems will be included unless these were so severe that the participant could not answer the questionnaires, or communicate their answers in an alternative way.