Trial Review

Did you know?

The ANZCTR now automatically displays published trial results and simplifies the addition of trial documents such as unpublished protocols and statistical analysis plans.

These enhancements will offer a more comprehensive view of trials, regardless of whether their results are positive, negative, or inconclusive.

The safety and scientific validity of this study is the responsibility of the study sponsor and investigators. Listing a study does not mean it has been endorsed by the ANZCTR. Before participating in a study, talk to your health care provider and refer to this information for consumers
Trial registered on ANZCTR


Registration number
ACTRN12623000350628
Ethics application status
Approved
Date submitted
15/03/2023
Date registered
5/04/2023
Date last updated
3/04/2024
Date data sharing statement initially provided
5/04/2023
Type of registration
Prospectively registered

Titles & IDs
Public title
Music Attuned Technology for Care via eHealth – MATCH (Study 3.2)
Scientific title
Music Attuned Technology for Care via eHealth – MATCH: Determining Digital Markers of Agitation (Study 3 - Stage 2)
Secondary ID [1] 309167 0
Nil known
Universal Trial Number (UTN)
Trial acronym
MATCH - Study 3.2
Linked study record
ACTRN12622000193774 is one of the MATCH Project studies, and it is a proof of concept study trialling the MATCH App with people living with Dementia in the Community. The MATCH technology aims to integrate the Community and Residential Aged Care versions of the App with the biomarkers to recommend music.

Health condition
Health condition(s) or problem(s) studied:
behavioural symptoms of dementia 329295 0
psychological symptoms of dementia 329296 0
Agitation 329297 0
Condition category
Condition code
Neurological 326242 326242 0 0
Dementias
Neurological 326243 326243 0 0
Alzheimer's disease
Neurological 326244 326244 0 0
Other neurological disorders

Intervention/exposure
Study type
Interventional
Description of intervention(s) / exposure
This study will be conducted in residential aged care (RAC) facilities in Victoria, Australia. We will fit consenting residents with wearable sensors and collect sensor data over a 10-day observation period (2 x 5-day observations separated by a 2-week break). Residents will be asked to wear the sensor devices for up to 8 hours per day.

During the first observation, the research team will observe participants to record episodes of agitation. During the second observation, the research team will continue to observe participants to record episodes of agitation. In addition, researchers will play manualised playlists of familiar/preferred and/or neutral music when a participant displays agitation symptoms and record any changes in agitation symptoms in response to the music. Five separate music playlists (5-10 songs per playlist, 1 playlist per day) will be curated by a music therapist prior to commencement of the study. Song choices will be based on information collected from the resident, their family and/or care staff regarding the resident’s familiar/preferred music, as well as any unpreferred music, specific songs to avoid, or past negative responses to music. Playlists will be created using Spotify and will be played on a wireless Bluetooth speaker. The playlists will be played in their entire duration unless participants respond negatively when the music is played (i.e. if agitation symptoms escalate or residents express wanting the music to stop), in which case the MATCH researcher will stop the music. Negative responses to music will also be recorded as adverse events. Data from this study will be used to determine the feasibility of using the different sensors from a clinical and technical perspective, as well as to develop the agitation detection algorithms by comparing clinical notes to sensor readings. Further, data from clinical notes, sensor readings and music metadata will be used to inform the development of an AI system to refine music recommendations to reduce agitation in residents living with dementia. The observation periods will be separated by a 2-week break in which researchers will develop models for agitation detection based on data from the first 5-day observation. Data from the second 5-day observation period will be used to test these models.
Intervention code [1] 325615 0
Behaviour
Intervention code [2] 325616 0
Treatment: Devices
Comparator / control treatment
No comparator will be used.
Control group
Uncontrolled

Outcomes
Primary outcome [1] 334116 0
Sensitivity and feasibility of sensor devices to capture biometric data that enables the prediction of agitation in residents with dementia, as well as changes in agitation levels in response to music, will be captured as a Composite Primary Outcome.

Biomarkers, such as respiratory rate, pulse rate and variability, motion-based activity, peripheral skin temp, electrical properties of skin, and galvanic skin response (perspiration), will be collected by the Empatica EmbracePlus sensor. Audio data (vocal patterns) will be collected by an audio sensor device (microphone) device worn by participating residents.

Clinical observations will be recorded using an adapted version of the Cohen Mansfield Agitation Inventory (CMAI). The CMAI includes 29 symptoms of agitation related to verbal, vocal, or motor activity. When the researcher observes an incident of agitation, the time of symptom onset and end will be recorded, as well as the types of behaviours observed and their severity based on perceived distress to the resident (mild, moderate or severe).
Timepoint [1] 334116 0
Biomarkers (physiology, motion and vocal patterns) will be continuously collected during the period of observation (2 x 5-day observations, <8 hrs per day)

Clinical observations of agitation symptoms will be recorded by an experienced clinician when agitation episodes occur during the observation period (2 x 5-day observations, <8 hours per day)
Secondary outcome [1] 419414 0
Reliability and safety of sensor devices: Any technical events (i.e. issues with the technical aspects of the sensors) will be recorded by researchers in a technical events log during the 10-day observation

Any adverse events that arise during the study will be recorded by researchers in an adverse event log.
Timepoint [1] 419414 0
Technical and adverse events will be recorded during the observation period (2 x 5-day observations)

Eligibility
Key inclusion criteria
• Resident (full-time, 24 hours/day) at a participating RAC
• Documented diagnosis of dementia (any aetiology)
• Dementia Severity Rating Scale score >18
• NPI-NH: Severity score >6
• Has given written informed consent (or verbal assent with written consent by proxy for those legally unable to provide consent themselves)
• Presence of at least 1 agitation symptom occurring at least several times a day (rated on the CMAI)
• Able to tolerate wearing sensors, based on care staff clinical knowledge about the resident
• Does not have severe aggression (NPI-NH score is less than 3 on agitation and aggression scale)
• Does not have a severe medical illness and/or life-threatening condition, as assessed by study doctors
• No known allergy to components in the Empatica (anodized aluminum, polycarbonate, SUS316L, 304 stainless steel, polyurethane; see https://support.empatica.com/hc/en-us/articles/115000195083-Embrace-materials)
• Does not have delirium
• Has no or low risk of pressure injuries (rated on Braden Risk Assessment score of 13 or higher)
Minimum age
No limit
Maximum age
No limit
Sex
Both males and females
Can healthy volunteers participate?
No
Key exclusion criteria
• Residents for whom there is a risk associated with them wearing the sensors
• Residents who refuse to wear the sensors
• Residents with hyperorality, who frequently place items in their mouth
• Residents whose mental state may be impacted by wearing the sensors (as assessed by care staff)
• Residents with known paranoid or delusional ideas
• Residents for whom the RAC staff overseeing their care believes participation would be contraindicated

Study design
Purpose of the study
Treatment
Allocation to intervention
Non-randomised trial
Procedure for enrolling a subject and allocating the treatment (allocation concealment procedures)
Methods used to generate the sequence in which subjects will be randomised (sequence generation)
Masking / blinding
Open (masking not used)
Who is / are masked / blinded?



Intervention assignment
Single group
Other design features
As this is a pilot feasibility study, we have no specific requirements regarding the number of participants. We aim to recruit up to 24 residents living with dementia. This sample size is deemed feasible from an implementation perspective and will provide sufficient data for further development, given monitoring will occur over a 10-day period (per participant).
Phase
Not Applicable
Type of endpoint/s
Efficacy
Statistical methods / analysis
Each sensor unit will be configured to fire a heartbeat signal every 24 hours. The sensor heartbeat is a binary signal that varies between normal and abnormal, indicating the hardware status of the sensor. A software tool will be developed to examine the hardware status for all sensor units and alert research team about potential faulty sensors on a daily basis. Data sanity check will be performed to detect corrupt or inaccurate records from the collected dataset. As the first step of data sanity check, the percentage of missing values for each of the sensor units will be calculated. The sensor units with a missing value percentage over 5% will be dropped from further analysis.
An agitation detection algorithm will be developed based on the data gathered and formulated as a supervised learning task. A learning model will be trained to map the physiological, activity, and audio data to the occurrence of agitation observed by clinicians. This algorithm will build a hypothesis from a set of “labelled” instances (the training set with the corresponding estimated agitation levels, based on observations from the modified CMAI form), which then can be used to make predictions about future unlabelled instances (i.e., predict agitation levels). To reduce variability in the results, a k-fold cross-validation technique will be used to assess the classification. Effect sizes will be calculated for the baseline scores for NPI-NH and Cohen-Mansfield Agitation Inventory.
The agitation levels estimated by the agitation detection algorithm will be used to build our music recommender system. This music recommender system predicts a set of songs to reduce the agitation level of a given user (i.e., intervention). Reinforcement learning techniques will be used to reward the recommendation model when the predicted songs contribute to the intervention process (i.e., our model will learn which songs should and shouldn’t be recommended).

Recruitment
Recruitment status
Recruiting
Date of first participant enrolment
Anticipated
17/04/2023
Actual
17/04/2023
Date of last participant enrolment
Anticipated
30/06/2024
Actual
Date of last data collection
Anticipated
30/09/2024
Actual
Sample size
Target
24
Accrual to date
14
Final
Recruitment in Australia
Recruitment state(s)
VIC

Funding & Sponsors
Funding source category [1] 313366 0
Government body
Name [1] 313366 0
Medical Research Future Fund (National Health and Medical Research Council)
Country [1] 313366 0
Australia
Primary sponsor type
University
Name
The University of Melbourne
Address
Faculty of VCA and MCM
University of Melbourne
234 St Kilda Road VIC 3006
Country
Australia
Secondary sponsor category [1] 315126 0
None
Name [1] 315126 0
Address [1] 315126 0
Country [1] 315126 0

Ethics approval
Ethics application status
Approved
Ethics committee name [1] 312585 0
The University of Melbourne - HASS 1
Ethics committee address [1] 312585 0
Office of Research Ethics and Integrity | Research, Innovation & Commercialisation
Level 5, Alan Gilbert Building, 161 Barry Street, Carlton
The University of Melbourne, Victoria 3010, Australia
Ethics committee country [1] 312585 0
Australia
Date submitted for ethics approval [1] 312585 0
22/07/2022
Approval date [1] 312585 0
03/10/2022
Ethics approval number [1] 312585 0
2023-24870-36694-8

Summary
Brief summary
An eHealth solution – Music Attuned Technology for Care via eHealth (MATCH) – was developed to support family carers of people living with dementia to use music intentionally to support care. To create scalable solutions for the growing number of people living with dementia, we developed a minimal viable product (the MATCH app) for the HOME setting, which will be adapted for the RAC setting in study 3.
MATCH represents a paradigm shift in music and dementia technology because it will: a) embed training programs that guide FCs and professional care staff in the strategic use of music; b) use sensor technology to capture behavioural markers to interpret agitation levels and auto-suggest music using algorithms that learn preferences of the person living with dementia and then suggest music they may like (recommender system); c) be able to continuously adapt the music to match and attune to arousal levels and reduce agitation; and d) be accessible to culturally and linguistically diverse groups (training videos will be available in multiple languages).
Stage 2 of this study aims to explore digital markers, using wearable sensor devices, as best indicators of agitation and other symptoms to inform the development of an AI system that can detect changes in agitation in response to music and refine music recommendations to reduce agitation symptoms in residents living with dementia.
We expect that data from the sensor devices will be able to support the detection and better management of agitation symptoms in people living with dementia.
Trial website
www.musicattunedcare.com
Trial related presentations / publications
Public notes

Contacts
Principal investigator
Name 125186 0
Prof Felicity Baker
Address 125186 0
Faculty of Fine Arts and Music
Conservatorium of Music – Parkville Campus
The University of Melbourne Victoria 3010
Country 125186 0
Australia
Phone 125186 0
+61 402 172 795
Fax 125186 0
Email 125186 0
[email protected]
Contact person for public queries
Name 125187 0
Tanara Vieira Sousa
Address 125187 0
Faculty of Fine Arts and Music
Conservatorium of Music – Parkville Campus
The University of Melbourne Victoria 3010
Country 125187 0
Australia
Phone 125187 0
+61 403159473
Fax 125187 0
Email 125187 0
[email protected]
Contact person for scientific queries
Name 125188 0
Felicity Baker
Address 125188 0
Faculty of Fine Arts and Music
Conservatorium of Music – Parkville Campus
The University of Melbourne Victoria 3010
Country 125188 0
Australia
Phone 125188 0
+61 402 172 795
Fax 125188 0
Email 125188 0
[email protected]

Data sharing statement
Will individual participant data (IPD) for this trial be available (including data dictionaries)?
Yes
What data in particular will be shared?
Non-identified data will be made available after the end of the MATCH Project, including manuscripts published in a Public Data Repository.
When will data be available (start and end dates)?
01/01/2027 - to unlimited timeline.
Available to whom?
To the general public
Available for what types of analyses?
No restrictions
How or where can data be obtained?
Assessing the Public public Data Repository or via email to the PI Professor Felicity Baker: [email protected]


What supporting documents are/will be available?

Doc. No.TypeCitationLinkEmailOther DetailsAttachment
18550Ethical approvalNA [email protected] 385532-(Uploaded-09-03-2023-15-05-43)-Study-related document.pdf



Results publications and other study-related documents

Documents added manually
No documents have been uploaded by study researchers.

Documents added automatically
No additional documents have been identified.