Trial Review

Registering a new trial?

To achieve prospective registration, we recommend submitting your trial for registration at the same time as ethics submission.

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
ACTRN12622000486729
Ethics application status
Approved
Date submitted
16/03/2022
Date registered
28/03/2022
Date last updated
28/03/2022
Date data sharing statement initially provided
28/03/2022
Type of registration
Retrospectively registered

Titles & IDs
Public title
Using speech characteristics to classify suicide risk in telehealth callers
Scientific title
Using speech characteristics to classify low versus high risk of suicide in 64 male and female telehealth counselling calls
Secondary ID [1] 306700 0
Nil known
Universal Trial Number (UTN)
U1111-1275-8186
Trial acronym
RELIANCE (REaL tIme ANalysis of voiCE)
Linked study record

Health condition
Health condition(s) or problem(s) studied:
suicide 325664 0
Condition category
Condition code
Mental Health 323013 323013 0 0
Suicide

Intervention/exposure
Study type
Observational
Patient registry
False
Target follow-up duration
Target follow-up type
Description of intervention(s) / exposure
Duration of observation is 1/7/2020 - 30/6/21
Call recordings made to a national telehealth suicide prevention service will be observed throughout the duration of each call. Callers identified as being at high risk of imminent suicide will be identified. Callers will by identified via screening questions ("are you at risk of suicide", and "do you have a plan that you will carry out"). High risk of suicide status will be initially assigned by the responding counsellor using an organisational risk framework, then confirmed retrospectively by a team of Associate Researchers and finally re-confirmed by the student investigator, using the Columbia Suicide Severity Risk rating (ratings 5 and 6). Calls are made form private homes, and in public spaces.
Intervention code [1] 323146 0
Early Detection / Screening
Comparator / control treatment
Call recordings made to a national telehealth suicide prevention service will be observed throughout the duration of each call. Callers identified as being at low risk of imminent suicide will be identified. Callers will by identified via screening questions ("are you at risk of suicide", and "do you have a plan that you will carry out"). Low risk of suicide status will be initially assigned by the responding counsellor using an organisational risk framework, then confirmed retrospectively by a team of Associate Researchers and finally re-confirmed by the student investigator, using the Columbia Suicide Severity Risk rating (ratings 1 and 2). Calls are made form private homes, and in public spaces.
Control group
Active

Outcomes
Primary outcome [1] 330780 0
High risk of suicide shall be determined via assessment using the Columbia Suicide Severity Rating Scale (CSSR-S) (Rating 5 and 6) (highest levels)
Timepoint [1] 330780 0
Assessment of high risk of suicide will be made at the conclusion of each call, and after listening to each call recording in full.
Secondary outcome [1] 407538 0
Emotional distress shall be assessed via the Distress Barometer (10 point Likert-style scale, 0=absence of emotional distress; 10 = highest level emotional distress).
Timepoint [1] 407538 0
Emotional distress shall be assessed at various timepoints across each call recording. Timepoints shall correspond to a variety of emotional distress ratings within each recording.

Eligibility
Key inclusion criteria
English speaking; call made within Australia; 18+ years
Minimum age
18 Years
Maximum age
No limit
Sex
Both males and females
Can healthy volunteers participate?
Yes
Key exclusion criteria
Calls made via teletext or requiring third party translation; speech difficulties (motor articulation); intoxication.

Study design
Purpose
Screening
Duration
Cross-sectional
Selection
Defined population
Timing
Retrospective
Statistical methods / analysis
Preprocessing of call recordings will include conversion to alternative format (8kHz; 16-bit PCM) compatible with RStudio. Recordings will be normalised and pre-emphasis applied to attenuate low frequencies and magnify high frequencies.
Analysis will involve:
1. L1-penalised Lasso Regression to reduce the number of extracted speech characteristics.
2. 3-level Mixed Effects Generalised Additive Modelling (GAMM) with speech characteristics extracted at within 40ms (50% overlapping) Blackman-windowed frames (level 1); Annotated Segments with Power Spectral Density characteristics (level 2); and individual callers with identified sex of caller (level 3). Binary regression with logit link will be used to discriminate between low and high risk of suicide response. AR1 autocorrelation will also be applied to level 1 predictors.
3. Component-wise Gradient Boosting well be used to automatically classify each recording to either low or high risk of suicide categories. The algorithm will be trained on 70% of the recordings, while the remaining 30% will be used for validation purposes.
The methods used at 2. and 3. will allow for significant non-linearity amongst the predictors by using splines to more accurately model the data.

Recruitment
Recruitment status
Completed
Date of first participant enrolment
Anticipated
Actual
8/07/2021
Date of last participant enrolment
Anticipated
Actual
15/07/2021
Date of last data collection
Anticipated
Actual
15/07/2021
Sample size
Target
500
Accrual to date
Final
840
Recruitment in Australia
Recruitment state(s)
ACT,NSW,NT,QLD,SA,TAS,WA,VIC

Funding & Sponsors
Funding source category [1] 311030 0
Charities/Societies/Foundations
Name [1] 311030 0
On The Line
Country [1] 311030 0
Australia
Funding source category [2] 311033 0
Government body
Name [2] 311033 0
Australian Federal Police
Country [2] 311033 0
Australia
Primary sponsor type
University
Name
Swinburne University of Technology
Address
John Street
Hawthorn
Vic. 3122
Country
Australia
Secondary sponsor category [1] 312353 0
None
Name [1] 312353 0
Address [1] 312353 0
Country [1] 312353 0

Ethics approval
Ethics application status
Approved
Ethics committee name [1] 310578 0
Swinburne University Human Research Ethics Committee
Ethics committee address [1] 310578 0
John Street,
Hawthorn
Vic. 3122
Ethics committee country [1] 310578 0
Australia
Date submitted for ethics approval [1] 310578 0
01/12/2020
Approval date [1] 310578 0
22/01/2021
Ethics approval number [1] 310578 0
20214340-5805

Summary
Brief summary
The primary hypothesis of the study is to automatically classify a sample of call recordings obtained from our telehealth partners according to risk of suicide. Risk of suicide is defined as low (levels 1 and 2 on the Columbia Suicide Severity Rating scale, CSSR-S) and High (levels 5 and 6 on the CSSR-R). Call recordings will be sourced from two industry partners including On The Line who administer the Suicide Call Back service and the Australian Federal Police who administer 000 emergency services in the Canberra, Australia region.
Trial website
Trial related presentations / publications
Public notes

Contacts
Principal investigator
Name 118146 0
Prof Denny Meyer
Address 118146 0
Centre for Mental Health
Swinburne University of Technology
John Street
Hawthorn
Vic. 3122
Country 118146 0
Australia
Phone 118146 0
+61 03 9214 4824
Fax 118146 0
Email 118146 0
[email protected]
Contact person for public queries
Name 118147 0
Ravi Iyer
Address 118147 0
Centre for Mental Health
Swinburne University of Technology
John Street
Hawthorn
Vic. 3122
Country 118147 0
Australia
Phone 118147 0
+61 456 565 575
Fax 118147 0
Email 118147 0
[email protected]
Contact person for scientific queries
Name 118148 0
Ravi Iyer
Address 118148 0
Centre for Mental Health
Swinburne University of Technology
John Street
Hawthorn
Vic. 3122
Country 118148 0
Australia
Phone 118148 0
+61 456 565 575
Fax 118148 0
Email 118148 0
[email protected]

Data sharing statement
Will individual participant data (IPD) for this trial be available (including data dictionaries)?
No
No/undecided IPD sharing reason/comment
Call recordings contain personally identifiable information and relate to sensitive information being discussed in a counselling context.


What supporting documents are/will be available?

No Supporting Document Provided



Results publications and other study-related documents

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

Documents added automatically
SourceTitleYear of PublicationDOI
EmbaseUsing Voice Biomarkers to Classify Suicide Risk in Adult Telehealth Callers: Retrospective Observational Study.2022https://dx.doi.org/10.2196/39807
N.B. These documents automatically identified may not have been verified by the study sponsor.