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The ANZCTR now automatically displays published trial results and simplifies the addition of trial documents such as unpublished protocols and statistical analysis plans.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
ACTRN12619000839101
Ethics application status
Approved
Date submitted
27/05/2019
Date registered
11/06/2019
Date last updated
18/06/2019
Date data sharing statement initially provided
11/06/2019
Type of registration
Prospectively registered

Titles & IDs
Public title
MINDtick app psychiatric care study-Digital signal indicators to enhance capacity of psychiatric services to predict distress in advance and intervene early.
Scientific title
MINDtick app psychiatric care study- Digital signal indicators to enhance capacity of psychiatric services to predict distress in advance and intervene early.
Secondary ID [1] 298162 0
nil known
Universal Trial Number (UTN)
Trial acronym
MINDtick
Linked study record

Health condition
Health condition(s) or problem(s) studied:
Mental illness 312715 0
Condition category
Condition code
Mental Health 311211 311211 0 0
Addiction
Mental Health 311212 311212 0 0
Anxiety
Mental Health 311213 311213 0 0
Depression
Mental Health 311214 311214 0 0
Other mental health disorders
Mental Health 311215 311215 0 0
Eating disorders
Mental Health 311216 311216 0 0
Psychosis and personality disorders
Public Health 311217 311217 0 0
Health service research

Intervention/exposure
Study type
Interventional
Description of intervention(s) / exposure
MINDtick is a smartphone-based application, that is available for download on both iPhone and Android devices. The app has been designed with the purpose of enabling a mechanism to gather continuous measures of symptoms related to mental illness in an individual’s natural daily environment. The app collects this type of information in two ways; 1) participant facing Ecological Momentary Assessments (EMA) measures; and 2) in-built sensor data.

Individuals who are currently receiving any form of psychological or psychiatric treatment for a mental illness will be recruited for the study.

Clinicians will be recruited primarily for the use of assisting in the recruitment of eligible participants, and first contact will be made by the two consultant psychiatrists within the research team. All clinicians from each of the two mental health services will be eligible for recruitment, and interested clinicians will be directed to the study website containing the Clinician Information Sheet, Consent form and Registration form. If a clinician registers to participate, a member of the research team will need to approve a clinician's registration in order to gain access any features of the website.


Eligible individuals will be approached by their treating healthcare professional during their scheduled visit, where the clinician will ask the patient to read carefully over the information sheet (PICF). Consent to participate is given twice: 1) online, through the website, and 2) additionally when they download the MINDtick application.
Additionally, patients will be asked to provide consent for their electronic health records to be accessed as a part of this project.

Consenting participants will be asked to use the MINDtick mobile application at least once a week for 6 months. Upon first entry into the app, participants will be prompted to complete once-off demographic questions (e.g., date of birth, gender, country of origin). MINDtick will prompt participant through a push-notification to answer all 8 EMA questions. The prompt reads: “[name of participant], do you have a minute to answer your health questions?”. Prompts to answer EMA questions occur on average 3 times a week, on a predefined random time schedule consisting of 09:00, 11:00, 13:00, 15:00, and 17:00, whereby only one prompt is delivered per day.
An online dashboard to graphically display gathered EMA data will be made accessible to consenting clinicians. The dashboard will be checked at least once a week during the study for the duration of the project to monitor participant engagement levels with MINDtick and trends in data for clinical use. The data displayed on the clinician dashboard is specific to the EMA data - that is, clinicians will not have access to passive sensor data collected from the patient’s mobile phone.
Intervention code [1] 314578 0
Prevention
Intervention code [2] 314579 0
Treatment: Other
Intervention code [3] 314580 0
Behaviour
Comparator / control treatment
care as usual as decided by health care professional and patient- not based upon specific guidelines.
Control group
Active

Outcomes
Primary outcome [1] 320197 0
Determine the feasibility of the smartphone-based application for continuous symptom and outcome monitoring in a sample of mental health patients.


The feasibility of this method of monitoring the mental health of recently discharged patients will be determined by an action-learning model that will also be used to monitor and improve uptake of the intervention. In order to achieve this, weekly patient enrolment rates (total and percentage of approached patients) at project sites during the study period will be monitored. Patients who declined to participate will be asked their reasons for doing so, with this data then being used to iteratively develop and implement corrective strategies relevant to subgroups that will enhance uptake.


Timepoint [1] 320197 0
weekly data collection, end of intervention (6 months) analysis.
Primary outcome [2] 320198 0
Determine the efficacy of the smartphone-based application for continuous symptom and outcome monitoring in a sample of mental health patients.

For efficacy, there will be a quantitative analysis of the extent to which changes in digitally recorded mobile phone data predict routinely collected assessment measurement scores at baseline, 3 months, and 6 months. This will involve correlational analyses of phone data with HONOS and K10 at 3 and 6 months.
Timepoint [2] 320198 0
baseline, 3 months, 6 months
Secondary outcome [1] 370829 0

Assess the efficacy of an online dashboard system for healthcare professionals to allow for the monitoring of symptoms in patients.

The efficacy of a dashboard for clinicians will be determined in an qualitative analysis of a subset of patient case records by expert clinicians who will rate whether information displayed by the dashboard informed or added value to the case note record. Case notes will be analysed by the lead qualitative researcher using NVIVO software. Using a thematic coding approach to identify themes and coding that arises from the data. These themes and codes will be structured before being fed back to the core group of healthcare professionals for confirmation, where they will be discussed and restructured until agreement is reached.
Timepoint [1] 370829 0
Intervention conclusion

Eligibility
Key inclusion criteria
Currently undergoing treatment in the recruiting service for a mental illness as defined the DSM-5 or ICD-10

Owns an Android or iOS compatible smartphone
Minimum age
18 Years
Maximum age
No limit
Sex
Both males and females
Can healthy volunteers participate?
No
Key exclusion criteria
1) does not own a smartphone
(2) insufficient command of the English language to be able to understand the instructions
(3) has a document cognitive impairment, substance abuse disorder, or hazardous use of alcohol
(4) has been judged as unable to participate in the trial by their participating treating health care professional.

Study design
Purpose of the study
Prevention
Allocation to intervention
Randomised controlled trial
Procedure for enrolling a subject and allocating the treatment (allocation concealment procedures)
central randomisation by computer
Methods used to generate the sequence in which subjects will be randomised (sequence generation)
Simple randomisation using a randomisation table created by computer software
Masking / blinding
Who is / are masked / blinded?



Intervention assignment
Parallel
Other design features
Phase
Not Applicable
Type of endpoint/s
Efficacy
Statistical methods / analysis
Statistical analysis will be performed after collecting data from all participants in this study. There will be no interim analysis. The analysis methodology will follow a hierarchical framework with multiple stages of information processing: from (i) app and sensor data (e.g. location change events); to (ii) low-level features (e.g. type of movement); to (iii) high-level behavioural markers (e.g. cognition, emotion and behaviour); to (iv) clinical measures (e.g. depression, anxiety, etc.). Statistical machine learning (ML) techniques will be utilised to jointly model the objective mobile stream data, the EMA responses, and the assessment measures. The aim is to develop predictive models both to estimate EMA responses from stream data, and to detect early changes in the assessment measures from stream data and EMA responses.

Baseline model: Chiefly, a hierarchical regression model will be used for prediction, with factors adjusted for population effects, latent class effects and individual effects. Other ML techniques will also be evaluated for efficacy, possibly including bespoke models, hidden Markov models (HMMs), decision trees and neural networks.

Feature engineering: As part of the hierarchical analysis framework, we will have a knowledge driven approach and also a data driven approach to identify suitable ML features. For the knowledge based approach we will consider previously reported features (digital phenotypes) from mobile data sources, covering the spectrum of behavioural, cognitive and emotional dimensions. In this approach, information from the app and sensor data streams (e.g. GPS location, encrypted call and SMS logs, accelerometer and gyroscope, etc.) will be selected as features when they have a demonstrated applicability to any of the above dimensions. For example, where the person is and how they are moving from location to location will be useful as features, since it is established that movement is related to mood fluctuations as an aspect of the emotional dimension. Data driven approaches for feature engineering will also be considered, such as clustering and self-organising maps.

Feature selection: A number of means to reduce the high dimensionality of potential features will be considered. For example, features may be discarded when: their variance crosses some predetermined threshold; or they have a low weighting in a regression model; or they lead to low probability values or poor predictive capability.

Model evaluation: We will try a variety of predictive models, such as bespoke models based on understanding of relationships, HMMs for temporal relationships, decision trees for sparse events, and neural networks for deep, nonlinear relationships. We will quantitatively compare the relative performance of each model, and select one or more models with best performance and ease of understandability. Each model will have its set of parameters learned from the training data set by optimising an appropriate objective function. The performance of each trained model will be evaluated for robustness against a cross-validation data set.

Recruitment
Recruitment status
Not yet recruiting
Date of first participant enrolment
Anticipated
30/06/2019
Actual
Date of last participant enrolment
Anticipated
3/12/2019
Actual
Date of last data collection
Anticipated
8/06/2020
Actual
Sample size
Target
180
Accrual to date
Final
Recruitment in Australia
Recruitment state(s)
SA

Funding & Sponsors
Funding source category [1] 302691 0
Government body
Name [1] 302691 0
NHMRC Medical Research Future Fund
Country [1] 302691 0
Australia
Primary sponsor type
University
Name
Flinders University
Address
College of Medicine and Public Health, Flinders University, Tonsley, GPO Box 2100, Adelaide, SA 5001,
Country
Australia
Secondary sponsor category [1] 302621 0
None
Name [1] 302621 0
Address [1] 302621 0
Country [1] 302621 0

Ethics approval
Ethics application status
Approved
Ethics committee name [1] 303305 0
The Southern Adelaide Clinical Human Research Ethics Committee
Ethics committee address [1] 303305 0
L6/C6 room 2019
Flinders Medical Centre
Flinders Drive, Bedford Park
5042
SA
Ethics committee country [1] 303305 0
Australia
Date submitted for ethics approval [1] 303305 0
04/05/2018
Approval date [1] 303305 0
06/06/2018
Ethics approval number [1] 303305 0
165.17

Summary
Brief summary
This project will aim to determine the feasibility of using smart phone based applications in a clinical population with severe mental illness, and to determine whether the information gathered from mobile devices can detect and predict deteriorating mental health. Two groups of participants will be recruited from mental health services across South Australia; primary healthcare professionals, and patients. Primary healthcare professionals (clinicians) will be recruited for the purposes of overseeing patient recruitment, and determining eligibility. Patients will be asked to download a smartphone app called Mindtick, and will be asked to use the app for the duration of the study. Through the app, they will be prompted (via push notification) a number of times in a week to answer 8 questions relating to a number of variables; (mood, sleep, cravings, level of functioning, and disruptions to life events). Additionally, the app will collect sensor data (GPS, call/text logs, etc.) from sensors already built into the smartphone. In addition to the mobile application, an online clinician dashboard has been developed for the purposes of clinician review to monitor and improve engagement with the intervention. The dashboard summarises patient app data, and presents this information in a graphical format for monitoring mental health patient’s symptoms and functioning. We will investigate whether fortnightly responses and statistics in mobile sensor data can predict early the change in the routinely collected traditional assessment measures observed at baseline, 3, and 6 months.
Trial website
Trial related presentations / publications
Public notes

Contacts
Principal investigator
Name 93210 0
A/Prof Niranjan Bidargaddi
Address 93210 0
Digital Psychiatry & Personal Health Informatics, College of Medicine and Public Health, Flinders University, Tonsley, GPO Box 2100, Adelaide, SA 5001,
Country 93210 0
Australia
Phone 93210 0
+61 8 7221 8840
Fax 93210 0
Email 93210 0
niranjan.bidargaddi@flinders.edu.au
Contact person for public queries
Name 93211 0
Lydia OAKEY-NEATE
Address 93211 0
Digital Psychiatry & Personal Health Informatics, College of Medicine and Public Health, Flinders University
1284 South Road, Tonsley SA 5042

GPO Box 2100 Adelaide SA 5001
Country 93211 0
Australia
Phone 93211 0
+61 8 7221 8264
Fax 93211 0
Email 93211 0
lydia.oakeyneate@flinders.edu.au
Contact person for scientific queries
Name 93212 0
Niranjan Bidargaddi
Address 93212 0
Digital Psychiatry & Personal Health Informatics, College of Medicine and Public Health, Flinders University, Tonsley, GPO Box 2100, Adelaide, SA 5001,
Country 93212 0
Australia
Phone 93212 0
+61 8 7221 8840
Fax 93212 0
Email 93212 0
niranjan.bidargaddi@flinders.edu.au

Data sharing statement
Will individual participant data (IPD) for this trial be available (including data dictionaries)?
No
No/undecided IPD sharing reason/comment


What supporting documents are/will be available?

Doc. No.TypeCitationLinkEmailOther DetailsAttachment
2191Ethical approval    377538-(Uploaded-27-05-2019-15-45-35)-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.