Trial Review

Technical difficulties have been reported by some users of the search function and is being investigated by technical staff. Thank you for your patience and apologies for any inconvenience caused.

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
ACTRN12623001056684
Ethics application status
Approved
Date submitted
22/08/2023
Date registered
29/09/2023
Date last updated
31/05/2024
Date data sharing statement initially provided
29/09/2023
Type of registration
Prospectively registered

Titles & IDs
Public title
Using Artificial Intelligence to support a self-help intervention for
perfectionism
Scientific title
Using Artificial Intelligence to support a self-help intervention for
perfectionism: a randomised controlled trial with undergraduate students
Secondary ID [1] 310434 0
Nil known
Universal Trial Number (UTN)
U1111-1296-9628
Trial acronym
Linked study record

Health condition
Health condition(s) or problem(s) studied:
Perfectionism 331209 0
Eating Disorders 331210 0
Depression 331211 0
Anxiety 331212 0
Stress 331213 0
Condition category
Condition code
Mental Health 327976 327976 0 0
Eating disorders
Mental Health 327977 327977 0 0
Depression
Mental Health 327978 327978 0 0
Anxiety

Intervention/exposure
Study type
Interventional
Description of intervention(s) / exposure
"Changing Perfectionism: An Artificial Intelligence (AI) Assisted Self-Help Guide" (Egan, Shafran, Wade, Johnson, Carlbring & Raghav). This is an 11 module self-help booklet (pdf) to target perfectionism, a shared risk factor for anxiety, depression and eating disorders. Module topics include identifying perfectionism, pros and cons of perfectionism, experiments to challenge perfectionism, and changing self-criticism. Participants will have access to one of two versions of the self-help workbook online for four weeks, which will be automatically released to them once they register for and consent to the study, and fill out the first survey. For the active intervention group, support will be provided by Artificial Intelligence tools (e,g, chatGPT). The workbook contains new instructions on how best to use artificial intelligence (AI) prompts to assist with questions that arise during the intervention. Example questions to ask AI are also given at the end of each module. The control group will receive the same booklet without the AI instructions, and with a request to avoid using AI for the workbook. Each module takes approximately 30 minutes to complete, totalling 5.5 hours for the 11 module workbook. Within the 4-week period, participants can progress at their own pace. We have added an adherence check regarding how many modules were completed to the post-intervention survey.
Intervention code [1] 326836 0
Treatment: Other
Intervention code [2] 326837 0
Prevention
Comparator / control treatment
Active control design. The control group now receives the 11-module perfectionism intervention booklet at the same time as the intervention group, but with the AI instructions removed. This group is also instructed not to use AI to assist this task. This is checked with a new question in the post-intervention survey.
The changes in design from case series to RCT occurred after only one participant took part in the case series in late 2023. This update has been requested prior to recruitment of any participants for the RCT.
Control group
Active

Outcomes
Primary outcome [1] 335830 0
Usage of AI Tools for Support
Will be assessed using a survey question developed specifically for this study:
Did you use Artificial Intelligence (AI) Tools as support during this intervention?
Never – rarely – sometimes – often – always (every module)
Timepoint [1] 335830 0
Immediately post intervention (4 weeks post baseline survey)
Primary outcome [2] 335933 0
Usefulness of AI Tools for Support
Will be assessed using a survey question developed specifically for this study:
How useful did you find AI Tools as support during this intervention
Not at all useful - slightly useful - somewhat useful - quite useful - extremely useful
Timepoint [2] 335933 0
Immediately post intervention (4 weeks post baseline survey)
Primary outcome [3] 335934 0
Useful elements of AI Tools for Support
Will be assessed using a survey question developed specifically for this study:
What did you like most/least about using AI Tools for support in this study?
Free text
Timepoint [3] 335934 0
Immediately post intervention (4 weeks post baseline survey)
Secondary outcome [1] 426175 0
Acceptability of AI instructions will be assessed using a survey question developed specifically for this study:
At the start of the workbook, we provided general tips on how to use AI Tools most effectively. How would you rate these instructions?
Not at all useful - slightly useful - somewhat useful - quite useful - extremely useful
Timepoint [1] 426175 0
Immediately post intervention (4 weeks post baseline survey)
Secondary outcome [2] 426176 0
Acceptability of AI instructions will be assessed using a survey question developed specifically for this study:
After each module, we provided sample questions for AI support. How would you rate these instructions?
Not at all useful - slightly useful - somewhat useful - quite useful - extremely useful
Timepoint [2] 426176 0
Immediately post intervention (4 weeks post baseline survey)
Secondary outcome [3] 426743 0
Acceptability of AI instructions will be assessed using a survey question developed specifically for this study:
Will be assessed using a survey question developed specifically for this study:
Do you have any suggestions for improving our instructions regarding use of AI Tools?
(Free text)
Timepoint [3] 426743 0
Immediately post intervention (4 weeks post baseline survey)
Secondary outcome [4] 426744 0
Acceptability re AI support.
Will be assessed using a survey question developed specifically for this study:
Were AI Tools better than no support (i.e., just reading the information?)
Yes - No - Unsure
Timepoint [4] 426744 0
Immediately post intervention (4 weeks post baseline survey)
Secondary outcome [5] 426745 0
Acceptability re AI support.
Will be assessed using a survey question developed specifically for this study:
If you had a choice, would you prefer
AI support
Automated reminders to continue working through modules
Human support (on demand – as needed)
Human support (regular, scheduled support)
No support (unguided)
Not sure
Timepoint [5] 426745 0
Immediately post intervention (4 weeks post baseline survey)
Secondary outcome [6] 426746 0
General feedback
Will be assessed using a survey question developed specifically for this study:
Do you have any other comments, suggestions or feedback regarding this study?
Timepoint [6] 426746 0
Immediately post intervention (4 weeks post baseline survey)
Secondary outcome [7] 426749 0
Clinical Perfectionism
measured using Clinical Perfectionism Questionnaire
Timepoint [7] 426749 0
Weekly progress measure during intervention: Baseline, 1,2,3,4-weeks post baseline.
Modifications
1. Removal of requirement for submitting weekly measure to team (prior to recruitment)
2. Removal of requirement for 1-month follow-up for participants enrolling from 4 April 2024
Secondary outcome [8] 426757 0
Perfectionism
using Frost Multidimensional Perfectionism Scale (Concern over mistakes/Personal standards subscales)
Timepoint [8] 426757 0
Baseline, immediately post intervention (4-weeks post baseline), 1-month follow-up.
Modification: Removal of requirement for 1-month follow-up for participants enrolling from 4 April 2024
Secondary outcome [9] 426759 0
Stress
Depression Anxiety Stress Scale (short form) - Stress subscale
Timepoint [9] 426759 0
Baseline, immediately post intervention (4-weeks post baseline), 1-month follow-up.
Modification: Removal of requirement for 1-month follow-up for participants enrolling from 4 April 2024
Secondary outcome [10] 426760 0
Anxiety
GAD-7 - Generalised Anxiety Disorder Scale
Timepoint [10] 426760 0
Baseline, immediately post intervention (4-weeks post baseline), 1-month follow-up.
Modification: Removal of requirement for 1-month follow-up for participants enrolling from 4 April 2024
Secondary outcome [11] 426761 0
Depression
PHQ-9 Patient Health Questionnaire
Timepoint [11] 426761 0
Baseline, immediately post intervention (4-weeks post baseline), 1-month follow-up.
Modification: Removal of requirement for 1-month follow-up for participants enrolling from 4 April 2024
Secondary outcome [12] 426762 0
Disordered Eating
EDE-Q7 Eating Disorder Examination Questionnaire -Brief version
Timepoint [12] 426762 0
Baseline, immediately post intervention (4-weeks post baseline), 1-month follow-up.
Modification: Removal of requirement for 1-month follow-up for participants enrolling from 4 April 2024
Secondary outcome [13] 434563 0
Use of AI tools as support will be assessed using a survey questions developed specifically for this study: Did you use Artificial Intelligence (AI) Tools as support during this intervention? If you were in the control group, but did use AI tools to help with the workbook- no problem! Your honest answer here is helpful for interpreting results
5-point Likert scale, Never – rarely – sometimes – often – always (every module)
Timepoint [13] 434563 0
Immediately post intervention (4 weeks post baseline survey)
Secondary outcome [14] 434564 0
Use of AI tools as support will be assessed using a survey questions developed specifically for this study: Did you use Artificial Intelligence (AI) Tools as support during this intervention? If you were in the control group, but did use AI tools to help with the workbook- no problem! Your honest answer here is helpful for interpreting results
5-point Likert scale, Never – rarely – sometimes – often – always (every module)
Timepoint [14] 434564 0
Immediately post intervention (4 weeks post baseline survey)
Secondary outcome [15] 434565 0
Module completion will be assessed using a survey question developed specifically for this study: We are interested to know how many modules you completed during the 4-week period. There is no judgement here – we know life is busy! Your honest answer helps us to know how much of the program is needed to be helpful, even if you didn’t complete it. You can estimate if you can’t remember.
I finished the following number of modules (0-11)
Timepoint [15] 434565 0
Immediately post intervention (4 weeks post baseline survey)
Secondary outcome [16] 434566 0
Module completion will be assessed using a survey question developed specifically for this study: We are interested to know how many modules you completed during the 4-week period. There is no judgement here – we know life is busy! Your honest answer helps us to know how much of the program is needed to be helpful, even if you didn’t complete it. You can estimate if you can’t remember.
I finished the following number of modules (0-11)
Timepoint [16] 434566 0
Immediately post intervention (4 weeks post baseline survey)

Eligibility
Key inclusion criteria
17-29 years of age, fluent in English, Self-identifying as a perfectionist (or think you might be: Unhealthy perfectionism = harsh self-criticism, fear of making mistakes, basing your self-worth almost entirely on achievements; different to striving for excellence, which is a good thing").
Minimum age
17 Years
Maximum age
29 Years
Sex
Both males and females
Can healthy volunteers participate?
Yes
Key exclusion criteria
Past involvement with perfectionism research.

Study design
Purpose of the study
Treatment
Allocation to intervention
Randomised controlled trial
Procedure for enrolling a subject and allocating the treatment (allocation concealment procedures)
Concealed (undertaken by Qualtrics software)
Methods used to generate the sequence in which subjects will be randomised (sequence generation)
Automatic (simple) randomisation in Qualtrics survey software
Masking / blinding
Open (masking not used)
Who is / are masked / blinded?



Intervention assignment
Single group
Other design features
Phase
Not Applicable
Type of endpoint/s
Efficacy
Statistical methods / analysis
N = 10 small case series to pilot approach for acceptability/feasibility
Analysis
Quantitative
- Within-person intervention effect sizes (Cohen's d) will be compared to benchmarks from previous studies utilising guided and unguided self-help perfectionism interventions
- Ratings of usefulness of approach and training
- prior usage of AI tools will be measured as a moderator of outcomes using the following baseline survey question specifically developed for this study: How often do you currently use Artificial Intelligence (AI) tools (Never-rarely-sometimes-often-always)
Qualitative
- Suggestions for refinement, organised by theme and frequency
For RCT, the following quantitative analysis will be used, and the qualitative analysis repeated
Linear mixed modelling with be used to compare outcomes between the two groups (waitlist, intervention) together with calculation of effect sizes (Cohen's d; 95% CIs). Power calculation shows that 27 participants per group are required ( alpha = .05, power = .80, attrition rates of T1 = 25% and T2 = 5%, d = .54, correlation = .8)

Recruitment
Recruitment status
Recruiting
Date of first participant enrolment
Anticipated
2/10/2023
Actual
16/10/2023
Date of last participant enrolment
Anticipated
31/10/2024
Actual
Date of last data collection
Anticipated
31/12/2024
Actual
Sample size
Target
54
Accrual to date
1
Final
Recruitment in Australia
Recruitment state(s)
SA,WA

Funding & Sponsors
Funding source category [1] 314641 0
University
Name [1] 314641 0
Flinders University (in-kind support)
Country [1] 314641 0
Australia
Funding source category [2] 314804 0
University
Name [2] 314804 0
Curtin University (in-kind support)
Country [2] 314804 0
Australia
Funding source category [3] 314805 0
University
Name [3] 314805 0
University College London (in-kind support)
Country [3] 314805 0
United Kingdom
Funding source category [4] 314806 0
University
Name [4] 314806 0
Stockholm University (in kind support)
Country [4] 314806 0
Sweden
Primary sponsor type
Individual
Name
Catherine Johnson
Address
College of Education, Psychology & Social WorkFlinders UniversityGPO Box 2100Adelaide SA 5001
Country
Australia
Secondary sponsor category [1] 316605 0
None
Name [1] 316605 0
Address [1] 316605 0
Country [1] 316605 0

Ethics approval
Ethics application status
Approved
Ethics committee name [1] 313662 0
Flinders University Human Research Ethics Committee
Ethics committee address [1] 313662 0
Flinders UniversityGPO Box 2100Adelaide SA 5001
Ethics committee country [1] 313662 0
Australia
Date submitted for ethics approval [1] 313662 0
20/08/2023
Approval date [1] 313662 0
21/09/2023
Ethics approval number [1] 313662 0
Ethics committee name [2] 315207 0
Curtin University Human Research Ethics Committee
Ethics committee address [2] 315207 0
https://www.curtin.edu.au/students/essentials/higher-degree-by-research/ethics-safety/human/
Ethics committee country [2] 315207 0
Australia
Date submitted for ethics approval [2] 315207 0
22/02/2024
Approval date [2] 315207 0
13/03/2024
Ethics approval number [2] 315207 0
HRE2024-0111

Summary
Brief summary
This pilot project tests whether the use of Artificial Intelligence (AI) Chatbots is an acceptable and helpful alternative to human support during an online intervention for perfectionism. Perfectionism is a risk factor for anxiety, depression and eating disorders. Participants (N=54) will be receive a self-guided perfectionism workbook to undertake 11 modules at their own pace over 4 week. Groups will be randomly assigned (27 in each group) to (1) receive instructions on how to use the AI Chatbot of their choice to support the workbook, or (2) requested to not use AI support for this task, Participants in the active group will feed back regarding acceptability of this approach. We will compare the impact on module completion between groups together with mental health outcomes: perfectionism, anxiety, depression, stress and eating disorder risk. Effects will also be compared to similar studies using human guidance.
Trial website
Trial related presentations / publications
Public notes
Slow recruitment in 2023, likely due to nearing end of academic year (one participant of 10 desired; Baseline measures collected, 17.10.23; Follow-up measures, 19.12.23 ). During 2023, we had one participant enrol and complete the study. (Baseline measures, 17.10.23; Follow-up measures, 19.12.23). Study re-evaluated and launched as an RCT for Semester 1, 2024 - higher numbers of participants traditionally sign up during this period allowing for expanded version of study. Recruitment will commence from 26 February, 2024 (start of Semester 1). Given the study design has changed, the single participant from the case series trial will not be utilised for data analysis

Contacts
Principal investigator
Name 128926 0
Dr Catherine Johnson
Address 128926 0
College of Education, Psychology and Social WorkFlinders UniversityGPO Box 2100Adelaide SA 5001
Country 128926 0
Australia
Phone 128926 0
+61 438500538
Fax 128926 0
Email 128926 0
catherine.johnson@flinders.edu.au
Contact person for public queries
Name 128927 0
Dr Catherine Johnson
Address 128927 0
College of Education, Psychology and Social WorkFlinders UniversityGPO Box 2100Adelaide SA 5001
Country 128927 0
Australia
Phone 128927 0
+61 438500538
Fax 128927 0
Email 128927 0
catherine.johnson@flinders.edu.au
Contact person for scientific queries
Name 128928 0
Dr Catherine Johnson
Address 128928 0
College of Education, Psychology and Social WorkFlinders UniversityGPO Box 2100Adelaide SA 5001
Country 128928 0
Australia
Phone 128928 0
+61 438500538
Fax 128928 0
Email 128928 0
catherine.johnson@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
Aggregated data will be publicly available via Open Science Framework


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
No additional documents have been identified.