Patient and public involvement

The current proposal represents a novel collaboration between the Centre for Addiction and Mental Health (CAMH) and the Community Addictions Peer Support Association (CAPSA), with project leads from both organisations. Following funding acquisition, a CAPSA Advisory Committee was formed to inform study goals and design, study assessment schedule, and knowledge translation and exchange. Advisors also directly informed the peer support protocol and training to be delivered. Following study launch, advisors remain engaged through quarterly progress reports and other communications and peer supporters meet monthly to provide feedback. Notably, peer supporters and other people with lived and living experience of substance use challenges have historically been involved in the development, implementation, and evaluation of BFO, with updates and improvements made to the programme based on regular feedback.21–23

Trial design

The study is a three-arm, parallel group, randomised controlled trial with evaluations conducted biweekly for 8 weeks during the intervention and monthly follow-ups conducted for 6 months after the end of the intervention. The trial has been registered with ClinicalTrials.gov. The protocol manuscript is compliant with the Standard Protocol Items: Recommendations for Interventional Trials statement.24

Recruitment and participants

Participants will be recruited from the CAMH Outpatient Addictions Programme through clinician and self-referral as well as the hospital research registry.

All participants will meet the following inclusion criteria: (1) 18 years of age and above; (2) fluency in English ; (3) understand and willing to comply with study requirements and provide informed consent; (4) meet diagnostic criteria for a current substance use disorder; (5) have used their primary substance of concern within the past 30 days; and (6) are registered with CAMH Addictions Programme outpatient services. Exclusion criteria include: (1) meet diagnostic criteria for current psychosis or mania; (2) psychiatric disorder requiring care more urgently than substance use, including acute intoxication or withdrawal requiring medical attention; (3) significant neurological disorder or physical illness likely to interfere with participation; (4) any known practical factors that would preclude participation, such as extended absence; (5) unable to read at a sixth-grade level.

Procedure

Participants will be approached following standard intake procedures in the CAMH outpatient services. All interested and potential participants will complete a brief telephone eligibility assessment, including an initial screening interview to assess basic demographic and clinical characteristics. Individuals who meet the preliminary eligibility requirements will then complete an in-person or video conference screening interview. Informed consent will be obtained and documented according to hospital standard operating procedures. Eligible participants are enrolled upon their pretreatment assessment, when they complete all baseline measures and are randomised to one of three study conditions. Following randomisation, participants will be approached to complete biweekly study assessments for 8 weeks, a post-treatment assessment and monthly follow-up assessments for 6 months regardless of treatment compliance or attrition status. All outcome assessments will be conducted by research team members blind to condition assignment. See figure 1 for the study flow chart. The study start date was 25 October 2021 and the anticipated study end date is 26 January 2024.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Participant flow chart. CAMH, Centre for Addiction and Mental Health.

Interventions

All participants will be offered standard care including ancillary services as needed, including psychiatric, pharmacologic, and emergency services. This pragmatic approach prioritises patient safety and has been applied in recent randomised controlled trials of online CBT in substance use.11

Experimental condition A: BFO

BFO is an online intervention designed to provide support to adults struggling with substance use. BFO enables individuals to access tailored clinical content specific to their needs, while maintaining anonymity. It provides psychoeducation and skills building exercises in evidence-based psychosocial intervention strategies, namely CBT and compatible approaches. This programme uses multimedia components through both computerised and mobile application delivery; no coaching or therapist support is provided. Participants actively engage with module content by working through individual exercises in a stepwise manner, which involves reviewing text, audio, and video information, and entering information in a variety of formats (e.g., checkboxes, open text fields) as they reflect on the content and progress towards their goals. The intervention strategies contained within BFO are provided via a six-domain biopsychosocial model, including difficult situations, negative thoughts, emotional impacts, unhelpful behaviours, physical sensations and lifestyle, which are closely aligned with the five factors used in basic CBT models (i.e., environment, thoughts, moods, behaviours, physical reactions25 26). Each domain of the biopsychosocial model corresponds to a module in the BFO programme which contains a psychoeducation ‘information’ strategy, as well as an ‘action’ strategy to facilitate behaviour change. Data captured during each users’ baseline assessment27 28 is used by the programme to populate the model and to provide feedback to the individual on their levels of functioning across the six domains via a traffic-light colour coding system, with green, amber and red, indicating ‘little’, ‘moderate’ and ‘significant’ impairment, respectively (see figure 2 below). The programme guides the user to concentrate on completing strategies aligned to aspects of their functioning with the greatest impairment.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Breaking Free Online interface using the traffic light colour coding system.

Data collection

Table 1 provides a summary of assessment measures as well as primary and secondary outcome measures administered during the trial. All self-report measures will be administered online. Interview measures may be administered in-person or via video conference, according to participant preference. To confirm eligibility, all participants will be clinically characterised by the Structured Clinical Interview for the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (SCID-529), which will be used to define the psychiatric diagnosis for alcohol and substance use disorders and assess the disorder severity over the past month. The SCID-5 has been demonstrated to have excellent reliability and validity.30 Participants will also complete the Contemplation Ladder,31 which is a single-item Visual Analogue Scale ranging from 0 ‘(No thought of quitting. I cannot live without the substance(s)’) to 10 (‘I have changed my substance use and will never go back to the way I used before’). It will be used to assess where participants currently are in thinking about and/or changing their substance use for each substance that they use.

Primary outcome

The primary outcome measure will be change in frequency (days) of substance use (primary substance of concern) over the past 30 days according to the Timeline Follow-Back (TLFB32), a retrospective interview, which assesses substance use over the previous month. The TLFB has accumulated evidence for its validity and reliability.33 34

Secondary outcomes

Secondary outcome measures will include the quantity of substance use (alcohol and cannabis only) over the past 30 days according to the TLFB. Further, the Substance Use Calendar is a self-report questionnaire that will be used to collect self-reports of alcohol and drug use over the previous 14 days, as well as amounts of substance use for the participants’ primary substance of concern, biweekly over the course of acute treatment and monthly over the course of follow-up. This additional measure will promote fulsome assessment of substance use frequency and quantity over both treatment and follow-up.

The Severity of Dependence Scale (SDS35) is a 5-item self-report questionnaire that will be used to assess the level of participants’ dependence on their primary substance of concern over the previous 30 days. Items are scored on a Likert scale of 0 (not at all) to 3 (always or nearly always), with a maximum total score of 15, with higher scores indicating more severe dependence. Research has supported SDS score validity36 and test–retest reliability.37

The Alcohol Use Disorder Identification Test (AUDIT38) is a 10-item self-report questionnaire that will be used to assess alcohol consumption amounts and frequency, alcohol-related issues, and drinking behaviour adapted to assess the past month to facilitate assessment of change (see 39 for an example). The items are scored on a scale of 0–4 with a maximum total score of 40, with higher scores indicating more severe alcohol use issues. AUDIT scores have demonstrated good internal consistency,40 sensitivity, and specificity.41

The Drug Use Disorders Identification Test (DUDIT42) is an 11-item self-report questionnaire that will be used to assess various drug-related issues and identify use patterns also adapted to assess the past month (see 43 for an example). The items are scored on a 3-point scale with total scores ranging from 0 to 44, with higher scores suggesting higher severity of substance use issues. DUDIT scores have demonstrated internal consistency and convergent validity.43

The Patient Health Questionnaire-9 (PHQ-944) is a 9-item self-report questionnaire that will be used to assess participants’ symptoms of depression over the previous 14 days. Items are scored on a scale of 0–3, with a total score range between 0 and 27 and higher scores indicating a higher level of depression symptom severity. PHQ-9 scores demonstrate good sensitivity and specificity, as well as internal consistency and reliability.45 46

The Generalised Anxiety Disorder 7-item (GAD-747) is a self-report measure that will be used to assess participants’ symptom severity of GAD over the previous 14 days. The items range on a scale of 0–4, with total scores ranging from 0 to 21 and higher scores indicating more severe GAD symptoms. GAD-7 scores demonstrate good reliability, validity and sensitivity.48

The WHO Quality of Life Brief (WHOQoL-BREF49) is a 26-item self-report questionnaire that will be used to assess participants’ quality of life over the previous 4 weeks. The items are scored on a 5-point Likert scale, with higher total scores indicating a higher quality of life. The WHOQoL-BREF has been demonstrated to have good internal consistency50 and validity.51

Additional measures will be administered to support the evaluation of healthcare costs and exploratory analyses of moderators and mediators of outcomes suggested by lived experience advisors. The Programme and Client Costs—Substance Abuse Treatment52 is a clinician-administered interview that will be used to monitor services provided to the participants as part of their standard treatment at CAMH over the past 7 days and associated costs. The Brief COPE Scale53 is a 28-item self-report questionnaire that will be used to assess frequency of use of different coping strategies in response to stressors. The Adult Hope Scale54 is a 12-item self-report questionnaire that will be used to assess current level of hope. The Positive and Negative Affect Schedule—Expanded Form55 is a 60-item self-report questionnaire that assesses positive and negative affect over the past 30 days, as well as additional specific affects. The Treatment Acceptability/Adherence Scale56 is a 10-item self-report questionnaire that assesses treatment acceptability and anticipated or actual adherence in response to a given treatment. The WHO Disability Assessment Schedule V.2.057 is a 12-item self-report measure assessing overall level of functioning over the past 30 days.

Finally, participants will complete a semi-structured interview post-treatment following a qualitative description design.58 59 Interviews will have a two-pronged focus including (1) barriers and facilitators to the implementation of online CBT and (2) a complementary ethnographic focus on ‘health work’ patients engage in.60 61 This dual focus permits a complex understanding of the impact of online CBT and explores ways to optimally deliver support during the care journey. Thematic analysis of interview data will be undertaken to offer a summative description of how patients give meaning to their experience in light of the intervention received.

Compensation

Participants will receive honoraria in the form of gift cards for study assessments in the following amounts (1) $C25 for completing the pretreatment and post-treatment study assessment, and (2) $C5 for each monthly follow-up assessment, (3) $C25 for completing the 6-month follow-up assessment and (4) a $C25 bonus for completing all assessments, for a maximum of $C125.

Withdrawal

Patients will be withdrawn prematurely from the study if, at any point during the trial, they ask to exit or they meet any of the study’s exclusion criteria. Patients will be considered for withdrawal from study treatment and assessment procedures if they evidence any urgent medical or mental health issues. Participant attrition and adverse events will be closely monitored and reported to the institutional ethics board as appropriate. Attrition is defined as failure to complete study outcome assessment measures (i.e., a complete loss to follow-up). The research team will track this electronically and also provide a fulsome description of missing data patterns during knowledge dissemination. Adverse events are defined as harmful side effects that a patient experiences when participating in a study, which may or may not be related to the treatment received, and which are classified as mild, moderate or severe. The research team queries adverse events at the beginning of every study visit, which are tracked on a detailed log and reported to the Institutional Review Board on an annual basis (except in the case of severe events, which are reported within 48 hours). A thorough assessment of participants’ baseline health will be completed; determination of any adverse events will be in reference to baseline health for significant changes.

Sample size

A power calculation for a randomised trial indicated that a sample size of n=60 participants per study condition (N=180 in total) will be required to locate a small-to-medium effect size (Cohen’s f=0.13) with a power of 80% (for main effects or interactions), an alpha level of 0.05, and assuming a pre–post correlation of 0.50. This power analysis was conducted using G*Power62 and was based on based on the power necessary to detect the interactions between pre–post changes in the primary outcome and the planned comparisons among the intervention groups (see proposed analyses below). We anticipate an attrition rate of 20%, and thus will randomise an attrition-adjusted sample size of 225 participants (75 per condition) to compensate for the impact of attrition on statistical power, although we will conduct an intent-to-treat analysis using data from all participants randomised (see below).

Randomisation

Randomisation will be stratified according to sex (male, female) and by diagnosis (Alcohol Use Disorder (AUD) only, AUD and/or other Substance Use Disorders [SUDs]) in randomly generated block sizes. Randomisation will be overseen by the CAMH Biostatistical Consultant. Randomisation is stratified to ensure comparable distribution of these participant variables across treatment groups, and to facilitate our capacity to conduct exploratory analyses across these important participant features. Allocation will be concealed from the study coordinator, with the use of sequentially numbered, opaque, sealed envelopes.

Statistical analyses

To evaluate efficacy, we will use a mixed effect model, wherein time will be entered as a within-subject factor with two levels (pretreatment, post-treatment), and study condition (between-subject factor) will be represented by two orthogonal contrasts comparing each intervention condition to the control condition (i.e., BFO vs CM and BFO+PS vs CM). The primary hypothesis will be tested via the interaction between time and each contrast. Random intercepts will be specified to capture variability across individual participants. We will follow the intention-to-treat principle and analyse all randomised participants. Mixed effect models handle missing values by maximum likelihood estimation, which is able to use all available information from randomised participants, and has been associated with the least bias in a comparison of methods for dealing with missing data.63 We plan to examine several predictors (e.g., demographics, treatment compliance, baseline severity of SUD, comorbid symptomatology) of both the missing data on the outcome and the outcome itself and will include any significant predictors as auxiliary variables in the analysis. When auxiliary variables are included in Full Information Maximum Likelihood analyses, bias due to missing data can be substantially reduced.64 We will examine patterns of missing data in the entire data set to determine if the Missing at Random assumption is tenable,65 although we will include auxiliary variables, if necessary, based on their correlation with missingness or the outcome itself. We will also examine whether the assumption of normality is met and use a robust estimator or model the outcome as a count variable with a Poisson or negative binomial distribution using a log link function where appropriate.

To evaluate change in secondary outcomes, we will follow the same procedure, and will also conduct post hoc models that include a contrast for BFO+PS versus BFO to directly compare the two intervention conditions. To evaluate durability of effects over the extended follow-up period, we will examine the interactions between treatment group contrasts and the trajectory of outcomes from post-treatment through the monthly follow-ups.

To evaluate the cost-effectiveness of the intervention for the primary outcome (change in frequency of substance use over the past 30 days), we will undertake an economic evaluation in line with the Canadian Agency for Drugs and Technologies in Health’s Guidelines for the Economic Evaluation of Health Technologies.66 We will take the perspective of public third-party payer (i.e., the healthcare system). We will collect all relevant costs associated with delivering each arm of the intervention. All costs will be reported in 2024 Canadian dollars, using the Consumer Price Index for Health and Personal Care.67 The time frame of the cost-effectiveness analysis will be up to the last follow-up and the primary outcome of the cost-effectiveness analysis will be the incremental cost-effectiveness ratio. To determine whether the intervention is considered cost-effective or not, we will assume a willingness to pay a threshold of $C50 000 per life year gained, as typically done within the Canadian context. Relevant sensitivity analyses will be performed to evaluate the robustness of the results. Exploratory analyses will include a cost-utility analysis, where we will determine quality-adjusted life-years in conjunction with treatment costs associated with each treatment condition.

To evaluate the mediational role of engagement, we will use path analysis within MPlus,68 applying maximum likelihood estimation and specifying condition as the independent variable (using the same two orthogonal contrasts), engagement at post-treatment as the mediator, and substance use frequency at follow-up as the outcome (controlling for pretreatment substance use frequency). We will assess engagement using a multifaceted measure in line with recommendations.69 We will use 10 000 bootstrap resamples to provide stable estimates of direct, indirect, and total effects. To compare pretreatment participant characteristics across conditions, analyses of variance will be conducted for continuous variables and χ2 tests for categorical variables. To compare intensity and type of standard care across conditions, similar analyses will be conducted.

Data statement

The project will use the standard CAMH Neuroinformatics Platform processes to make data available, which adhere to the highest standards of privacy and security, including anonymisation frameworks and data sharing processes.