The trial is a two-arm, parallel group (1:1 ratio) superiority RCT that will determine the efficacy and cost-effectiveness of the Keep it Real web-based program, compared to an ICW. Assessments are completed at baseline, and at 3, 6, 9, and 12 months follow-up. The trial has ethics committee approval, is registered (ACTRN12618001107213), and follows Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) research protocol guidelines (see supplementary materials) as well as Consolidated Standards of Reporting Trials (CONSORT) eHealth guidelines (see Figure 1) [42,43].

This project will utilize a secure, automated web-based project and data management system for RCTs. A two-step recruitment and informed consent process will be used. Young people aged 16-25 years who have used cannabis in the past month will be first asked to complete an online survey on Cannabis and Psychotic Experiences, and repeat the survey 6 and 12 months later. They are recruited online via university student emails and posts and paid advertising to social media (eg, Facebook) and substance use-related websites (eg, Cannabis Information Service) containing a link to an online consent form and survey. Snowballing techniques will be used and participants will be offered a reimbursement incentive of Aus $10 for referring a friend to the study. Consenting participants (via a checkbox) self-complete all baseline and follow-up assessments online. Survey logic automatically screens participants for eligibility. Participant contact details (name, email address, mobile number) and the IP address of respondents will be manually checked to identify repeat or duplicate surveys. Participants who give contact details that appear false, for example, multiple similar names or email addresses, will be excluded. All baseline, 6-, and 12-month survey completers will be entered in a draw to win 1 out of 10 Aus $100 gift vouchers.

Young people are asked if they are interested in trialing the Keep it Real website (via a checkbox) on the online survey consent form, and again at the end of the baseline survey. Those who express interest in trialing the website and score a minimum of 18 on the CAPE-15, which is indicative of at least three PEs “sometimes” or one PE “nearly always” in the past 3 months, are eligible to participate in the RCT. They are also required to have internet access and to provide a mobile number and email address. People who use alcohol or other illicit drugs are included as long as cannabis is the most frequently used drug (other than tobacco). Exclusion criteria will be (i) history of traumatic brain injury or organic brain disease, (ii) current acute suicide risk, and (iii) insufficient English fluency. Eligible participants are automatically sent the consent form for the RCT, and those who provide consent (via a checkbox) are randomized and sent an email and SMS link to either the Keep it Real web-based program or information-only control website. They are not informed about which website program they are allocated to. The study information sheet states that both programs provide information on PEs, psychosis, and cannabis use. They differ by their level of interactivity.

Young people who are ineligible for the RCT remain participants in the original survey study and are provided with immediate access to the Keep it Real website. Those in the RCT are automatically sent follow-up surveys at 3, 6, 9, and 12 months postbaseline via SMS and email. Noncompleters will be contacted via telephone by outcome assessors blind to treatment allocation to remind them to complete the survey online or over the phone. Participants are reimbursed Aus $20 for completing each follow-up survey (maximum of Aus $80). If a participant chooses to withdraw from the study, they will still be able to continue using Keep it Real or the ICW. The trial statistician will be blind to treatment group allocation.

All measures are self-reported and the follow-up frequency is provided in Table 1. Demographic variables include sex, gender, occupation, education, employment, living arrangements, urbanicity, relationship status, migration status, ethnicity, estimated hours of internet use/day, family and personal history of mental and substance use disorders, and history of traumatic brain injury or organic brain disease.

The primary outcome measures for cannabis use and psychotic like experiences are described below:

A number of secondary outcomes will also be measured, as outlined below:

Health-related quality of life will be assessed using the 5-dimension version of the European Quality of Life 5-level (EQ-5D-5L) scale [61]. Information on other health care service use including medical and psychological treatment for substance use and other issues in the month prior to each assessment point will be collected.

A number of other measures will also be included in the survey:

A computerized random number sequence generator incorporating random permuted blocks will be used to allocate participants to: (i) Keep it Real or (ii) ICW. Stratification will be by cannabis usage (daily or less), sex, age (16-20 years and 21-25 years), and psychosis screen result (present vs not present). The randomization sequence is concealed within the secure website and the allocation is automatically released when full eligibility criteria are met.

Young cannabis users who meet RCT inclusion criteria and consent to participate are automatically randomized to either the Keep it Real or ICW program, and are sent an email and SMS link to the relevant program. Both programs are accessible via login on any computer, tablet, or mobile device with an internet connection. They are password protected and are sited on parallel, secure servers. Both programs are self-guided (fully automated), and all website modules are unlocked and can be accessed by the user at any time. SMS reminders to access both programs will be sent on day 7, 14, and 21 after baseline. Participants’ “backend” usage data will be automatically captured by the website, including logins, modules accessed/completed, inserted text, and dates of data entry. All participants have continuous access to the Get Help section of the website, which provides the contact details of mental health and AOD services (including crisis care).

Weekly meetings monitor project implementation, clinical (including study withdrawals, adverse events), and research integrity (including data safety, important harms or unintended effects, blindness violations). Email or telephone consultations will be used to determine urgent issues, and a record of decision precedents will be kept to ensure consistency. Suicide and psychosis safety protocols identify and manage risk at all survey time points. Participants with a positive psychosis screen or who report a low level of suicide risk (MINI Suicidality Scale score of 1-5) receive a message in the online survey and an email providing details of appropriate support services and helplines. They are also asked to contact the research team if they have additional questions or would like assistance finding support. Those who report a moderate (Suicidality Scale score of 6-9) or high (10+) level of suicide risk are asked if they are receiving adequate support and if their feelings have recently improved (due to the 1-month time frame of the measure). If the participant answers yes to both questions, no further action is taken. If they answer yes to one of these questions or request further support, they are called within 48 hours by a research assistant to assess how further support can be provided (using a call script). Those who report a suicide plan or a suicide attempt within the past month are contacted by a trained clinical psychologist, who assesses the level of risk and helps the participant connect with appropriate supports.

Power calculations to determine sample size were based on the effects found in the Keep it Real pilot study, which found a moderate effect size for reductions in the frequency of cannabis use (d=0.46). Due to the uncontrolled nature of the pilot study, a conservative approach using small to moderate effects was used. To measure a moderate effect size of f=.20, alpha (α) set at 0.025, and power set at 0.95, we would require 191 participants per group (a total of 382 participants). If baseline covariates accounting for about 40% of the variability in the outcome model were included in the calculations, then we would have 0.99 power to detect a small to moderate treatment effect. We predict a 20%-30% attrition rate at 12 months based on our pilot data and previous work and will therefore need to randomize a total of 498 participants.

The pilot study recruited 1089 past-month cannabis users over a period of 6 months (using the same recruitment strategies) to identify 860 monthly cannabis users meeting our PE inclusion criterion (CAPE-15≥18). Of these, 295 (34%) expressed interest in participating in the Keep it Real pilot trial (66 [7%] no interest; 503 [58%] no response) and 213 (72% of those interested; 25% of the total sample) were randomized. We will, therefore, need to recruit 1992 cannabis users in order to randomize 498 participants (see Figure 1).

The independent variable is a treatment condition with two levels: (i) Keep it Real and (ii) the ICW. The primary outcome variable is the frequency and average amount of cannabis/day (in 1/4 gram standard units) in the past month (on the OTI) and the frequency of PEs (CAPE-15, including subscales) in the past 3 months. Secondary outcome variables include the typical quantity and frequency of alcohol, tobacco, cannabis, and other drug use and related problems in the past 3 months (ASSIST), cannabis knowledge, PE-related distress (CAPE-15), cannabis intoxication experiences (CEQ-I), severity of cannabis dependence (SDS), depression (PHQ-9)/anxiety (GAD-7) symptoms, suicidality (MINI), mental well-being (MHC-SF), and functioning (MAFS). To determine whether there are group differences in the primary and secondary outcome measures at 3, 6, 9, and 12 months, a series of mixed effects models for repeated measures (MMRM) will be employed. The within-groups factor will be time (baseline, 3, 6, 9, 12 months) and group (Keep it Real, ICW) will be the between-subjects factor to examine the timexgroup interaction. This technique can also control for potential confounds (eg, other drug use and related problems) and examine potential baseline moderators of intervention effects (eg, age, gender, childhood trauma [CTQ], impulsivity [SUPPS-P], urbanicity, cannabis knowledge, lifetime psychotic disorder). Missing data will be handled using full information maximum likelihood estimation and an intention-to-treat analysis is performed.

A two-pronged economic analysis comparing Keep it Real and ICW will be conducted, taking a societal perspective, including costs of treatment provision and intervention (including website development and maintenance) as well as productivity costs. The analysis will display the cost-effectiveness in both natural units and as QALYs (based on the EQ-5D-5L). Models will assess Keep it Real’s offset on reduced health care service use and productivity losses are avoided. In a second modeled cost-effectiveness analysis (CEA), the estimated change in cannabis consumption from the trial will be extrapolated to the general Australian population of 16-25-year olds. Uncertainty will be assessed using a combination of nonparametric bootstrapping for the effect size (cannabis frequency) and Monte Carlo simulation for all other variables (with sampling uncertainty).

Large numbers of young cannabis users report PEs, which increase their risk of developing psychotic, substance use, and depressive/anxiety disorders. Current services are failing to engage these young people—and would struggle to meet the need if they did seek help. An effective, safe, low-cost, nonstigmatizing accessible intervention is needed. Our pilot data showed that the Keep it Real program resulted in substantial reductions in both PEs and cannabis use, and young people found the program engaging and easy to use. This study protocol describes a large RCT that will determine if Keep it Real is more efficacious and cost-effective than the minimal web-based information they may otherwise receive.

Engaging large numbers of young nonhelp-seeking cannabis users into web-based treatment is challenging, as they are unlikely to view their cannabis use as problematic or responsible for their PEs. However, the pilot study clearly demonstrated the feasibility of recruiting young cannabis users with PEs (79% of survey completers), with 72% of the 34% who expressed interest agreeing to be randomized to the trial. The pilot study also highlighted the acceptability of Keep it Real, as young people gave the program an average rating of 4 out of 5 for overall objective quality, and the functionality, aesthetics, and information subscales and a 3 out of 5 rating for engagement. Moreover, the pilot study provided preliminary evidence for the efficacy of Keep it Real as well as valuable data on the participation/retention rates and effect sizes to inform sample size which was used to ensure that this RCT is well powered.

Young people will be asked if they were interested in participating in the RCT prior to receiving feedback on their PEs. While this recruitment strategy will ensure that all young people are given the opportunity to participate, it may increase the risk of selection bias by potentially excluding those with little insight into their PEs. Nevertheless, the inclusion of people with a lifetime history of a psychotic disorder and the stratification of the randomization sequence by psychosis screen result may help reduce such risk of bias.

The inclusion of a range of potential control variables and moderators of treatment outcomes is a further strength of the study. However, mediators were not included due to concerns about potential assessment reactivity, the existing length of the assessment schedule, and participant burden. Keep it Real participants are able to monitor their PEs and substance use within the program. The extent of monitoring observed in this study will provide an indication of young cannabis users' willingness to engage in regular assessments for future studies exploring potential mediators of change.

The number of modules in the Keep it Real and ICW programs are not matched, but the ICW program contains the same number of information sheets as the Keep it Real program. While it could be argued that the smaller amount of content in the ICW may be insufficient to produce an effect, this control condition was designed to control access to a web-based program and the minimal amount and type of information users may otherwise receive online. Nevertheless, it is possible that the overlap in content may wash out any differential treatment effects between programs.

Web-based programs are ideal platforms for delivering self-guided treatments to nonhelp-seeking young people with PEs. Keep it Real is the first web-based program to target PEs in cannabis users. The program content is based on evidence-based CBT treatments for people at ultrahigh risk for psychosis (which includes a subgroup with PEs) [41] and motivational interviewing interventions for cannabis use [37,38]. Keep it Real will be freely available and users will have unlimited 24/7 access to the program, including the PE and substance use monitoring and feedback, enabling the user to engage in continuous self-management. The program also has broad applicability as it can be used as a stand-alone treatment or as an adjunct to usual treatment for help-seeking cannabis users with PEs accessing clinical services. Different components of the program can also be used to address cannabis, alcohol, methamphetamine, or heroin use, PEs, cannabis intoxication effects, and to provide coping skills training. If Keep it Real demonstrates a substantial and cost-effective impact on cannabis use and PEs, its low unit cost and scalability would mean that it has substantial potential to make a population-wide impact on both immediate and longer term consequences of cannabis use and their significant social, economic, and health costs.