This is an open-access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Journal of Medical Internet Research, is properly cited. The complete bibliographic information, a link to the original publication on http://www.jmir.org/, as well as this copyright and license information must be included.
After alcohol and tobacco, cannabis is the most widely used psychoactive substance in many countries worldwide. Although approximately one in ten users develops serious problems of dependency, only a minority attend outpatient addiction counseling centers. A Web-based intervention could potentially reach those users who hesitate to approach such treatment centers.
To test the efficacy of a Web-based self-help intervention with and without chat counseling—Can Reduce—in reducing the cannabis use of problematic cannabis users as an alternative to outpatient treatment services.
Altogether, 436 participants were recruited by various online and offline media for the Web-based trial. A total of 308 of these were eligible for study participation and were randomly allocated in an unblinded manner to either self-help with chat (n=114), self-help without chat (n=101), or a waiting list control group (n=93). The fully automated self-help intervention consisted of eight modules designed to reduce cannabis use, and was based on the principles of motivational interviewing, self-control practices, and methods of cognitive behavioral therapy. Additional individual chat counseling sessions were based on the same therapeutic principles. The sessions were conducted by trained counselors and addressed participants' personal problems. The main outcomes were the frequency (number of days) and quantity of cannabis use (number of standardized joints) per week, as entered into the consumption diary at baseline and at the 3-month follow-up. Secondary outcomes included self-reported symptoms of cannabis use disorder, severity of cannabis dependence, risky alcohol use, and mental health symptoms. Intervention participation and retention were extracted from the user progress data and the consumption diary, respectively.
Can Reduce participants were older (U=2.296,
Web-based self-help interventions supplemented by brief chat counseling are an effective alternative to face-to-face treatment and can reach a group of cannabis users who differ in their use and sociodemographic characteristics from those who enter outpatient addiction treatment.
International Standard Randomized Controlled Trial Number (ISRCTN): 59948178; http://www.isrctn.com/ISRCTN59948178 (Archived by WebCite at http://www.webcitation.org/6bt01gfIr)
Web-based self-help programs that aim to reduce cannabis use might help to reach cannabis users who do not want to enter available outpatient addiction counseling services due to their fear of being stigmatized or their need to distance themselves socially from drug counselors [
Treatment demand statistics from Swiss in- and outpatient addiction treatment centers demonstrated a linear increase—from 2006 (9.9%) to 2012 (14.7%)—in new treatment entry cases for whom cannabis was the main problem substance [
An initial meta-analysis included diverse studies that mainly investigated computer- and some Web-based interventions to reduce cannabis consumption and found a small overall effect size (g=0.16, 95% CI 0.09-0.22,
The combination of a fully automated self-help intervention based on the approaches of Rooke et al [
Thus, the current study aims to investigate and compare the efficacy of Web-based self-help interventions—in combination with or without tailored chat counseling based on CBT, MI, and BSM—in reducing cannabis use in problematic cannabis users.
Study participants were recruited by a press release, several websites from local outpatient treatment centers, and from nightlife prevention websites that were linked to the Can Reduce website [
Study inclusion and exclusion criteria are depicted in
Inclusion and exclusion criteria and rationales.
Participant criteria | Rationales | |
|
|
|
|
Minimum age of 18 years | To ensure a minimal age of participation |
|
Read and understand German | To ensure understanding of interventions |
|
Internet access and a valid email address | To ensure participation |
|
Using cannabis at least once a week over the 30 days prior to study entry | To include at least occasional users |
|
|
|
|
Current serious psychiatric disorders or history of psychosis, schizophrenia, bipolar type I disorder, or significant current suicidal or homicidal thoughts | To avoid exacerbation of serious symptoms of these severe psychiatric disorders |
|
Other pharmacological or psychosocial treatments for cannabis use disorders | To avoid confounding treatment effects |
|
For women: pregnancy and breastfeeding | To avoid serious complications resulting, for example, from withdrawal symptoms |
The Web-based self-help intervention, Can Reduce, was based on classical CBT approaches for treating cannabis dependence [
Can Reduce is the first self-help intervention for problematic cannabis users in Switzerland. It was developed by the authors of this publication from the ISGF and the ARUD. Both institutions are located in the Canton of Zurich, Switzerland. Study participation was free of charge. The self-help part of Can Reduce was developed according to the experiences of an earlier study in problematic cocaine users [
The protocol of the RCT was approved by the Ethics Committee of the Canton of Zurich (KEK-StV-Nr. 15/13) and was carried out in compliance with the Helsinki Declaration. Before giving informed consent, participants were informed of the following: (1) the rationale of the study, (2) study inclusion and exclusion criteria (see
There were three different study arms. The first consisted of the Web-based self-help intervention, Can Reduce, in combination with up to two individual chat counseling sessions based on MI and CBT approaches that considered the data the participants entered into the self-help intervention and individual requests. The second study arm consisted of the same intervention but without chat counseling. Study arms 1 and 2 received weekly automated motivational emails to remind the user to log in and fill out the consumption diary. Study arm 3 consisted of a classical waiting list and people in this arm received access to the self-help intervention after 3 months.
The following modules, organized into three main parts, were offered as a Web-based self-help intervention (study arms 1 and 2) and—as long as the participant did not feel an urgent need to skip to a specific module—it was recommended that they should be worked through in the order shown in
Part 1: Introduction
Registration process
Explanation of the "standard cannabis joint" concept and choice of the personal standard cannabis joint (see
Examination of the pros and cons resulting from a change in cannabis consumption patterns and further principles of motivational interviewing to address motivation, followed by setting an appropriate target value for overall cannabis use, which is to be reached by the end of the intervention
Explanation of the
Explanation of the emergency button for immediate responses to frequently asked questions and access to emergency contacts
Part 2: Key Modules (participants are encouraged to complete these modules in the order presented below; see
Module 1: Strategies for goal achievement
Module 2: Identifying risk situations
Module 3: Dealing with cannabis craving
Module 4: Dealing with relapses
Part 3: Further Modules (participants are encouraged to complete at least two, in any order)
Module 5: Tobacco smoking during the reduction in cannabis use
Module 6: Saying "no" to foster refusal skills
Module 7: Dealing with burdens
Module 8: Preserving achievements
The additional (up to two) chat counseling sessions with a scheduled duration of 20 to 30 minutes in study arm 1 supported behavioral change according to MI, discussed the modules of the Web-based self-help part based on MI and CBT, and reviewed the development of the consumption diary. Invitations to chat sessions were sent by the counselors according to a predefined procedure between weeks 1 and 2 for the first and between weeks 4 and 6 for the second chat session. The chats took place within the website in a small box at the bottom right corner, while keeping the content of the webpage in view (see
The chat counselors received quarterly supervision sessions and consisted of trained MI counselors, mainly psychologists or psychiatrists with advanced or completed further education, with at least one year of experience in treating cannabis-abusing patients face to face. Specific quality standards were developed for addiction chat counseling and implemented for this study in the chat counselor supervision based on the study on the development of a European Union framework for minimum quality standards and benchmarks in drug demand reduction treatment quality standards [
To optimize and manage their interactions with clients, counselors had access to a specific user management area to add arranged chat dates, define statuses, and add personal comments about their clients. With this tool, counselors could follow their clients’ progress in reducing their cannabis use through clearly arranged charts, and look up previous chat histories. Specific lists helped counselors track their clients (eg, a list with
The Web-based self-help intervention and the subsequent tailored chat counseling aimed to reduce cannabis use. However, those participants who sought cannabis abstinence were also encouraged to make step-by-step reductions until full abstinence was reached. In accordance with the counselor supervision group, we deviated from the study protocol [
Participants randomized to the waiting list had the opportunity to participate in the Web-based self-help intervention 3 months after registration.
Screenshot of the Can Reduce Web-based intervention, showing the decision on the standard cannabis joint prior to the first consumption diary entry.
Main menu of the Can Reduce Web-based intervention's study arm 1 with self-help plus chat counseling that took place within the website in a small box at the bottom right corner.
This study aimed at comparing the efficacy of a Web-based self-help intervention alone or combined with chat counseling in the reduction of the cannabis use of problematic cannabis users within a three-arm randomized controlled trial with assessments at baseline and 3-month follow-up (see
We hypothesized that Web-based interventions—which are more interactive—would be more effective than less interactive interventions in reducing cannabis use among problematic cannabis users. We tested the following detailed study hypotheses with respect to the main outcome (ie, the reduction of the weekly cannabis used between the baseline and the 3-month follow-up):
Tailored chat-based counseling in combination with Web-based self-help for the reduction of cannabis use (study arm 1) is more effective than the waiting list control condition (study arm 3).
Web-based self-help for the reduction of cannabis use (study arm 2) is more effective than the waiting list control condition (study arm 3).
Chat-based counseling in addition to Web-based self-help for the reduction of cannabis use (study arm 1) exhibits a trend to be more effective than Web-based self-help alone (study arm 2).
The primary outcome measure was the recorded quantity of cannabis use in the previous 7 days, quantified in individually standardized cannabis joint sizes, and as specified in the consumption diary (see
The following secondary outcome instruments were applied:
1. The Cannabis Use Disorders Identification Test (CUDIT), which is a 10-item questionnaire [
2. The Severity of Dependence Scale (SDS), which is a five-item questionnaire that measures the severity of cannabis dependence. Each of the five items is scored on a 4-point scale (0-3). The total score is obtained by adding the ratings on all five items. High scores indicate high levels of dependency [
3. The Cannabis Withdrawal Scale (CWS) [
4. The Cannabis Craving Symptoms questionnaire (CCS-7), which is a seven-item questionnaire [
5. The Fragebogen Substanzanamnese (FDA), which is a questionnaire that ascertains the number of years of consumption over the lifetime, the past month’s consumption, and the manner of consumption for the Diagnostic and Statistical Manual of Mental Disorders’ substances of abuse. This measure was derived from the Europe Addiction Severity Index [
6. The short version of the Mental Health Inventory (MHI-5) [
None of the secondary outcome instruments has yet been specifically validated for Internet use. Intervention satisfaction for all modules, the diary, the chat, the knowledge base, the instant help, and the overall satisfaction was ascertained on a 4-point scale, ranging from
Study measurements and instruments.
Assessments/instruments | Baseline | 1 week | 3 weeks | 6 weeks | 3-month follow-up |
Sociodemographics | x |
|
|
|
|
MHI-5a | x |
|
|
|
x |
Quantity of cannabis useb | x | x | x | x | x |
Frequency of cannabis useb | x | x | x | x | x |
CUDITc | x |
|
|
|
x |
SDSd | x |
|
|
x | x |
FDAe | x |
|
|
x | x |
aMental Health Inventory (MHI-5).
b7-day point prevalence values of the quantity (in common standard joints) and frequency (the number of days on which cannabis is used) of cannabis use were derived from the consumption diary for the preceding 7 days.
cCannabis Use Disorders Identification Test (CUDIT).
dSeverity of Dependence Scale (SDS).
eFragebogen Substanzanamnese (FDA).
Based on results of the study of Rooke et al [
Once participants had completed their baseline assessment, they were randomized by a computer program in a 1:1:1 ratio to one of three parallel groups. As the participant information offered full transparency on the three study arms in our nonblinded design, we anticipated a risk that some participants might register another account, in an effort to change their assignment and access a different study arm. In that case, the participant remained in the initially assigned study arm for the rest of the day, as based on his or her IP address.
Data were analyzed according to the intention-to-treat principle. For the ITT analyses, in departure from the study protocol, we applied multiple imputation procedures of R (R Foundation for Statistical Computing, Vienna, Austria) in Amelia II that have been demonstrated to outperform other imputation methods [
Three months after the baseline assessment, participants were invited by email to log in and complete the final study assessment; they were reimbursed with €40 (via an online voucher or an online charitable donation). The follow-up assessment was performed in three steps. First, participants were invited via email to participate in the assessment. Up to three reminders were sent. Those participants who failed to complete the 3-month follow-up despite these reminders were contacted via telephone and offered an interview by study collaborators. Those participants who refused a telephone interview were offered an interview on the primary outcome only. Finally, 117 out of 308 participants (38.0%) could be followed up with.
CONSORT-EHEALTH trial flowchart: overview of the participant flow for this trial.
Baseline characteristics of participants.
Characteristics | Study arm 1a (n=114) | Study arm 2b (n=101) | Study arm 3c (n=93) | Total (n=308) | χ2, ANOVAd, or Kruskal-Wallis test |
|
|
|
|
|
|
|
χ2 2=4.3 (n=308) | .12 | |
|
Female | 35 (30.7) | 24 (23.8) | 17 (18) | 76 (24.7) |
|
|
|
Male | 79 (69.3) | 77 (76.2) | 76 (82) | 232 (75.3) |
|
|
Age in years, mean (SD) | 28.4 (9.6) | 30.2 (9.2) | 31.0 (11.1) | 29.8 (10.0) |
|
.15 | |
|
|
|
|
|
χ2 2=3.9 (n=308) | .14 | |
|
≤20 years | 24 (21.1) | 12 (11.9) | 18 (19) | 54 (17.5) |
|
|
|
21-25 years | 31 (27.2) | 19 (18.8) | 13 (14) | 63 (20.5) |
|
|
|
26-30 years | 16 (14.0) | 29 (28.7) | 19 (20) | 64 (20.8) |
|
|
|
31-35 years | 17 (14.9) | 18 (17.8) | 15 (16) | 50 (16.2) |
|
|
|
36-40 years | 14 (12.3) | 10 (9.9) | 11 (12) | 35 (11.4) |
|
|
|
41-45 years | 6 (5.3) | 5 (5.0) | 7 (8) | 18 (5.8) |
|
|
|
46+ years | 6 (5.3) | 8 (7.9) | 10 (11) | 24 (7.8) |
|
|
|
|
|
|
|
χ2 10=8.6 (n=308) | .57 | |
|
Not specified | 4 (3.5) | 3 (3.0) | 5 (5) | 12 (3.9) |
|
|
|
Primary school | 18 (15.8) | 12 (11.9) | 11 (12) | 41 (13.3) |
|
|
|
Apprenticeship | 43 (37.7) | 38 (37.6) | 41 (44) | 122 (39.6) |
|
|
|
Secondary school | 19 (16.7) | 13 (12.9) | 17 (18) | 49 (15.9) |
|
|
|
Technical college | 18 (15.8) | 26 (25.7) | 13 (14) | 57 (18.5) |
|
|
|
University | 12 (10.5) | 9 (8.9) | 6 (7) | 27 (8.8) |
|
|
|
|
|
|
|
χ2 6=8.1 (n=308) | .23 | |
|
Canton of Zurich | 52 (45.6) | 33 (32.7) | 42 (45) | 127 (41.2) |
|
|
|
Other cantons | 53 (46.5) | 61 (60.4) | 48 (52) | 162 (52.6) |
|
|
|
Germany | 8 (7.0) | 5 (5.0) | 3 (3) | 16 (5.2) |
|
|
|
Other countries | 1 (0.9) | 2 (2.0) | 0 (0) | 3 (1.0) |
|
|
CUDITe, mean (SD) | 19.8 (5.8) | 19.7 (6.4) | 19.1 (6.2) | 19.6 (6.1) |
|
.69 | |
SDSf, mean (SD) | 7.7 (3.5) | 7.5 (3.6) | 7.3 (3.2) | 7.5 (3.4) |
|
.69 | |
MHI-5g, mean (SD) | 54.0 (19.3) | 53.9 (20.0) | 55.1 (22.6) | 54.3 (20.5) |
|
.90 | |
|
|
|
|
|
|
|
|
|
Cannabinoids | 9.6 (7.4) | 10.9 (7.6) | 12.6 (10.0) | 10.9 (8.4) |
|
.04h |
|
Risky alcohol usei | 2.5 (5.6) | 2.6 (5.3) | 2.7 (6.4) | 2.6 (5.7) |
|
.97 |
|
Cocaine | 1.1 (4.3) | 1.4 (3.8) | 0.8 (1.8) | 1.1 (3.4) |
|
.52 |
|
Amphetamines | 0.7 (2.0) | 1.1 (3.1) | 0.6 (2.0) | 0.8 (2.4) |
|
.40 |
|
|
|
|
|
|
|
|
|
Cannabinoids | 112 (98.2) | 100 (99.0) | 93 (100) | 305 (99.0) | Not computable (no variance) | N/Aj |
|
Risky alcohol usei | 40 (35.1) | 26 (25.7) | 31 (33) | 97 (31.5) | χ2 2=2.6 (n=226) | .28 |
|
Tranquilizers | 7 (6.1) | 8 (7.9) | 5 (5) | 20 (6.5) | χ2 2=0.6 (n=215) | .74 |
|
Cocaine | 7 (6.1) | 14 (13.9) | 10 (11) | 31 (10.1) | χ2 2=2.7 (n=223) | .26 |
|
Amphetamines | 16 (14.0) | 13 (12.9) | 14 (15) | 43 (14.0) | χ2 2=0.2 (n=221) | .90 |
|
Hallucinogens | 6 (5.3) | 4 (4.0) | 4 (4) | 14 (4.5) | χ2 2=0.6 (n=210) | .76 |
|
Heroin | 0 (0) | 1 (1.0) | 0 (0) | 1 (0.3) | χ2 2=1.8 (n=201) | .40 |
|
Methadone | 1 (0.9) | 1 (1.0) | 3 (3) | 5 (1.6) | χ2 2=1.9 (n=197) | .40 |
|
Others | 4 (3.5) | 1 (1.0) | 1 (1) | 6 (1.9) | χ2 2=3.1 (n=198) | .21 |
aSelf-help with chat.
bSelf-help without chat.
cWaiting list.
dAnalysis of variance (ANOVA).
eCannabis Use Disorders Identification Test (CUDIT) scores range from 0 to 40 with a cutoff of >8 for a cannabis use disorder.
fSeverity of Dependence Scale (SDS) scores range from 0 to 15 with a cutoff of ≥4 for cannabis dependence.
gMental Health Inventory (MHI-5): higher values represent improved symptoms. MHI-5 values range from 0 to 100 with a cutoff of <70 for clinically relevant symptoms.
h
iRisky alcohol use was defined as five or more standard drinks per day on at least three days per week. A standard drink was defined as 5 cl spirits, 15-20 cl wine, or 33-45 cl beer.
jNot applicable (N/A).
Module completion rate for study arms 1 (self-help with chat) and 2 (self-help without chat).
Study retention based on the weekly completion of the consumption diary for study arms 1 (self-help with chat) and 2 (self-help without chat) between baseline and week 6, including 3-month follow-up completion rate.
The differences in cannabis use between baseline and the 3-month follow-up, as expressed by the mean number of cannabis use days per week and based on the imputed data, differed between self-help without chat versus self-help with chat (beta= -0.75, SE = 0.32,
Cannabis use days per week according to the consumption diary between baseline and 3-month follow-up for all three study arms based on the nonimputed dataset.
Weekly quantity of cannabis used in number of standardized cannabis joints between baseline and 3-month follow-up for all three study arms based on the nonimputed dataset.
There were no significant differences in the group comparisons in the secondary outcomes (see
Number of participants and mean and standard deviation changes from the imputed (50 imputations) and complete case datasets between baseline and 3-month follow-up.
Outcomes | Study arm 1 |
Study arm 2 |
Study arm 3 |
||||
|
|
Baseline | Follow-up | Baseline | Follow-up | Baseline | Follow-up |
|
|
|
|
|
|
|
|
|
Imputed data | 6.0 (1.6) | 4.6 (2.1) | 6.0 (1.6) | 5.3 (1.8) | 6.3 (1.0) | 5.3 (1.8) |
|
Complete cases | 6.1 (1.6) | 3.8 (3.0) | 6.1 (1.7) | 5.5 (2.3) | 6.7 (0.9) | 5.3 (2.5) |
|
|
|
|
|
|
|
|
|
Imputed data | 22.3 (14.8) | 13.3 (12.0) | 23.1 (23.1) | 14.4 (11.8) | 25.8 (18.7) | 18.6 (17.7) |
|
Complete cases | 23.0 (15.1) | 10.9 (13.8) | 25.1 (25.2) | 14.2 (13.3) | 23.6 (13.2) | 20.7 (23.7) |
|
|
|
|
|
|
|
|
|
Imputed data | 19.8 (5.8) | 16.6 (7.1) | 19.7 (6.4) | 15.6 (6.7) | 19.1 (6.2) | 16.6 (6.4) |
|
Complete cases | 19.8 (5.8) | 12.6 (8.4) | 19.7 (6.4) | 13.0 (7.4) | 19.1 (6.2) | 16.0 (7.2) |
|
|
|
|
|
|
|
|
|
Imputed data | 7.7 (3.5) | 6.3 (3.3) | 7.5 (3.6) | 6.2 (3.1) | 7.3 (3.1) | 6.3 (3.3) |
|
Complete cases | 7.7 (3.5) | 5.3 (3.8) | 7.5 (3.6) | 6.0 (3.3) | 7.3 (3.1) | 5.9 (3.8) |
|
|
|
|
|
|
|
|
|
Imputed data | 53.9 (19.3) | 58.1 (18.2) | 53.9 (20.0) | 60.4 (19.1) | 55.1 (22.6) | 59.4 (19.4) |
|
Complete cases | 53.9 (19.3) | 62.4 (19.8) | 53.9 (20.0) | 63.4 (20.4) | 55.1 (22.6) | 64.6 (18.3) |
|
|
|
|
|
|
|
|
|
Imputed data | 3.4 (6.2) | 2.8 (2.8) | 2.4 (5.0) | 2.2 (3.0) | 4.5 (7.9) | 3.3 (4.0) |
|
Complete cases | 3.4 (7.0) | 1.6 (2.6) | 2.5 (5.8) | 1.0 (2.6) | 4.5 (8.7) | 2.1 (4.7) |
aBased on the weekly number of cannabis use days according to the consumption diary.
bBased on the weekly number of standard cannabis joints according to the consumption diary.
cCannabis Use Disorders Identification Test (CUDIT) scores range from 0 to 40 with a cutoff of >8 for a cannabis use disorder.
dSeverity of Dependence Scale (SDS) scores range from 0 to 15 with a cutoff of ≥4 for cannabis dependence.
eMental Health Inventory (MHI-5): higher values represent improved symptoms. MHI-5 values range from 0 to 100 with a cutoff of <70 for clinically relevant symptoms
Results for the between-study arma comparisons from the linear (and logistic) regression models and calculated effect sizes based on the imputed dataset (50 imputations).
Characteristics | beta | SE |
|
|
Cohen's d (95% CI) | |
|
|
|
|
|
|
|
|
(Intercept) | -3.95 | 0.58 | -6.76 | <.001 |
|
|
Arm 1 vs arm 3 | 0.70 | 0.32 | 2.16 |
|
|
|
Arm 2 vs arm 3 | -0.05 | 0.33 | -0.16 | .87 | -0.14 (-0.43 to 0.14) |
|
(Intercept) | -3.25 | 0.56 | -5.79 | <.001 |
|
|
Arm 2 vs arm 1 | -0.75 | 0.32 | -2.39 |
|
|
|
|
|
|
|
|
|
|
(Intercept) | -14.50 | 2.24 | -6.46 | <.001 |
|
|
Arm 1 vs arm 3 | 4.73 | 2.50 | 1.89 |
|
0.09 (-0.19 to 0.36) |
|
Arm 2 vs arm 3 | 3.77 | 2.42 | 1.56 | .12 | 0.06 (-0.22 to 0.35) |
|
(Intercept) | -9.78 | 1.92 | -5.09 | <.001 |
|
|
Arm 2 vs arm 1 | -0.96 | 2.43 | -0.39 | .69 | 0.01 (-0.26 to 0.28) |
|
|
|
|
|
|
|
|
(Intercept) | -10.39 | 1.61 | -6.46 | <.001 |
|
|
Arm 1 vs arm 3 | 0.24 | 1.29 | 0.19 | .85 | 0.09 (-0.18 to 0.37) |
|
Arm 2 vs arm 3 | 1.19 | 1.20 | 0.99 | .32 | 0.21 (-0.07 to 0.49) |
|
(Intercept) | -10.14 | 1.68 | -6.05 | <.001 |
|
|
Arm 2 vs arm 1 | 0.95 | 1.16 | 0.82 | .41 | -0.12 (-0.39 to 0.14) |
|
|
|
|
|
|
|
|
(Intercept) | -4.68 | 0.64 | -7.34 | <.001 |
|
|
Arm 1 vs arm 3 | 0.03 | 0.58 | 0.05 | .96 | 0.08 (-0.19 to 0.36) |
|
Arm 2 vs arm 3 | 0.10 | 0.56 | 0.17 | .86 | 0.07 (-0.21 to 0.35) |
|
(Intercept) | -4.65 | 0.65 | -7.19 | <.001 |
|
|
Arm 2 vs arm 1 | 0.07 | 0.55 | 0.13 | .90 | 0.02 (-0.25 to 0.28) |
|
|
|
|
|
|
|
|
(Intercept) | -43.91 | 4.33 | -10.15 | <.001 |
|
|
Arm 1 vs arm 3 | 0.96 | 3.44 | 0.28 | .78 | 0.01 (-0.27 to 0.28) |
|
Arm 2 vs arm 3 | -1.38 | 3.42 | -0.40 | .69 | -0.09 (-0.38 to 0.19) |
|
(Intercept) | -42.95 | 4.14 | -10.37 | <.001 |
|
|
Arm 2 vs arm 1 | -2.34 | 3.28 | -0.71 | .48 | 0.11 (-0.16 to 0.38) |
|
|
|
|
|
|
|
|
(Intercept) | -2.84 | 0.56 | -5.09 | <.001 |
|
|
Arm 1 vs arm 3 | 0.32 | 0.63 | 0.52 | .61 | -0.10 (-0.38 to 0.17) |
|
Arm 2 vs arm 3 | 0.83 | 0.73 | 1.14 | .25 | -0.16 (-0.44 to 0.12) |
|
(Intercept) | -2.52 | 0.46 | -5.46 | <.001 |
|
|
Arm 2 vs arm 1 | 0.51 | 0.61 | 0.84 | .40 | 0.06 (-0.20 to 0.33) |
aStudy arm 1: self-help with chat; study arm 2: self-help without chat; study arm 3: waiting list.
bBased on the weekly number of cannabis use days according to the consumption diary.
cSignificant and borderline significant differences and effect sizes are in italics.
dBased on the weekly number of standard cannabis joints according to the consumption diary.
eCannabis Use Disorders Identification Test (CUDIT) scores range from 0 to 40 with a cutoff of >8 for a cannabis use disorder.
fSeverity of Dependence Scale (SDS) scores range from 0 to 15 with a cutoff of ≥4 for cannabis dependence.
gMental Health Inventory (MHI-5): higher values represent improved symptoms. MHI-5 scores range from 0 to 100 with a cutoff of <70 for clinically relevant symptoms.
Dropouts at follow-up did not differ from completers with respect to the following baseline variables: gender (
Significantly more participants could be followed up who received at least one chat session compared to those who could not be contacted at the 3-month follow-up (17.0% vs 5.5%, chi-square [df 2]= 7.5,
Dropouts did not differ between the three study arms with respect to gender (
Although not intended as an outcome measure, we also offered cannabis abstention in the study protocol for those participants who wished to achieve this [
Number of participants in three study arms at each time point.
Study time point | Study arm 1 |
Study arm 2 |
Study arm 3 |
Week 1 | 9 (7.9) | 12 (11.9) | N/Aa |
Week 6 | 8 (7.0) | 9 (8.9) | N/A |
Follow-up | 9 (8.8) | 2 (2.0) | 4 (4) |
aNot applicable (N/A).
Self-reported abstinence between groups and with the corresponding logistic regression.
Abstinence at follow-up | beta | SE |
|
|
ORa (95% CI) |
(Intercept) | 0.04 | 0.02 | 1.88 | .06 |
|
Arm 1 vs arm 3 | 0.76 | 0.61 | 1.25 | .21 | 2.14 (0.86-5.30) |
Arm 2 vs arm 3 | -0.80 | 0.88 | -0.91 | .36 | 0.45 (0.11-1.78) |
(Intercept) | -2.34 | 0.33 | -7.07 | <.001 |
|
Arm 2 vs arm 1 | -1.56 | 0.79 | -1.98 |
|
0.21(0.02-2.33) |
aOdds ratio (OR).
bBorderline significant difference is shown in italics.
Participants in study arm 1 who received at least one chat session exhibited lower changes in their entries in the consumption diary. This meant that they took longer to complete the consumption diary and exhibited higher retention (change in mean 0.3 vs 0.5; beta = -0.28, SE = 0.12,
Participants in study arm 1 who did not receive a chat session for whatever reason did reduce their frequency of cannabis use more (change in mean 1.9) than participants in study arm 2 (change in mean 0.7) who did not have the possibility for a chat session due to their allocation (beta = -1.97, SE = 0.60,
At the 3-month follow-up, 88.0% of participants (103/117) stated that they had not contacted any other treatment services (7 participants in study arm 1, 2 in study arm 2, and 5 in study arm 3). A total of 5.1% (6/117) had contacted a psychiatrist, 2.6% (3/117) a family doctor, 1.7% (2/117) a psychologist, 1.7% (2/117) a different Internet counseling service, and 1 person (0.9%) a drug counselor. During the whole study period, 5 out of 308 (1.6%) participants contacted one of the outpatient addiction clinics from the ARUD Centers for Addiction Medicine. None of them had to be treated as an emergency case or had to be referred to an inpatient treatment service. Moreover, none of the involved counselors or researchers are aware of any adverse or serious adverse event related to the Can Reduce study that was reported by other addiction counseling services.
The Can Reduce study could reach a different group of cannabis users who do not enter outpatient addiction treatment services. They are older and consume much more cannabis than outpatient service users. The finding that we reached cannabis users with more entrenched problems (eg, daily users) is not consistent with the common perception that those using online interventions have less severe problems than those entering outpatient services. We assume that this finding was most probably due to an age effect. Older users consume longer and possibly also more than younger ones but might feel more stigmatized if they enter an outpatient addiction service, due to their greater responsibilities and roles in social relationships, at work, and in society in general.
Can Reduce participants allocated to the self-help with chat study arm reduced their frequency of cannabis use more than those in the other two arms. Even cannabis abstinence was higher among those who received additional chat counseling relative to those who received self-help only at follow-up. There was a trend (
As only one-quarter received at least one chat session, the question arose as to what was actually responsible for the superiority of the self-help with chat study arm. The subgroup analyses showed that those participants in study arm 1 who did not receive a chat session reduced their frequency of cannabis use more than those who received self-help only from the beginning (study arm 2). Thus, even an invitation to a chat session and the knowledge that there is a possibility to have a chat appointment might have improved this main outcome for cannabis use. To the best of our knowledge, there are no similar studies in the literature that have reported a comparable effect. However, our result is in line with the first point of the Supportive Accountability model [
However, those participants who actually received at least one chat counseling session in study arm 1 still performed better in their reduction of cannabis use and completed more self-help modules than their counterparts who did not receive a chat session in the same study arm. This result is in line with a further point of the Supportive Accountability model [
If we compare the current results with former studies about the reduction of cannabis use with similar therapeutic approaches, it stands out that participants in the Can Reduce self-help without chat study arm performed worse than those in the Australian
We observed a borderline significant effect in the abstention rates between the self-help with chat and the self-help without chat study arms. As we did not initially expect that enough participants would maintain their abstinence, we omitted abstinence as an outcome measure in the study protocol [
Setting a goal for cannabis consumption was implemented as described in the study protocol [
The strengths of the Can Reduce study are that the intervention is theory based and pretested, that this Web-based intervention was able to reach cannabis users who otherwise would not have sought help, and that we were able to disentangle the effects of chat counseling additional to self-help for the reduction in cannabis use in frequent cannabis users, three-quarters of whom used cannabis daily. This study also possesses limitations that merit consideration. First, we did not biologically validate cannabis consumption for financial reasons, as we did not want to limit participation to participants who were willing to provide, for example, saliva samples, and as we did not want to limit external validity. Second, we did not succeed in attaining a better 6-week follow-up as intended in the study protocol, which limits the explanatory power of the short-term effects of Can Reduce. However, the 3-month follow-up rate (117/308, 38.0%) was comparable to similar studies with problematic cannabis users in Europe [
In conclusion, the Can Reduce study demonstrated that Web-based interventions possess the potential to reach heavy cannabis users who differ from those who enter outpatient addiction treatment services. We further conclude that offering brief chat counseling in addition to Web-based self-help can significantly increase success in the reduction of cannabis use in the different groups of cannabis users investigated
Can Reduce participant information and informed consent page.
CONSORT-EHEALTH checklist V1.6.1 [
addiction, care and therapy information
analysis of variance
Arud Centers for Addiction Medicine
behavioral self-management
cognitive behavioral therapy
Cannabis Craving Symptoms questionnaire
Cannabis Use Disorders Identification Test
Cannabis Withdrawal Scale
Fragebogen Substanzanamnese (questionnaire for the assessment of substance use history)
Swiss Research Institute for Public Health and Addiction
intention-to-treat
Mental Health Inventory
motivational interviewing
not applicable
odds ratio
randomized controlled trial
Severity of Dependence Scale
waiting list
Funding for this study was provided by Infodrog, the Swiss Office for the Coordination of Addiction Facilities, Switzerland (Grant No. 5012/13/ZH/Cannabis Control). The funding institution had no role in the development or evaluation of the interventions. The authors wish to extend particular appreciation to the psychology master's students Emilija Nikolic and Manja Djordjevic for helping to conduct the telephone follow-up interviews.
None declared.
MPS was responsible for the study design and the final manuscript. AW and MPS performed the analyses and prepared the first draft of the paper. All authors developed the intervention of study arms 1 and 2 and SH supervised the analyses. AW programmed and implemented the study website, Can Reduce. All of the authors approved the final version of the manuscript submitted for publication.