We adhered to the Standard Protocol Items Recommendations for Interventional Trials (SPIRIT) guidelines in the preparation of this protocol (see Fig. 1 and Additional file 1 for the SPIRIT Figure and Checklist, respectively).
Setting and participants
Project coordination will take place at “Riddargatan 1: Clinic for Alcohol and Health”, an outpatient alcohol clinic located in Stockholm, specializing in treating AUDs and part of the Stockholm Center for Dependency Disorders. We collaborate with a number of PC clinics within Stockholm County in the study. Participants are informed about the study through the leaflet “Feel better with less alcohol”, distributed at the registration desk of the PC study sites to all patients > 18 of age under a time period. Patients are also recruited via advertisements placed in waiting rooms at the participating PC clinics and via advertisements on the internet and in local newspapers. Participants must be listed on and have their GP in one of the collaborating PC clinics.
Inclusion criteria:
Participants have to meet criteria for alcohol dependence, based on the International Statistical Classification of Diseases and Related Health Problems - Tenth Revision (ICD-10). Participants also have to meet the threshold for hazardous drinking, which is defined as 6 or more points for women and 8 or more points for men on the Alcohol Use Disorders Identification Test (AUDIT) (20).
Exclusion criteria:
Individuals with serious mental illness, including substance use disorder other than alcohol or nicotine, persons deemed in need of specialist treatment in psychiatry or dependence care, cognitive impairment and lack of knowledge of the Swedish language will be excluded.
The web-based treatment platform
We have built a web-based treatment platform with questionnaires for initial assessment and follow up and a series of treatment modules based on cognitive behavioral therapy (CBT) and motivational enhancement therapy, here called internet based cognitive behavioral therapy (iCBT). Prior to starting the RCT, a pilot study (n=3) was conducted involving an intervention identical to that used in the present study. The purpose was to test the study protocol, the internet assessment procedure and the web-based treatment platform to obtain feedback from the patients on the feasibility.
Study design, randomization and blinding
This is a multi-center, two-group, parallel, RCT to compare the effects of two 12-week interventions (internet cognitive behavioral therapy (iCBT) plus treatment as usual (TAU) or TAU only designed to demonstrate statistical superiority. The proposed flow of participants through the trial is shown in Fig. 2. Assessments and biomarkers are completed by all participants at baseline (pre randomization) and post treatment at week 13 and week 52. The randomization was done in blocks of 20 according to a fully automated and concealed procedure in the online platform that was programmed in the content management system Drupal (drupal.org) by the fourth author. Type of treatment (iCBT+TAU vs TAU) is not blinded for neither the participants nor the research nurse. Hence unblinding of participants and the research nurse will not be necessary during the trial. Responsible researchers as well as data analysts will not have access to the codes concerning which experimental condition the participants participate until after completion of analysis of the primary outcome measure. Hence, the risk of bias in the interpretation of the study´s results is minimal.
Procedure
Recruitment
The personnel at the PC clinics are asked to encourage all patients attending the PC clinic under the time period to logon to the website indicated on the leaflet. Individuals interested in participating click on a link to the study registration site. Here the participants are provided with information about the study, they sign an informed consent, leave their data on demographics and contact information. Participants provide an e-mail address for communication.
Screening and baseline assessment
Initially a web-based screening assessment takes place. Individuals that fulfill the inclusion criteria (alcohol dependence and hazardous consumption) are automatically invited to complete baseline assessment. Individuals that do not fulfill the inclusion criteria are excluded and guided where to get further help.
After completion of self-assessment on the website an extended telephone screening with a research nurse takes place to assure data quality and completeness. In the telephone conversation the research nurse goes through the exclusion criteria. Individuals that meet exclusion criteria are advised to seek appropriate action.
Randomization and treatment allocation
The research nurse randomizes all eligible participants via the online platform by activating a link that only she has access to. The participants are informed by the research nurse about their group allocation and are scheduled for a blood test for biomarkers and an appointment with the GP at the PC clinic, ideally within 2-3 weeks. The GPs give all participants feedback on the assessment and biomarkers and design a treatment plan based on current routines at the PC clinic. For those randomized to iCBT, the internet program will be made available the following Monday after randomization. If considerable somatic or psychiatric complications occur during the study the research personnel or the GPs can take the initiative to withdraw the patient from the study and if applicable make a referral to an appropriate care giver. Following completion of the study no post-trial care is planned as part of the study, but since all patients are in contact with their GPs further treatment may be provided within the framework of TAU.
Outcome measures
Primary outcome is change in alcohol consumption in mean grams per week and mean number of days with heavy drinking (that is >4/5 drinks for women and men) the last 30 days from baseline to12 and 52 weeks post inclusion. A drink contains 12 grams of alcohol.
For all secondary outcomes change from baseline to 12 and 52 weeks post inclusion applies; difference in mean AUDIT score, difference in mean number of ICD-criteria for alcohol dependence, difference in mean scores in EQ5D, difference in mean HADS-scores and difference in biomarker levels including liver enzymes and PEth.
Baseline and follow-up assessments
Questionnaires: Time Line Follow Back 30 days (20) is used to assess mean weekly alcohol consumption and Heavy drinking days. Severity of hazardous drinking and dependence is assessed with Alcohol Use Disorders Identification Test (AUDIT) (21) and the ICD-10 criteria for alcohol dependence (The World Health Organization, 1992). Symptoms of anxiety and depression are assessed with Hospital Anxiety and Depression Scale (HADS) (22). The EQ 5D-5L questionnaire measure health related quality of life (23). Satisfaction with treatment is assessed, only at 12 weeks follow up, with Client Satisfaction Questionnaire (CSQ) (24).
Biomarkers: Blood is analyzed for standard biomarkers of heavy drinking and liver pathology at each assessment; levels of phosphatidylethanol (PEth); gamma-glutamyl transferase (GGT); aspartate amino transferase (AST) and alanine amino transferase (ALT).
Biological specimens will not be stored as a part of the study. Results from blood samples are provided via the patient file system at the PC clinics and blood samples are destroyed after analysis.
Treatment as usual (TAU)
The GPs design a treatment plan based on current routines for treating alcohol problems at the PC clinic. The fact that all participants are offered contact with a GP enables participants in need of specialized care of any kind to be referred by the GP. All GPs are offered a brief training in giving feedback on the assessment and biomarkers and the use of medication for treatment of alcohol dependence. Together this serves as the base of TAU which is provided to the participants in both study arms.
Internet Cognitive Behavioral Therapy (iCBT)
The intervention group will be given iCBT in addition to TAU. The iCBT program is based on the self-help material used in previous studies on the internet (18, 25) and in primary care (13). iCBT includes five elements: 1. Motivation to cut down or stop drinking. 2. Goal-setting for alcohol consumption per occasion and weekly consumption and self-control strategies, for example not to have alcohol at home and strategies to drink more slowly. 3. Chart risk situations and conduct behavior analysis on occasions with too much alcohol. 4. Plan alternative activities to drinking alcohol. 5. Relapse prevention and crisis plan and mapping of life situations or stressors that can lead to relapse. The intervention is open-ended, meaning that participants can log on as often and for as long as they want.
For each assignment an informational text is provided and a home assignment is included. iCBT is an extended self-help intervention with automated feedback where participants repeatedly are reminded to start and complete every assignment.
Follow-up assessments
Patients will be contacted by automatically created messages in the web-based treatment platform and by the research nurse when necessary during week 13 and 52 in order to conduct the follow-up assessments and blood tests. At week 13 the Client Satisfaction Questionnaire (CSQ) is added. Assessments completeness are checked automatically in the treatment platform and if not completed participants can´t proceed to the next step. Participants will be contacted by the research nurse and reminded if assessments are not completed using a combination of messages from the treatment platform, phone calls, emails, and text messaging in order to improve retention in treatment and prevent loss to follow up over the follow-up period. The research nurse will also remind the participants to leave the blood tests during week 13 and 52 and collect the results from the patient files.
Program engagement
Program engagement will be measured by automatic data collection on number of assignments accessed, tasks completed. Association between program engagement metrics and treatment outcomes will be explored.
Statistical analyses and sample size calculations
Baseline participant characteristics stratified by treatment group will be presented. Group differences will be tested using t-test for continuous variables and chi-square tests for categorical outcomes. Multiple imputation will be used to replace missing internal values where appropriate. Within-group changes in the primary outcome (alcohol consumption) and secondary outcomes (ICD-10 criteria, AUDIT-scores, depression, anxiety, EQ5D and biomarkers) at 12 and 52 weeks will be analyzed according to intention-to-treat principle, using linear mixed models and Poisson regression where appropriate. The models will further be adjusted for baseline covariates and treatment adherence. Per-protocol analyses will also be presented. All tests will be two-sided and p-values of less than 0.05 considered statistically significant.
The study is designed to demonstrate statistical superiority. An effect-size of 0,61 has been shown in previous Internet studies on high consumers in the general population (26). Due to a lack of studies on alcohol dependent patients, our estimate is based on a more conservative effect-size of 0,4, which requires the estimated sample size of 100 participants in each arm to get a power of 80% and a statistically significance 0,05 when using a t-test. Given an estimated dropout in iCBT of 30%, 260 participants are included in the study (27).
No interim analyses are planned.
Patient and public involvement
The need for the investigation of the efficacy of the iCBT-intervention was informed by clinical experience and by previous research conducted within our research group. We collected information from GPs on how to perform the iCBT-intervention with as little extra workload for the GPs as possible, when planning the study. We also conduct semi-structured interviews with a number of the GPs involved in this study on factors that can facilitate and hamper implementation of iCBT in primary care, results which will be published in a separate paper. Patients were not involved in the design or execution of the trial. Participants will have the possibility to access their individual study-data through the intervention platform. Results from the study will be disseminated through journal articles, conference presentations and reports to the funding agencies. Burden of the intervention was not assessed by patients prior to commencing the trial; patient´s satisfaction with the treatment in the study will however be collected as part of the postintervention assessment.