Supportive Text Messages to Reduce Mood Symptoms and Problem Drinking- Randomised Controlled Pilot Trials

Depression and Alcohol Use Disorders (AUDs) are leading causes of disability worldwide associated with significant treatment challenges especially when they occur concurrently (1). Concurrent challenges include: Higher rates of lifetime drug dependence (2); worse outcomes among those entering treatment for alcohol and drug misuse (3); higher relapse following AUD treatment among adolescents (4) and adults (5); greater severity of suicidality in adults (6) higher likelihood of suicide attempts (7) and completed suicides (8). Because of this we seek to develop a complimentary enhanced, efficient, innovative and cost effective treatment strategy aimed at reducing the burden these disorders impose on service users, care givers, their communities and Alberta Health Services. Using mobile phone technology (9). In a pilot trial In Ireland, participants with concurrent depression and AUD who completed an in-patient dual diagnosis treatment program were randomised to receive twice daily supportive text messages or a two weekly thank you text message for three months. That trial showed that supportive text messaging improved symptoms of depression (Beck Depression Inventory scores) compared with those who received only standard care. There was also a trend towards increased Cumulative Abstinence Duration (CAD) in the supportive text message group (10). The pilot trial in Ireland did not explore the effects of supportive text messages on the individual disorders when they occur alone, although it would be reasonable to assume that they would benefit equally. A recent study in JAMA Psychiatry (11) (randomising patients leaving residential treatment for alcohol use disorders to receive either a smart-phone application to support recovery in addition to treatment as usual or only treatment as usual) reported that for the 8 months of the intervention and 4 months of follow-up, patients in the smart phone application support group reported significantly fewer risky drinking days than did patients in the control group. In a review of text messaging for clinical and healthy behaviour interventions, Wei et al. (2011) found that among sixteen randomized controlled trials, ten reported significant improvement with interventions and six reported differences suggesting positive trends (12).

Overall Aim: To conduct a randomised controlled pilot trial to determine the effectiveness of supportive text messages in reducing mood symptoms and problem drinking in patients with Depression and/or AUD respectively.

Specific objectives:

1. To compare mean changes in Beck's Depression Inventory (BDI) from baseline for patients with Depression receiving supportive text messages with those not receiving any supportive text messages.
2. To compare the Cumulative Abstinence Duration (CAD) from baseline for patients with AUD receiving supportive text messages with those not receiving any supportive text messages..

Hypotheses:

Supportive text messages will:

1. Reduce the change BDI scores from baseline in patients with Depression by at least a mean of 30% for patients in the intervention group compared to those in the control group.
2. Increase the CAD in patients with AUD by at least 30%.

Study design and setting: This will be a longitudinal, prospective, parallel design, two arm, placebo-controlled single-rater-blinded randomized clinical trial with a recruitment period of 6 months and an observation period of 3 months for each participant with two strata based on primary diagnosis of Major Depressive Disorder or AUD. Patients with AUD will be recruited from those completing the Northern Addiction Residential Treatment programme located in Grande Prairie while patients with Depression will be recruited from those assessed for wait-listing onto the Cognitive Behaviour Therapy (CBT) program run by the Adult Mental Health Team in Fort McMurray. Ethics approval will be sought and participants will all provide informed written consent prior to randomization.

Recruitment: All patients completing the Northern Addiction Residential Treatment programme located in Grande Prairie and those who are assessed by a psychiatrist in Fort McMurray who fulfil the inclusion criteria below will be invited to participate. This will minimize the source of referral as a confounding factor. Information leaflets will be provided and those consenting will be recruited. The Structured Clinical Interview for the Diagnostic and Statistical Manual of Mental Disorders {DSM-5} (SCID) will be used to screen and confirm the diagnosis of all study participants prior to inclusion in the study.

Randomisation and blinding: Randomisation will be stratified by primary diagnoses (Depression or AUD) using permuted blocks to ensure balance (1:1) between treatment and control groups within each condition under study. The randomisation codes will be transmitted by an independent statistician only to the researcher sending text messages, immediately after the participant has signed the informed consent to a dedicated password protected phone line with a secure online backup for these messages.

The researcher who sends the text messages will not be involved in the outcome assessments except to contact those receiving the supportive text messages for feedback on the benefits of the text messages. Participants cannot be blinded and treatment allocation will be made explicit to them. Outcome assessors will be blinded to treatment group allocation. To ensure this, treatment allocation will be concealed from the outcome assessor by blocking access to the database which contains the randomization code, and all participants in the study will be asked not to reveal their treatment allocation to their assessor. Outside the assessments, outcome assessors will not participate in discussions of participants in study forums. To test the success of blinding we will ask the assessor to guess the allocation group for each participant at 3 months follow-up. Moreover, these assessors will also not be involved in data analysis. After data collection is complete all data will undergo a blind review for the purposes of finalizing the planned analysis. Every effort will be taken to avoid incomplete data and methods of dealing with unanticipated missing data will be decided upon during the blind review of the data.

Intervention: Patients in all the intervention groups will receive twice daily supportive text messages for 3 months. The messages will be tailored towards improving mood, compliance with medication or targeting abstinence from alcohol in accordance with the primary aims of our study. As far as possible, the majority of the text messages will reflect lessons patients usually learn during CBT and addiction counselling sessions. They will be sent at specified times by a centralised computer program which will be set up and monitored by one of the research assistants. Some sample text messages are outlined in Table 2. To safeguard patient confidentiality no identifiable patient information will be transmitted via the text messages. The patients will also receive a phone call every two weeks from the research assistants who does the treatment allocation to confirm that they are still receiving the text messages and also to thank them for taking part in the study. Patients in the control group will receive no text messages but will also receive a phone call from the same researcher every two weeks to thank them for taking part in the study. The aim of this will be to help increase the retention rate for the study.

Sample size: Consistent with the idea that this is an exploratory study, the research will use the data that can be elicited from the participants that can be enrolled within the existing resources, a principle described by Haynes et al. (2005) as using "the patients I can get" (13). The study will therefore be limited to a pilot trial with a sample size of 120, with about 30 patients in each of the two primary diagnosis (Depression or AUD) based intervention groups and another 30 patients in each of the two primary diagnosis based in the nonintervention group.

Follow-up assessments: At 3 months, a blinded researcher (most probably a research psychologist) will contact all patients over the phone and assist them to complete a range of assessment tools relating to the primary and secondary outcome measures. Administrative data related to cost of health services utilisation and days absent from work for each patient will also be sourced and compiled. Those with AUD will also be directed to a dedicated laboratory for blood tests related to the secondary outcome measures.

Statistical methods: Analysis of both primary and secondary outcome variables will be done on an intention-to-treat basis, using generalised linear mixed models (GLMM). Patients who withdraw from the study or are lost to follow-up will have their primary, endpoint and administrative data included in the analysis as per consent provided prior to randomization unless such consent is explicitly withdrawn. Missing data in other outcome variables can be adequately accounted for using GLMM. Analyses will be stratified by diagnosis group, providing that there is no evidence of an interaction between treatment group and diagnosis group (i.e. the effect of treatment differs according to diagnosis group). Subsequent analysis will examine differences between the two treatment groups within each quarterly period, using time as a 4 category variable, since the effect of time is unlikely to be linear, and using an interaction between treatment group and time. Economic decision modelling approach will be used to test the sensitivity and acceptability of the incremental cost and cost-effectiveness results (14, 16).