BA Trial for Co-Occurring Depression and Substance Use

RATIONALE Over 50% of treatment-seeking substance users present with clinically significant depression symptoms (Johnson, Neal, Brems, \& Fisher, 2006). This combination of problems tends to complicate treatment (e.g. Carroll et al., 1993) resulting in greater functional impairment (Johnson et al., 1995) and poor health outcomes (Hasin et al., 2002). Treatment protocols for co-occurring mental health and substance misuse have traditionally focused on treating these issues separately, however there is now increasing consensus that integrated approaches to treatment are more effective (Davidson et al., 2010) and subsequently a growing interest into how targeted mental health support can be incorporated into standard drug and alcohol treatment.

CBT is deemed to be one of the most effective psychological treatments for depression, though may be too complex and time-consuming for effective delivery in routine CDAT. Emerging evidence suggests that a more suitable intervention in this context may be Behavioural Activation (BA); a brief, manualised intervention which targets depression symptoms by increasing engagement in rewarding, values-based activities (Lejuez, Hopko, Acierno, Daughters, \& Pagoto, 2011). Indeed, previous studies have found that BA is associated with reductions in both depression symptoms and substance use (Daughters et al., 2018; Daughters et al., 2008; Delgadillo et al., 2015), as well as improved treatment retention among substance misuse service users with comorbid depression (Magidson et al., 2011). The only known trial of BA in UK Community Drugs and Alcohol Treatment (CDAT) was conducted by Delgadillo and colleagues (2015), however this study was under-powered and employed qualified mental health practitioners to deliver the intervention which may be too costly or unfeasible to implement in routine practice. More research is needed to establish the efficacy of BA in CDAT services, and given that BA may be delivered effectively with minimal training by staff who are not formally trained as psychological therapists (Ekers et al., 2008), it seems that the most suitable and cost-effective way of integrating BA into routine drug and alcohol treatment would be to upskill the existing workforce.

Accordingly, given the high prevalence and impact of comorbid depression among addiction service users, along with evidence to suggest that BA is an effective and acceptable intervention for treating this population, the present study will expand on previous research by conducting a trial to evaluate the efficacy of BA delivered by trained drugs and alcohol workers in a sample of patients accessing treatment for substance use. The central research question is: How effective is Behavioural Activation for treating depression in substance users compared to treatment as usual in Community Drugs and Alcohol services?

AIMS

The proposed study aims to:

1. Evaluate the efficacy of brief one-to-one manualised BA delivered by trained drugs and alcohol workers in a CDAT setting in terms of depression outcome
2. Evaluate the efficacy of delivering manualised BA in this setting in terms of substance use and engagement (attendance and dropout rates) outcomes

HYPOTHESES

Participants who receive the BA intervention will show improvements in:

1. Depression symptoms
2. Substance use
3. Treatment Engagement (sessions attended, % dropout)

CLINICAL IMPLICATIONS If the BA intervention is found to be more effective than the usual care offered in CDAT services, this trial will provide evidence for the intervention to be implemented in routine practice. Even if the BA intervention is not found to be more effective, the results of this study will contribute to understanding of what works and what doesn't work in this setting, which can be used to inform the development of future interventions aimed at addressing co-occurring substance use and depression.

DESIGN This study is a pragmatic, parallel group, randomised controlled trial assessing the efficacy of BA compared to treatment as usual (TAU) for patients with depression who are accessing CDAT services. Outcome measures will be taken at baseline, 3-, 6-, 12- and 24-weeks to enable the comparison of relative efficacy.

METHOD Participants will be recruited from CDAT services in the UK. They will be working age adults (18-65) with mild to moderate depression who are actively using substances. The study aims to recruit 142 participants in total.

PROCEDURE Drugs and alcohol workers in CDAT services will employ a step-wise mental health screening strategy as part of routine practice in order to identify potential participants. This strategy involves routinely completing TOP and brief mental health questionnaires sequentially in order to identify patients with common mental disorders.

Potential participants will be provided with an information sheet and asked for permission to pass their details on to the research team.

Consenting participants will be contacted by a member of the research team within one week to conduct eligibility screening and obtain informed consent. Participants will be aware of confidentiality arrangements for the study, that their participation is voluntary and that they have the right to withdraw at any time.

Once consent to participate has been established, the identified clients' details will be passed on to an independent administrator for randomisation.

Participants will be randomly allocated into one of two arms ('BA Intervention' or 'Treatment as Usual'). Participants randomised to Treatment as Usual will continue to receive their usual care at the CDAT service. Participants randomised to the BA intervention will receive 6 x weekly one-to-one sessions of BA plus 2 booster sessions at 2-3 and 4-5 weeks post-treatment. BA will be delivered by drugs and alcohol workers who have completed 2 days of training in delivering the BA intervention. Staff delivering the intervention will also receive an extra hour of supervision once per month with a supervisor-trained supervisor which will be focused on delivery of the BA intervention.

To ensure fidelity to the BA treatment protocol, drugs and alcohol workers delivering the BA intervention will complete a checklist at the end of each session indicating their adherence to the session protocol, with the checklist for each session highlighting specific session objectives. Additionally, drugs and alcohol workers will be asked to audiotape at least two of their sessions at random over the course of the study. Audiotaped therapy sessions will be rated using the rating scales from Magidson and colleagues' (2011) study by an independent evaluator with an understanding of behavioural activation strategies. Independent evaluators will be Psychological Wellbeing Practitioners with experience of delivering BA, the Study Coordinator, or students under the supervision of the Study Coordinator. Issues of non-adherence will be considered a training implication representative of areas for improvement to be addressed via standard supervisory procedures. Non-adherence will not preclude data inclusion for the purposes of this study; instead it will be addressed as a limitation of the study related to therapist experience.

DATA ANALYSIS Screening, recruitment, random allocation and treatment engagement (including dropout) data will be summarised using a CONSORT diagram. The integrity of randomisation, demographic and clinical characteristics will be assessed by comparing between groups and between recruitment sites using appropriate statistical tests. Loss to follow-up has been estimated at 20% based on a previous feasibility study of BA conducted in this setting (Delgadillo et al., 2015). Intention-to-treat analyses will be conducted using Last Observation Carried Forward method in order to account for the loss to follow-up. Secondary to this, we will analyse the data only with participants who completed all of the follow-up assessments in order to check that findings were not unduly affected by attrition.

We will calculate and report standardised effect sizes for depression outcome (PHQ-9) using Cohen's d. Severity of depression at 12-week follow-up will be considered the main outcome for the calculation of effect sizes; although we will also report 6- and 24-week follow up data for comparison. Severity of depression at 12-week follow-up will be compared between-groups using Analysis of Covariance with participant's pre-treatment scores on the PHQ-9 used as covariates. This is to counter potential baseline variance in depression that may influence results and was considered robust for relatively small sample sizes in a similar study of BA (Ekers et al., 2011).

For both intervention and control groups, we will present the proportions of participants who had reliable and clinically significant change (RCSC) in depression status at 6, 12 and 24 weeks. The RCSC outcome definition (Jacobson \& Truax, 1991) has been recommended as a standard reporting strategy for all published research involving psychological interventions (Evans et al., 1998).

Pre-post changes scores will also be calculated and reported for substance use outcomes (Percent Days Abstinent and Severity of Dependence). Percent days abstinent will be considered the primary substance use outcome. We will calculate and report standardised effect sizes for percent days abstinent at 12-week follow-up using Cohen's d, with 6- and 24-week follow-up data reported for comparison. Percent days abstinent at 12-week follow-up will then be compared between groups using Analysis of Covariance, with participant's pre-treatment scores on the SDS entered as a covariate to account for potential baseline variance in severity of dependence that may influence results. Changes in SDS scores pre-post treatment will be compared between groups using Student's t-test.

For both intervention and control groups, we will present the proportions of participants who had reliable and clinically significant change (RCSC) in percent days abstinent and severity of dependence at 6, 12 and 24 weeks based on guidelines by Jacobson and Truax (1991).

The mean number of key-working sessions offered to participants in the Treatment as Usual arm during the 6-week intervention period will be reported. Percentage of sessions attended will then be compared between groups using Student's t-test.

SAMPLE SIZE CALCULATION Based on previous research conducted in the USA, an effect size of 0.5 was used to calculate the required sample size (Daughters et al., 2018; Daughters et al., 2008). Standard sample size calculation based on a medium effect size (0.5) with 2 independent sample means indicates that 128 participants (64 per group) would be required to achieve 80% power using a significance level (p) of 0.05. After accounting for the possibility that an estimated 10% of participants may drop out of the study (i.e. withdraw consent), this means that we will need to recruit 142 participants in total (71 per group). A previous feasibility study conducted by Delgadillo and colleagues (2015) indicated a screened to recruited ratio of 4:1 in this setting, therefore we will aim to screen a minimum of 568 participants sequentially over a 2-year period in order to achieve an adequate sample size for this study.