Can Cognitive-bias Modification Training During Inpatient Alcohol Detoxification Reduce Relapse Rates Post-discharge?

Overview

It is well-established that many substance misusers experience impairment in cognition (thinking skills), particularly those needed to regulate and monitor behaviour and ensure that goals are achieved. According to the dual-process model, addiction arises from an imbalance in 'bottom-up' processing i.e., overactive automatic (impulsive) processes that drive behaviours and impaired 'top-down' controlling processes that stop behaviours associated with negative consequences. As a result, the individual becomes more sensitive to cues in their environment (e.g., alcohol images) that trigger the addictive behaviour. Cognitive-bias modification (CBM) is a novel, computer-based training paradigm that trains the brain to pay less attention to negative/harmful cues and more attention to positive or neutral cues. This approach minimizes the overactive 'bottom-up' processes and improves the 'top-down' control processes of unhealthy behaviors which enables the addicted individual to make better decisions. Recently, CBM has been used with addicted population to alter the tendency to approach alcohol, with one German study showing that a 4-session training programme was associated higher rates of abstinence at one-year (Wiers et al., 2011). The current study examines whether a novel computer based training programme alters cognitive biases (the tendency to approach alcohol related stimuli) in alcohol-dependent inpatients, and examine whether this enables them to be better at decision-making more generally, and its impact on craving and post-discharge abstinence rates. The study will also explore whether individual differences in impulsivity and sensitivity to reward and punishment determine response to the training programme. This will be achieved using a parallel-groups randomized superiority trial design involving approximately 80 patients attending inpatient withdrawal programmes in Victoria. The findings are likely to have implications for the design and delivery of psychosocial interventions delivered during early recovery from alcohol-dependence to optimise treatment effectiveness.

Detailed Description: According to the dual-process model of addiction (Gullo, Loxton, \& Dawe, 2014), addictive behaviour is the result of an imbalance between a strong, impulsive processing system and a relatively weak, reflective processing system. Due to this imbalance, impulsive preferences are rewarded, reflected in an increased sensitivity to the stimuli of addiction (i.e., attentional bias) and an automatic tendency to engage with the stimuli of addiction (i.e., approach bias; Wiers et al., 2007). Clinical trials of CBM have begun to emerge for individuals with alcohol use disorders (Fadardi \& Cox, 2009). Wiers et al. (2009) developed the alcohol approach/avoidance task (alcohol-AAT), where participants respond with an approach behaviour (pulling a joystick) or an avoidance behaviour (pushing a joystick) to pictures of addiction-related and neutral stimuli. Following four sessions of this training task patients displayed better treatment outcomes one-year later (Wiers et al., 2011). A study by Eberl et al. (2013) found that 12 sessions of approach-bias modification training was associated with higher rates of abstinence at one-year relative to controls. Underlying the relationship between cognitive biases and addiction, within the dual-process model paradigm, are impulse-control processes. However, the relationship between impulsivity and cognitive biases remains unclear. Indeed, individuals with weak inhibition skills tend to have a bias toward automatic information processing (Gladwin et al., 2011) and impulsivity is thought to play a role in the degree to which cognitive biases influence outcomes (Peeters et al., 2012). The development of this understanding will benefit assessment of CBM treatments for addiction patients. Rationale: Despite intensive psychosocial interventions, most alcohol-dependent patients relapse within weeks if not days of leaving inpatient detoxification. The study therefore examines whether an alcohol approach-bias modification training programme during detox can reduce craving and relapse in alcohol-dependent inpatients. By dampening the automatic tendency to approach alcohol-related stimuli, individuals are allowing time to make more informed behaviour choices (i.e., improving their decision-making ability). Its impact will be examined through abstinence rates at 2-weeks and 3-months relative to those receiving sham training. The findings are likely to have implications for the design and delivery of psychosocial interventions delivered during early recovery from alcohol-dependence, aiming to optimise treatment effectiveness. Aims: 1. To determine if 4-sessions of CBM, using an alcohol approach-bias modification training program, delivered during inpatient withdrawal, reduces craving and relapse rates and other drinking-related outcomes at 2-weeks and 3-months post-discharge. 2. To determine if CBM, using alcohol approach-bias modification training, improves decision-making after four training sessions. 3. To determine if CBM effects (i.e., using alcohol approach-bias modification training) are moderated by impulsivity. Hypotheses: 1. CBM (using alcohol approach-bias modification training) will lead to significantly higher rates of abstinence at 2-weeks and 3-months and larger reductions in days to relapse, percentage of heavy drinking days and craving score, relative to sham training. 2. CBM (using alcohol approach-bias modification training) will be associated with significantly greater improvement in decision-making (Iowa Gambling Task performance) after the fourth training session compared to sham training. 3. Patients with greater impulsivity will show significantly greater responses to CBM (using alcohol approach-bias modification training). Methodology: The study will be a parallel-groups randomized superiority trial comparing CBM (using alcohol approach-bias modification training) versus sham training (i.e., control condition) in alcohol-dependent participants following residential detoxification treatment. The primary outcome variables will be self-reported alcohol use (relapse versus abstinence) 2-weeks and 3-months post discharge and secondary outcomes will include: days to relapse, percentage of heavy drinking days, alcohol craving and Iowa Gambling Task performance (i.e., decision-making) immediately post-intervention. Assignment: Randomisation will be carried out according to the ICH Guideline by an independent statistician not involved in the day-to-day conduct of the trial. Following baseline assessments, participants will be assigned randomly to the treatment condition using a 1:1 ratio and randomly permuted blocks. Participants: The target sample is defined as alcohol-dependent patients seeking treatment at the detoxification inpatient unit in Wellington House, Box Hill, Victoria. Setting: Wellington House, an inpatient detoxification service; part of Turning Point, a state-wide Alcohol and other Drug (AOD) service that incorporates Eastern region based services. Measures: Standard demographic questionnaire including: age, gender, history of alcohol use, previous alcohol treatment, employment, income source, housing arrangements, family situation, mental health. Current medication regimen, including psychotropic medications. All benzodiazepines will be converted to standard dose of diazepam. Baseline drug use and clinical measures: Alcohol/Drug use: Timeline Followback Interview (TLFB; Sobell \& Sobell, 1992): as a measure of frequency of alcohol/drug use in the past 4 weeks. Baseline alcohol dependence: Severity of Alcohol Dependence Questionnaire (SADQ; Stockwell et al, 1983). General cognitive functioning: Montreal Cognitive Assessment (MoCA): a brief cognitive screening tool of general cognitive functioning (Nasreddine et al, 2005). Appetitive motivational behaviour/ impulsivity: The Sensitivity to Reward (SR) part of the Sensitivity to Punishment and Sensitivity to Reward Questionnaire (SPSRQ; Torrubia et. al., 2001); specifically measures appetitive motivational behaviour or impulsivity. Impulsiveness, venturesomeness, empathy: The I7 Impulsiveness questionnaire (Eysenck, Pearson, Easting \& Allsopp, 1985) is a 54-item measure consisting of three scales: impulsiveness, venturesomeness, and empathy. Craving: Assessed using 100mm visual analogue scales (VAS). The VAS provides an index of urge to drink using two dimensions (positive and negative urge; Dawe \& Gray, 1995). Craving: The Alcohol Craving-Short-Form-Revised (ACQ-SF-R: Singleton, 2000) is a 12-item measure of craving for alcohol among alcohol users in the current context (right now). Depression and anxiety: Seven depression and seven anxiety items of the shortened Depression Anxiety and Stress Scale (DASS; Lovibond \& Lovibond, 1995). Decision-making: A computerised version of the Iowa Gambling Task (IGT; Bechara et al., 2000): measures decision-making under conditions of uncertainty and risk of punishment. Procedures: Recruitment and consent: Participants will be approached by a researcher no earlier than day three of admission. All Participants will be provided with a Plain Language statement explaining the purpose of the study and what will be required of participants. The study will also be explained verbally by the researchers. If the participant chooses to become part of the study a written consent form (i.e., Patient Information and Consent Form) will be completed. The researcher will also seek consent to contact others and document relevant telephone/contact details so that self-reported outcome can be assessed 2-weeks and 3-months post-discharge. Baseline, intervention and post-intervention outcome assessment: Following the consent taking, the baseline assessment will be completed as described in the measures section. Participants will then be randomised to one of two conditions, cognitive bias training or sham training. On the morning of days 3/4 to days 6/7 (i.e., four consecutive days) of admission, participants will undergo the ABM (intervention). Each training session is expected to last approximately 15 minutes. Participants will be given the opportunity to take an optional break during the intervention. After the final (i.e., fourth) training session, participants will complete the post training assessment (decision-making task and craving measures). Two-weeks and 3 months post discharge the researcher will telephone the participant to use the timeline follow-back instrument to document alcohol consumption since discharge and ask questions about engagement in treatment since discharge. Statistical Analyses: The primary hypothesis uses a binary outcome variable (abstinent versus relapse at 2-weeks) following completion of the protocol (four sessions) and will be examined using chi-square analyses. Secondary outcomes which are continuous variables (i.e., days to relapse, percentage of heavy drinking days, mean number of standard drinks, craving scores, IGT score etc.) will be examined using mixed effects repeated measures models (MMRM) including type of intervention (cognitive bias modification training versus sham training) as the independent variable and craving, decision-making, days to relapse and other alcohol consumption outcomes as dependent variables. The third hypothesis will be tested with a multiple regression model including sensitivity to reward and impulsivity scores at baseline, as predictors of outcome (abstinence, days to relapse etc..). Power calculations based on a hypothesized medium effect size indicate that a sample size of 72 participants (36 in each group) is required to test the study hypotheses with 80% power.