Controlled drinking : a viable treatment goal in alcohol use disorder?

Abstract

Alcohol use disorders (AUD) has one of the largest treatment gaps among psychiatric diagnoses, with a treatment coverage of 10 to 20 %. One of the key contributing factors is the lack of treatments aiming for controlled drinking (CD) instead of abstinence. Although a large number of trials investigated CD as the outcome, there are major limitations to the existing studies, such as small sample sizes, non-adequate control conditions, and heterogenous definitions of CD outcomes. Further, very few studies have investigated clinically relevant predictors of outcomes specifically in CD in sufficiently large sample sizes. Efficacy studies have been the main focus in research on CD and very few studies have investigated patient perspectives on the treatments offered for a CD goal. Lastly, there are few validated clinically relevant measures for the assessment and evaluation of impaired control over alcohol consumption, which is predictive of outcome in treatment for controlled drinking.

The primary aim of the thesis was to investigate if a CD goal was viable in a treatment seeking sample of individuals of patients with AUD. This aim was broken down in the four following studies. Study I was a randomized controlled superiority trial including 250 individuals with alcohol use disorder. We hypothesized that Behavioral Self-Control Training (BSCT), a five-session cognitive and behavioral treatment, would be superior to Motivational Enhancement Therapy (MET), which was a four-session treatment based on Motivational Interviewing in reducing weekly alcohol consumption. Linear mixed models were used to analyze primary and secondary outcomes alongside with a Bayes factor analysis for the primary outcome. No differences were identified between groups for the primary outcome of mean weekly alcohol consumption at 26 weeks (primary endpoint). The secondary outcome proportion of weeks with hazardous drinking defined in line with the former Swedish low-risk drinking recommendations of >9/>14 weekly standard drinks was found to be statistically different between groups.

Study II was a prospective cohort study, in which we investigated differences at 52 weeks post inclusion between BSCT and MET for both alcohol consumption and related consequences as well as predictors of treatment outcome. Linear and logistic mixed regression models were used for the outcomes at 52 weeks, and linear and logistic regression models for the predictor analyses. BSCT was superior to MET for the change between baseline to 52 weeks for the outcome of CD, which was defined as drinking levels in line with the new low-risk drinking levels of less than10.0 weekly standard drinks per week for both women and men(p=0.048). In the sample altogether, 50.5 % succeeded with a CD level at 52 weeks. A total of 57 % of individuals in BSCT attained a level of CD, as opposed to 43 % in MET, which was a statistically significant difference. The clinical characteristics in individuals with CD levels compared to those not attaining a CD goal showed that there were substantial differences om clinically relevant outcomes, such as quality of life and reduction of risk and harm, favoring those with CD.

Further, Study II showed that predictors for obtaining CD and reducing weekly alcohol consumption was a lower baseline alcohol consumption. Predictors of symptom reduction in AUD were lower severity of AUD at baseline and a lower self-rated impaired control over alcohol consumption. Women were also found be significantly more successful in attaining CD, at both 12, and 52 weeks post inclusion. In conclusion, the study suggested that females and receiving BSCT was more favorable for attaining a long-term goal of CD. Baseline levels of AUD, baseline consumption and impaired control were corroborated as predictors of outcome.

Study III was a qualitative interview study with the aim to investigate how patients experienced MET as their treatment for AUD and a goal of controlled drinking. Fifteen patients (8/7 female/male) were recruited at the 26-week follow-up. Interview data was analyzed with thematic analysis. Five themes were identified: the therapist conveyed the MI-spirit, the therapist did not guide on how to reach the goal, participants were committed to change before starting treatment, participants were uncertain if treatment was enough to maintain change, and significant others were not wanted in sessions. One conclusion from this study was that there may be a need for modification of the MET manual, to support some patients in attaining a CD goal. The format may need to be prolonged to support patients sense of self-efficacy to change. Lastly, significant others were important for the support of change without necessarily being present in sessions.

Study IV was a psychometric study which evaluated the psychometric properties of the Swedish version of the Impaired control scale. We aimed to investigate dimensionality, reliability, convergent and divergent validity, measurement invariance and sensitivity to change for the evaluation of psychometric properties. The analyses of dimensionality by a principal component analysis of Rasch residuals indicated some multidimensionality created by two items (12, 22) (I can stop/ I would be able to stop/ before getting completely drunk). Tests of convergent and divergent validity showed that failed control had the strongest associations to impaired control and alcohol use disorder while the attempted control part was not associated with the construct of impaired control or alcohol use disorder. The conclusion from this study was that the failed control scale was the most valid measure of impaired control, and was sensitive to change when measured again post treatment. This makes the ICS (except the attempted control subscale) suitable as an assessment- and treatment evaluation instrument in AUD treatment of drinking.

The main conclusions of the thesis were that CD was a viable goal to the majority of individuals in a sample of treatment-seeking individuals with AUD with a low level of psychiatric comorbidity, one year after receiving psychological treatment aiming for CD. When comparing outcomes on alcohol- and related consequences as well as quality of life, individuals with a CD level showed more favorable outcomes than those who were not at this consumption level. Baseline alcohol consumption was corroborated as a key predictor of outcome in treatment for CD, for the outcomes of CD and mean weekly consumption, while AUD severity and impaired control were corroborated for the prediction of reduction of AUD symptoms. Although not proven to be superior for the primary outcome, there was supporting evidence that patients who received BSCT were shown to attain a CD level in more cases both at 12 weeks and at the one-year follow-up. Further, women were far more successful in attaining a CD goal. In order to increase self-efficacy to change in patients receiving MET, the manual may need modification on how to adapt treatment to a CD goal, as well as delivering MET in more flexible format. Lastly, the Swedish version of the Impaired control scale (except the attempted control subscale) can be used in health care settings and in research for the assessment and evaluation of treatment for CD.