The role of temperamental self-regulation in predicting treatment outcome in alcohol use disorders

BACKGROUND:  
Alcohol Use Disorders (AUDs) are highly prevalent psychiatric disorders and often have a poor treatment outcome in terms of high drop-out rates and relapses. Low Effortful Control (EC) seems to be a core risk factor associated with both the initiation and continuation of AUDs. EC refers to regulative temperament and reflects cognitive control that develops later in life parallel with the maturation of the prefrontal cortex. AUDs are associated with a wide range of cognitive deficits.
We investigated whether EC, measured by means of self-report and behavioral measures, can predict treatment outcome (relapse, clinical symptomatology) in men and women with AUD.
When a low level of EC indeed turns out to be a significant predictor of relapse and lack of decrease in clinical symptomatology, treatment interventions aiming at strengthening EC could result in better treatment outcomes in AUD patients.
METHODS: 
The sample consisted of 75 adult patients with AUD (68.9% males, mean age 45,49 years (SD=9.96) admitted at a treatment unit for addiction. To assess EC we used the Effortful Control Scale from the Adult Temperament Questionnaire (self-report) as well as five behavioral tasks of the Cambridge Neuropsychological Test Automated Battery: Cambridge Gambling Task (CGT), Stop-Signal Task (SST), Intra-Extradimensional Set Shift (IED), Spatial Working Memory (SWM) and Spatial Span (SSP). To assess clinical symptomatology, we used the Symptom Checklist-90.
RESULTS: 
We performed binary logistic regression analyses with both EC and CANTAB measures as predictors and relapse/no relapse (during treatment and after 3 months FU) as dependent variables. According to these analyses only poor attentional set shifting and flexibility (IED) was significant in predicting relapse (p<0.05) after 3 months FU.
Repeated measures ANOVA’s showed that for men treatment was effective (i.e. symptomatology decreased) from baseline, over end of treatment, till 3 months FU; for women symptomatology decreased from baseline till end of treatment but increased slightly again at 3-months FU.
Linear regression analyses with EC/CANTAB measures as predictor and SCL-90-total difference scores (T2-T1) as dependent variables were performed. Neither measures were  able to predict reduction of total clinical symptomatology.
CONCLUSIONS
In our study we investigated whether EC could predict relapse and clinical symptomatology reduction (SCL-90) during treatment and after 3 months FU in patients with AUD. Only poor attentional set shifting and flexibility (IED) was significant in predicting relapse after 3 months FU. 
Further EC nor CANTAB measures could predict treatment outcome in terms of reduction in clinical symptomatology.
One hypothesis for these findings is that our sample of inpatients with AUD at a specialized addiction service is too homogeneous, all presenting lower levels of self-regulation. 
Future research should thus focus on larger samples and less homogeneous population.