Identifying Multiple Mechanisms of Change in Alcoholism Treatment

Cognitive-behavioral therapy (CBT) is an effective treatment for alcohol dependence, but
little is known about how CBT works to achieve these effects. Although several possible
mechanisms have been proposed to explain the effects of CBT, it is rare that more than one
mechanism is studied. However, it may be the case that similar outcomes (e.g., abstinence)
may be reached through multiple paths. Therefore, essential to conducting work on
behavioral change mechanisms is distinguishing different courses or paths and moderating
influences (Kazdin, 2007). In the present study, we will focus on 2 key mechanisms posited
to underlie the effectiveness of CBT specifically, increasing self-efficacy and
self-confidence and reducing positive outcome expectancies for alcohol use, and 2 key
mechanisms posited to underlie the effectiveness of a wide range of therapeutic
interventions, increasing the therapeutic alliance and reducing/regulating negative
emotional states. For the present study, participants will be 72 alcohol dependent men and
women who agree to participate in a 12-week trial of CBT for alcohol dependence. In
addition, comprehensive research assessments will be conducted with patients at baseline,
end of treatment, and 3-months posttreatment.

Specific Aim 1. To examine the within-treatment week-to-week relationship between ratings
of 4 key therapeutic mechanisms and alcohol involvement (operationalized as percent days
abstinent {PDA} and drinks per drinking day {DDD}) during treatment. It is hypothesized
that self-efficacy, outcome expectancies, therapeutic alliance (as rated by the patient and
therapist), and negative affect assessed after any given treatment session will each predict
alcohol involvement during the following week. Exploratory analyses also will investigate
the relationship of a given week's alcohol involvement on each of the four key variables as
assessed the following week.

Specific Aim 2. To examine profiles of the four key mechanisms over the course of treatment
in relation to alcohol involvement during treatment and during the 3-month follow-up period.
These analyses will be descriptive/exploratory. It is expected that profiles will emerge
with the key mechanisms that are associated with decreased alcohol involvement (for example,
when the alliance is relatively strong throughout treatment or when the alliance grows
stronger over the course of treatment, or when positive outcome expectancies grow weaker
over the course of treatment, etc.). Similarly, it is hypothesized that profiles will
emerge that are associated with little improvement in alcohol involvement (for example, when
negative affect is relatively high throughout treatment or when self-efficacy weakens over
the course of treatment).