Alcohol, Inflammation, and Mindfulness Study (AIM)

This study was designed to evaluate the effect of two outpatient, individually-administered, psychosocial interventions (Relapse Prevention vs. Mindfulness-Based Relapse Prevention). Participants are recruited via mass media advertisements, flyers, and local outreach to clinical providers that do not address substance use. We will recruit 226 individuals. Our previous experience with clinical trials, as well as the experience of our research team members with these interventions support the feasibility of recruiting approximately 50 patients per year. This sample size provides adequate statistical power for the research questions undertaken.

Participants are screened over the telephone and invited into the laboratory for a baseline assessment. Immediately after the baseline assessment, they are randomized to a therapist and then to a treatment condition. At the end of each year, the distribution of participants across conditions and therapists are evaluated for imbalances. If attrition has resulted in an imbalance across conditions, the randomization procedure will be adjusted to ensure equal numbers of participants in RP and MBRP.

Within a week of completing the baseline assessments, participants are seen by their individual therapist for a motivational interviewing, goal-setting session. After the goal-setting session, the eight-session MBRP or RP treatments are administered. To ensure that therapists adhere to treatment manuals, a checklist is provided consisting of the primary content addressed each session (four to seven primary content pieces each session). Further, questionnaires assessing mindfulness practice and facets of mindfulness (e.g., being nonjudgmental, nonreactive) are administered at each assessment. At the most recent RPPR, these questionnaires validated that participants in the MBRP condition were increasing in some facets of mindfulness during treatment (and up to three months after treatment).

We estimate that the completion of assessments and procedures across the duration of the trial, excluding therapy sessions, will total about 15 hours of time. To compensate subjects for their time and travel costs, each subject will be paid $300 if they complete all sessions.

We will conduct extensive analyses of any attrition encountered in the project to determine bias. However, note that in each of the analyses described herein, we will have the capability of using modern approaches to the handling of missing data including full information maximum likelihood estimation of missing data within Proc Mixed in SAS and within EQS.

The distributional properties of all continuously scaled variables will be examined for skewness and kurtosis to determine the need for normalizing transformations or for alternations to our analysis plan (i.e., generalized estimating equations or robust estimation) prior to the primary analyses.

To confirm the validity of random assignment, pretest equivalence of the two treatment conditions across demographics, drinking history, smoking, and all other baseline measures will be assessed via t-tests on continuous items and c2 tests of categorical items. We will use the Bonferroni approach to correct for alpha inflation with a familywise alpha of .05. Any variables on which the two groups are unequal at pretest will be covariates in all further analyses.

Attrition analyses will be conducted after each follow-up data collection effort to provide assurance that differential attrition by treatment condition has not occurred. Following previously published procedures, a series of ANOVAs of treatment (MBRP versus RP) X retention (retained, not retained) will be conducted on continuous baseline measures. Significant treatment X retention interactions identify measured variables on which differential attrition may have occurred. The logit model analog procedure will be applied to categorical baseline measures to test for differential attrition on categorical variables such as gender and ethnicity. We will conduct these analyses to assure that differential attrition by treatment condition does not account for any of the effects of the treatments.