Research That Matters (January 17 - 20, 2008)

Background and Purpose: Within the alcoholism field, there is mounting evidence supporting an important relationship between treatment adherence and outcomes in both pharmacological and behavioral treatment. However, little is known about what combinations of treatment factors produce both good adherence along with better outcomes. The COMBINE Study (combining medication and behavioral treatment for alcohol problems), an NIAAA-funded, landmark study was designed to test the efficacy of two promising medications, naltrexone and acamprosate and two kinds of behavioral interventions, a low intensity medication management (MM) and a moderate intensity combined behavioral intervention (CBI). The latter intervention is particularly applicable to social workers treating alcohol problems in specialized settings. The present paper reports on the findings on medication and treatment adherence at 16-weeks. A major aim of this study was to determine whether differences in medication and treatment adherence impact separately or concomitantly on the various treatment combinations to produce different outcomes.

Methods: The study sample (n= 1383) was recruited from hospitals and clinics which typically serve individuals with alcohol problems. Medication adherence was defined as the ratio of number of pills (active medication or placebo) taken by the number of pills expected to be taken. With regard to treatment adherence, data were regularly collected on session attendance. Two primary outcomes were percent days abstinent (PDA) and time to relapse (TTR). A linear mixed model was used to estimate the effects of adherence on PDA while proportion hazard model was used to examine indicator effects on TTR.

Results: Nonadherent participants in this study did more poorly than adherent participants. By 16 weeks, the mean PDA for medication adherents was 82% vs. 72% for nonadherents (p < 0.0001). Adherence also had significant effect on TTR. The odds of relapsing were significantly decreased if the patient remained adherent to the prescribed pills (active or inactive medications) (O.R .453, CI 0.346- 0.594, p <0. 0001). Concerning the relationship between medication adherence and behavioral treatment, findings indicate a three-way interaction between adherence, study medications and CBI. Namely, the odds of relapsing were greater for medication nonadherents if they were not exposed to CBI (O.R. 0.568, CI 0.328- 0.984, p = 0.04). However, among medication adherents, CBI did not significant increase the risk of relapse.

Conclusions and Implications: Among medication nonadherents, particularly those assigned to an inactive medication condition, exposure to CBI provided a beneficial effect in reducing the median relapse time; those not taking medication had the highest relapse rates unless exposed to CBI, which then lowered the relapse rate. Thus, CBI may be more beneficial than MM in addressing issues often experienced by nonmedicated patients such as craving, stress, and poor social skills, factors related to recovery from alcohol problems.