Drinking in the year after treatment as a predictor of three-year drinking outcomes *.

TREATMENT FOR ALCOHOL-USE DISORDERS (AUDs), at least in the United States, and self-help organizations such as Alcoholics Anonymous (AA) aim to help individuals to achieve and sustain abstinence from alcohol and other drug use. Consistent with this goal, the ability to maintain reductions in alcohol use and related behaviors during the initial period following an episode of treatment may be critical, because the first year following AUD treatment has been recognized clinically as a highrisk period for relapse. Defined broadly, relapse is the reinstatement of alcohol use that is associated with problems (Maisto et al., 1998). Furthermore, it is often assumed that relapse during this high-risk period is associated with poorer longer-term functioning (Maisto et al., 2003; Marlatt and Donovan, 2005).

Drinking stability and longer-term functioning

A number of studies published over the last three decades suggest that there is a positive association between shorter-term (up to 2 years after alcohol treatment initiation or completion) intervals of abstinence from alcohol and later (up to 10 years) alcohol use and related areas of functioning (Humphreys et al., 1997; review of studies by McKay and Weiss, 2001; Polich et al., 1980; Vaillant and Milofsky, 1982). One of these studies also suggested that sustained shorter-term moderate alcohol use predicts positive longer-term outcomes among some patients (Polich et al., 1980). However, there have been only several studies that have more directly and in more detail addressed the relationships between shorter-term alcohol use and later functioning.

In two of these studies, Maisto and colleagues (1998, 2002) found that patients who sustained 12 months of complete abstinence from alcohol immediately following the initiation or completion of AUD treatment were drinking significantly less at follow-up than those patients who had done any drinking during this period. Additionally, first-year abstainers showed superior marital functioning and reported fewer alcohol-related hospitalizations at the 30-month follow-up (Maisto et al., 1998) and improved self-efficacy and psychological outcomes at 24 months posttreatment initiation (Maisto et al., 2002). In both of these studies, differences between the two drinker groups pretreatment were controlled for in analyses of differences between the groups at follow-up. Thus, a period of sustained abstinence immediately following the initiation or completion of AUD intervention may be important for healthier longer-term functioning, although the minimum duration of abstinence necessary to predict more favorable longer-term functioning remains unknown.

Data from Stout (in Clifford et al., 2006) begin to address the question regarding minimum duration of abstinence. The analyses reported concerned the relationships among longest intervals of abstinence, drinking, and heavy drinking, respectively, during the baseline and 1-year period posttreatment initiation, respectively (six predictors), and these same drinking intervals at Year 3 posttreatment initiation. The baseline and 1-year drinking interval data were categorized into approximate thirds for testing in the prediction models. These models also included a priori defined covariates that might be associated with the dependent variables, as well as the treatment condition (case monitoring or control) to which participants were randomly assigned. The results of these analyses showed that the longest abstinence interval in Year 3 was significantly predicted by the longest abstinence interval in Year 1 (positive relationship) and the longest drinking interval in Year 1 (negative relationship). The longest drinking interval and the longest heavy drinking interval in Year 1 predicted the longest heavy drinking interval in Year 3, and the only significant predictor of the longest drinking interval in Year 3 was the longest drinking interval in Year 1. Furthermore, participants who had shorter intervals of abstinence in Year 1 had a median longest abstinence interval of about 1 month at Year 3. However, participants who were abstinent for all of Year 1 tended to be abstinent all of Year 3.

First-year moderate drinking

As is apparent from the literature described thus far, there are far more data on the relationship between posttreatment intervals of abstinence and longer-term functioning than there are on the association between posttreatment moderate drinking and longer-term functioning. This question is not only of academic interest; periods of sustained moderate, nonproblem alcohol use in some percentage of individuals following AUD treatment (even following programs that emphasize abstinence as the substance use outcome goal) occur often (e.g., Miller et al., 1992). Moreover, the recent emphasis of harm reduction as an underlying substance use disorder treatment philosophy (Single, 1995) places additional significance on the relationship between shorter-term moderate drinking and longer-term functioning.