Relation of client variables to continuance in five types of alcohol treatment settings.

Multiple regression analyses were used to assess the relationship of 29 client variables to continuance in five types of alcohol treatment settings (outpatient, inpatient, social detoxification, halfway houses with highly structured treatment regimens, and halfway houses with minimal structured treatment; N = 3,240). Socioeconomic variables were most often related to continuance; however, the pattern of the relationship between predictor variables and continuance varied greatly from setting to setting. The strongest multiple regression coefficients were found for outpatient and inpatient settings and the weakest were found for the two types of halfway house settings. It was concluded that archive-derived variables had significant, but limited, power to predict continuance and that other approaches to investigating continuance are needed.