Looking forward and back to relapse: implications for research and practice.

In this commentary, the three principal investigators of the Relapse Replication and Extension Project (RREP) reflect on clinical and research implications of study findings from the three collaborating sites. A primary purpose of RREP was to study the reliability and validity of a taxonomy of relapse antecedents originally proposed by Marlatt two decades ago. Under the best of research conditions, with extensive training and practice, it was difficult to achieve reliability of coding with the original three-level system, although with only two levels of classification more reasonable albeit variable reliability was found. Modifications may improve the taxonomy's reliability, but RREP data indicate that a more appropriate strategy is to measure possible antecedents of relapse by continuous scales such as those provided by Annis, Heather and Litman. There is reasonably consistent evidence for two common antecedents of relapse: negative emotional states, and positive emotional states in a social context. Antecedents of relapse show only modest consistency within individuals from one occasion to the next. The causes to which clients attribute relapses may exert a significant effect on future drinking episodes. Stable and internal attributions, such as are commonly associated with a dispositional disease model, may serve to perpetuate relapse. From the RREP studies, the availability of coping skills appears to be a potent protective factor, and ineffective coping a consistent predictor of relapse. Implications for clinical research and practice are considered.