What is a drinking episode?

Objective: The phrase "drinking episode" is used informally in many ways. However, a scientific understanding of the factors affecting the length of drinking episodes and the way treatment components affect these episodes requires rigorous operational definitions, supported by evidence for the appropriateness of these definitions. Method: Daily drinking data from two studies (Project MATCH and BETA) involving a total of 1,955 subjects are examined by survival analysis methods to determine the prognostic significance of different durations of postdrinking abstinence. The dependent measures are "time to next drink" and "time to heavy drinking." Results: Curves relating postdrinking abstinence to subsequent drinking indicate that 1 day of abstinence has Little prognostic significance. As the duration of abstinence increases from 1 up to 60 days, longer abstinence has a decelerating but still positive association with time to subsequent drinking/heavy drinking. There is no apparent threshold point beyond which further abstinence has no further effect. Inflections in the curves suggest, however, that intervals of 1, 2 or 4 weeks of continuous abstinence may be important milestones. These general patterns seem to hold up across samples despite significant quantitative differences across studies. Conclusions: These results suggest that two different definitions of "drinking episode" may be useful in examining treatment effects on drinking behavior. These analyses help to provide a foundation for further quantitative research on treatment effects on addictive behaviors over time.