Drunkenness arrests: predictors of recurrence and effect of detoxication treatment

Predictors of the recurrence of drunkenness arrests and the effect of detoxication treatment on outcome during a 12-14-month follow-up were studied among 560 arrestees. Arrestees were assigned on the basis of the last digit of their day of birth to a treatment group (odd digits) or a control group (even digits). Treatment was compulsory and consisted of a bath, rest, food and care provided by social workers, nurses and a physician at a detoxication station. In regression analysis, the number of subsequent arrests was significantly related to housing status, age, marital status and occupational status but not to treatment. Survival analyses indicated that the first and the second new arrest were significantly associated, in addition to the above-mentioned predictors, with sex but not with treatment. Relative risks between selected categories of predictor variables, estimated using Cox's proportional hazard models, varied between 1.2 and 2.5. Compulsory detoxication does not decrease the recurrence of drunkenness arrests.