The problem of high levels of client drop-out in drug and alcohol treatment is frequently reported in the literature. In the course of conducting an evaluation of a specialist women's treatment service, the inadequacy of the data on women-specific predictors of treatment drop-out was highlighted. Using a retrospective design, the characteristics of 160 women who left treatment less than 5 days after admission were compared to the 160 women who stayed longer than 5 days and were enrolled in the evaluation study. The findings of this study suggested that women who were employed, had a history of sexual assault (especially in adulthood), nominated alcohol as their drug of choice, were not married, older than 25 years of age and had demonstrated a sympathy with the agency's treatment philosophy were less likely to drop-out of treatment. In addition, for lesbian women, women with a history of sexual assault in childhood, and those with dependent children, attendance at a specialist women's service reduced the incidence of treatment drop-out.