Abstract
Attrition is a common problem in outpatient mental health care settings, and can be understood as situations in which clients end treatment before achieving a clinically significant response. This archival study used a longitudinal method to look at the nature of attrition in an outpatient clinic, utilizing data for 3,728 clients, using the OQ 45.2. A Cox regression proportional hazards model was used in order to better understand who is likely to attrit when considering: (1) demographic groups, (2) diagnostic categories, and (3) process variables (e.g. overall and recent symptom change), using hazard ratios. A pattern emerged, with younger clients and those reporting less education and lower incomes tending to be more likely to end treatment. Consistent with the large scale STAR*D treatment of depression study, clients with more social and economic challenges demonstrated more risk. Adults diagnosed with a substance or OCD-related disorder showed the most elevated risk. Clients who demonstrated overall improvement and, in particular, a recent status change were more likely to remain. Engagement strategies are discussed, with the goal of better supporting recovery. Findings suggest that attrition is something that can be anticipated, identified, and reduced.