Predicting attrition from substance misuse treatment using the Inventory of Interpersonal Problems

A simple, self-report scale to predict who would leave substance misuse treatment against medical advice was developed and tested. Scale items were drawn from the Inventory of Interpersonal Problems which measures interpersonal distress in the same format that the SCL-90-R uses to measure personal distress. Potential items were selected using data from an initial sample of 66 patients at a substance misuse clinic. Factor analytic techniques were then used to decide which of these items to include in the scale. It was tested using 98 patients not included in the initial sample. Logistic regression analysis confirmed that patients with high scores on the scale were significantly more likely to leave treatment than were patients with low scores.