Predicting drop-out during the systems training for emotional predictability and problem solving (STEPPS)
background Drop-out is a complex problem in mental health care and in stepps. Research has revealed a variety of predicting factors and has produced contradictory results.
aim To investigate whether the information available at the start of stepps can pinpoint predictors of drop-out.
method The rom data for 150 patients were used to test the link between the following factors: age, gender, education, employment, substance abuse, anxiety, hostility, interpersonal relations, responsibility and social concordance with drop-out. The method used for testing was logistic regression analysis.
results Factors that contributed significantly to the prediction of drop-out were gender and employment status. These factors made up 16% of the explained variation (R2 Nagelkerkes) in drop-out. Gender was the strongest predictive factor. Concerning the other factors, no differences were found between groups (drop-out and non-dropouts).
conclusion In its present form stepps does not suit a large number of the male participants. Drop-out during stepps is hard to predict on the basis of rom-questionnaires. Future research should focus on preconditions and marginal conditions that influence patients to complete their training.