Predicting treatment outcome by combining different assessment tools: Toward an integrative model of decision support in psychotherapy.

Abstract

Different types of evidences must be integrated to derive predictions about the possible outcome of psychotherapeutic interventions. We use the Minnesota Multiphasic Personality Inventory–2 (MMPI–2; Butcher, Dahlstrom, Graham, Tellegen, & Kaemmer, 1989) as a tool for gathering assessment data and propose a new method of statistical prediction to apply to psychotherapy. We apply artificial neural network (ANN) technology to a pool of clinical data gathered from the MMPI–2 profiles of 150 patients to derive indications about the outcomes of the interventions. ANNs are nonlinear computational devices, directly inspired by the principles of biological neural computation, that process information to learn patterns as a result of exposure to a set of representative “training cases.” The performance of the paper’s ANN in forecasting successful and unsuccessful treatment cases shows a mean rate of correct classification of 81%. This suggests that ANNs may be considered a potentially useful tool for supporting clinical practice and for deriving prognostic indicators for psychotherapy outcome.