Mental health computerized adaptive testing

Context: Mental health measurement relies upon antiquated approaches to the analysis, design and administration of psychiatric instruments. Objective: This study examines the utility of the combination of Item Response Theory (IRT) and computerized adaptive testing (CAT) for the calibration, administration, and scoring of psychiatric measurement systems. Design: To illustrate the application of IRT-based CAT to a problem in psychiatric measurement, we selected the Mood and Anxiety Spectrum Scales (MASS). The MASS consists of 626 items covering mood spectrum, panic-agoraphobic spectrum, obsessive-compulsive spectrum, and social phobic spectrum. Based on a balanced incomplete block design, 36 forms each consisting of 154 items were administered to 800 participants who were in outpatient treatment for a mood or anxiety disorder at Western Psychiatric Institute and Clinic (WPIC, Pittsburgh, PA). Results: A bi-factor IRT model provided an excellent fit to the observed data, and 90 % of the items were included in the item bank. Simulated CAT administration for subjects that had completed the entire test, revealed a 96 % average reduction in the number of items administered for CAT (24 items) relative to traditional scale administration (615 items). The correlation between tota