Mental Health CAT (MH-CAT) for Community-Based Use

DESCRIPTION (provided by applicant): Mental disorders are among the most prevalent conditions in the United States, and their burden for the individual and the society is tremendous. Nevertheless, they are still widely under-diagnosed in community based health care systems. One way to help general practitioners (GP) identify and monitor mental health problems is to use self-administered patient questionnaires; however, the instruments that are currently available are either very lengthy or lack the measurement precision sought for individual decision making, and results from different tools are hardly comparable. The advent of computer adaptive testing (CAT) methods promises to solve these problems, but to facilitate their wide-spread use in clinical practice, appropriate software solutions must be developed and constantly supported. We propose to build a Mental Health CAT (MH-CAT) as part of a comprehensive software solution suitable for daily use in community-based settings. The MH-CAT will be a highly sensitive, easy-to-use first stage screening tool to help GPs identify and monitor patients in need of special mental health care services. The MH-CAT will be downward comparable to two of the most frequently-used fixed-length questionnaires (i.e., the CES-D and the MHI-5, representing one scale of the SF-36). To avoid floor effects and to make the tool meaningful for patients with and without mental health problems, we will include items measuring disease specific symptoms, as well as positive mental health states. In Phase I we propose to: (1) construct an item bank covering a bipolar euthymic-depressive symptoms continuum; (2) develop a comprehensive and flexible CAT software solution, including a physician report; (3) provide a cross-calibration approach to compare scores between the MH-CAT, MHI-5 and CES-D; and (4) demonstrate the usability of our CAT software in a field test in a community-based setting. To achieve these aims, we will use methods based on item response theory (IRT) to evaluate existing data sources of 13,547 patients and 256 different items from QualityMetric and RAND Corporation, as well as data from 60 patients included in the field test. In Phase II, we will test the diagnostic sensitivity and specificity of the MH-CAT compared with static questionnaires, extend the measurement to assess further mental health constructs (anxiety, distress, cognitive impairment), include additional features (i.e., assessment of treatment side-effects) to facilitate integration and widespread application of the tool, and test the impact of using the instrument.