Development of a computerized adaptive test for depression.

Area of Science:

• Psychiatry and Mental Health
• Medical Informatics
• Psychometrics

Background:

• Psychiatric diagnosis heavily relies on patient self-report, necessitating improved accuracy and efficiency.
• Current depression assessment methods can be burdensome for patients and clinicians.

Purpose of the Study:

• To develop a Computerized Adaptive Test for depression (CAT-DI) to enhance measurement precision and reduce patient and clinician burden.
• To validate the CAT-DI against established depression assessment scales and diagnostic interviews.

Main Methods:

• A case-control study involving 1614 participants with and without depression.
• Development of the CAT-DI, a novel adaptive testing instrument for depression.
• Convergent validity assessed using Hamilton Rating Scale for Depression, PHQ-9, and CES-D; diagnostic classification via Structured Clinical Interview for DSM-IV.

Main Results:

• The CAT-DI achieved a depression severity estimate with 0.3 SE using an average of 12 items, correlating at r=0.95 with a 389-item scale.
• Demonstrated high diagnostic accuracy for major depressive disorder with 0.92 sensitivity and 0.88 specificity.
• Showed strong correlations (r=0.8) with other depression rating scales and provided excellent discrimination across the full depression severity spectrum.

Conclusions:

• Computerized Adaptive Testing (CAT) fixes measurement uncertainty, allowing variable item administration, unlike traditional fixed-item scales.
• The CAT-DI significantly reduces the number of items required for depression assessment.
• This adaptive approach leads to increased precision in measuring depression severity.