Smith G R, Witt A S, Golding J M
Department of Psychiatry, University of Arkansas for Medical Sciences, Little Rock 72205-7199.
Control Clin Trials. 1991 Aug;12(4 Suppl):180S-188S. doi: 10.1016/s0197-2456(05)80022-4.
This study examined the relationship of function-specific mental health measures, anxiety and depression subscales of the Mental Health Index (MHI), to psychiatric diagnoses of generalized anxiety disorder and major depression, as assessed by the Diagnostic Interview Schedule (DIS). Focusing on the clinical relevance of the MHI scales, we evaluated their performance in screening for DIS-diagnosable disorders in a sample of primary care patients with multiple unexplained somatic symptoms. Receiver Operating Characteristic curves and examples of sensitivity, specificity, and predictive value for different cutoff scores on the MHI are presented. When sensitivity is set at approximately 90%, positive predictive value exceeds 50%. Because the base rates of disorder in this sample are high, however, the screening measure has little independent usefulness. Similarly, cutoff scores yielding negative predictive value of approximately 90% fail to detect most of the true-negative cases. Future research should examine the use of these scales in other high-risk populations.