Screening for Anxiety and Depression in Community Mental Health: An Evaluation of a Series of Brief Screens
Catherine Greeno, PhD, University of Pittsburgh, Shaun M. Eack, MSW, University of Pittsburgh, and Jonathan B. Singer, LCSW, University of Pittsburgh.
Purpose: Accurate diagnosis is one of the most important components to any psychological treatment. Unfortunately, evidence suggests that routine clinical diagnoses made in community mental health settings are not reliable and often inaccurate. The utilization of standardized methods of psychiatric assessment can improve the accuracy of diagnoses made in such settings. However, the vast majority of standardized assessments require both large amounts of training and time to effectively administer, which hampers their application to community settings. The identification of brief, yet valid methods to standardize psychiatric assessment could be of substantial value to social workers in community mental health settings. The purpose of this research was to examine, through a series of studies, the ability of three commonly employed measures of anxiety and depressive symptomatology, the Beck Anxiety and Depression Inventories and the Patient Health Questionnaire, to aid in the diagnosis of two of the most common psychiatric disabilities: anxiety and depression. We sought to accomplish this by comparing the diagnostic accuracy of these instruments to a well-known gold standard (Structured Clinical Interview for DSM-IV [SCID]). Methods: Data from Studies 1 and 3 (N = 288) were subjected to receiver operating characteristic curve analyses to identify optimal diagnostic cut-points for the Beck Anxiety Inventory (BAI) and Beck Depression Inventory (BDI) for diagnosing anxiety and depression, respectively. The diagnostic accuracy of these instruments at various cut-points was validated in comparison with the SCID. Data from Study 3 were used to verify independent diagnoses made by the Patient Health Questionnaire (PHQ) with those made by the SCID. Results: Receiver operating curve analyses indicated that the BAI held the most utility for diagnosing panic disorder. However, even for this diagnosis, its utility was limited, as with an optimal cut-point of 14, the BAI failed to diagnose 10% of cases with panic disorder, and over-diagnosed 30% of non-panic disordered individuals. The BDI proved to hold the greatest utility for diagnosing major depressive disorder. However, similar to the BAI, with an optimal cut-point of 16, the BDI failed to identify 25% of assessed cases with major depression, and over-identified 26% of non-depressed cases. The diagnostic utility of the PHQ was also limited, as nearly half (44%) of all individuals who had been diagnosed with an anxiety disorder by the SCID were not detected by the PHQ. Likewise, the PHQ failed to detect 35% of individuals who had a depressive disorder diagnosed by the SCID. Implications: These findings indicate that the BAI, BDI, and PHQ all hold limited utility for aiding the diagnosis of anxiety and depression in community settings. Clinical social workers need to be cautious when employing these instruments as diagnostic aids in community settings. Social work researchers need to continue to work to develop methods of standardizing psychiatric diagnosis that are both valid and applicable to community settings. The development of such methods is vital to improving the care individuals receive in community mental health clinics.