Development of reliable and valid screening methods for depression in multiple sclerosis

Abstract

There is a lack of standardized screening methods to detect depression in multiple sclerosis (MS), especially in light of the ability to treat depression in MS and high prevalence rate. Following a consensus conference on depression in MS, researchers recommended the use of the Beck Depression Inventory-II (BDI-II) as a well-validated depression screening instrument. Although it has been widely used with MS patients, there is a paucity of MS-specific empirical data on appropriate "cut-off scores." Thus, the primary objective of this study was to utilize a pre-existing depression database in MS to identify an optimal screening cut-off score on the BDI-II that would indicate the presence of a level of depression that typically warrants treatment and may be used as a short screen in the MS practitioner's office. Using Receiver Operating Characteristic (ROC) curve analysis, optimal sensitivity and specificity were determined to create a clinical cutoff for a treatable depression in this population on the BDI-II as compared to the "gold standard" for assessing depression, the Structured Clinical Interview for DSM diagnosis (SCID-I). The two items on the BDI-II reflecting depressed mood and anhedonia were treated as a subscale in accordance with Mohr's work (2007). When comparing patients with Major Depressive Disorder (MDD) to all other participants using these two items, the area under the ROC curve was found to be .764, with a sensitivity of 73.0% and specificity of 81.7% at a cutoff equal to or greater than 1.5. Comparing participants with MDD to all other individuals using only the two items on the BDI-II that reflect depressed mood and anhedonia produced the greatest sensitivity, specificity, and optimal cutoff. Thus, this assessment tool provides a quick and efficient means of detecting MDD in MS patients who need treatment, without the need for expensive resources such as personnel (e.g. nurses, doctors, or therapists).