The use of the Beck Depression Inventory to screen for depression in the general population: a preliminary analysis

Abstract

Objective: The aim of the present paper is to study the performance of Beck’s Depression Inventory (BDI) as a screening instrument for depressive disorders in a general population sample.

Methods: 1250 subjects, from 18 to 64 years old, were randomly selected from the Santander (Spain) municipal census. A two-stage method was used: in the first stage, all individuals selected completed the BDI; in the second, ‘probable cases’ (BDI cut-off≥13) and a random 5% sample of the total sample with a BDI score less than 13 were interviewed by psychiatrists using the Schedules for Clinical Assessment in Neuropsychiatry (SCAN), which generates diagnoses of depressive disorders.

Results: Our data confirm the predictive value of the selected cut-off point (12/13): 100% sensitivity, 99% specificity, 0.72 PPV, 1 NPV, and 98% overall diagnostic value. The area under ROC (AUC) was found to be 0.99. There were no statistical differences in terms of sex or age. We conclude that the BDI is a good instrument for screening depressive disorders in community surveys.

Introduction

There are now more than 50 instruments available for the detection of depressive disorders (Wittchen and Essau, 1990); one of the better known is Beck’s Depression Inventory (BDI) (Beck et al., 1961). Although initially developed to measure the intensity or depth of the depressive symptomatology in patients with psychiatric disorders, it is now widely used as a screening instrument to detect depression in clinical practice and research projects (Beck et al., 1988). Self-report questionnaires, such as the BDI, are particularly suitable in these contexts; in fact, this instrument is generally considered one of the best screening tools available for assessing depression in both psychiatric and medical settings (McDowell and Newell, 1996). There is, however, certain evidence suggesting that the BDI performs better as a screening instrument in psychiatric than in normal populations, compared with clinical evaluation. We have to point out, however, that this evidence is not derived, as it should be, from studies based on a non-biased sample of normal populations, using a reliable structured psychiatric interview for clinical evaluation (e.g. the PSE-9, the SCAN or the SCID). To our knowledge, there is no published study examining the BDI’s predictive values for the detection of depression in a representative sample of the working-age general population that uses a reliable psychiatric interview as a validation instrument. Thus, if we wish to adopt this questionnaire as a screening instrument for research use, we should first try to verify its criterion validity with an appropriate methodological design.

The objective of the present study was to assess, in a representative sample of the general population and using the Schedules for Clinical Assessment in Neuropsychiatry (SCAN; Vázquez-Barquero et al., 1994) system as the criterion against which the predictive value of the different cut-off points of the Spanish version of the BDI will be explored.