Should physicians screen for depression in elderly medical inpatients?: Results of a decision analysis.

Abstract

We wish to determine whether or not elderly medical inpatients should be screened for depressive disorder using either 1) a self-rated depression scale (Geriatric Depression Scale), 2) "usual clinical assessment," or 3) neither, assuming that treatment with tricyclic antidepressants (TCAs) is the primary mode of intervention. Based on recent data from epidemiological studies on the prevalence and course of depression, the test characteristics of available screening tests, and the efficacy and side-effects of traditional antidepressants, decision analysis is used to help decide whether or not clinicians should screen for depression in this setting. These calculations indicate that if screening is done solely to identify depressed patients for treatment with TCAs, then the highest utility lies in not screening; however, the difference in utilities between that decision and the decisions to either screen with GDS or screen by usual clinical assessment was only .04 units on a 0 to 100 scale, making the decision virtually a toss-up. Furthermore, even a small variation in one of several clinical factors or test characteristics could give screening a higher utility. In particular, if psychotherapy is considered as the primary intervention, then the utility of screening exceeds that of not screening. Characteristics of the screening test, clinical setting, types and safety of available treatments, each impact on the usefulness of screening and must be kept in mind when diagnosing and treating depressed medically ill elders hospitalized in acute care settings.