Depression: to screen or not to screen?

There are few more difficult diagnoses to make than depression in sick older patients in a general hospital. There is no diagnostic test and confounding physical symptoms, the sheer fact of being ill in the inhospitable and frightening environment of a hospital away from home, possibly for weeks, and still facing an uncertain recovery makes interpretation of symptoms very difficult. When does a miserable and frightening experience become a mental disorder? But, we know that depression in this situation is common and adversely affects outcome, and so its presence matters. The association of depression with physical illness is strong, and multi-morbidity is the norm that has implications for research and medical education [1]. In a general hospital, the prevalence of depression in older people may be as high as 20–30% (2–3 times higher than in the community) depending on the diagnostic criteria used. The number of older people with depression is increasing quickly as the population ages [2], and depression is predicted to become the highest global cause of Disability Adjusted Life Years by 2030 [3]. To screen for depression in this hospitalised population would involve large numbers of patients, and any screening instrument would need to be easy and quick to administer with good sensitivity and specificity if it is to identify those in need of treatment yet not produce large numbers of false-positive cases that would require a longer diagnostic assessment. Furthermore, there needs to be sufficient evidence that treating depression would improve outcome. The paper by Baillon et al. [4] investigates an elderly medical population using a two-stage procedure supplementing the two screening questions recommended by NICE [5] with the 15-item Geriatric Depression scale (GDS). The diagnostic gold standard was the ICD-10 criteria for depression excluding those who scored <24 on the Mini Mental State Examination. Twenty-two per cent met the ICD-10 criteria for depression. Biological symptoms of depression in a physically ill population are particularly difficult to interpret and both NICE questions and the GDS purposefully avoid them as much as possible which usually improves specificity but at a loss of sensitivity. The GDS is the most validated screening instrument for depression in the hospitalised elderly population. The primary purpose of a good screening instrument is not to miss cases. On this level, the report by Baillon et al. found that answering positively to just one of the NICE questions was very effective with a sensitivity …