Accuracy of general practitioner unassisted detection of depression

Objective: Primary care is an important setting for the treatment of depression. The aim of the study was to describe the accuracy of unassisted general practitioner judgements of patients' depression compared to a standardised depression-screening tool delivered via touch-screen computer. Method: English-speaking patients, aged 18 or older, completed the Patient Health Questionnaire-9 (PHQ-9) when presenting for care to one of 51 general practitioners in Australia. General practitioners were asked whether they thought the patients were clinically depressed. General practitioner judgements of depression status were compared to PHQ-9 results. Results: A total of 1558 patients participated. Twenty per cent of patients were identified by the PHQ-9 as being depressed. General practitioners estimated a similar prevalence; however, when compared to the PHQ-9, GP judgement had a sensitivity of 51% (95% CI [32%, 66%]) and a specificity of 87% (95% CI [78%, 93%]). Conclusions: General practitioner unassisted judgements of depression in their patients lacked sensitivity when compared to a standardised psychiatric measure used in general practice.