An accurate diagnosis is essential for the management of late-life depression in primary care. This study aims to (1) provide information on the agreement on depression diagnoses between general practitioners (GPs), dimensional tools (Geriatric Depression Scale [GDS], Hospital Anxiety and Depression Scale [HADS]) and a categorical tool (Structured Clinical Interview for DSM-IV criteria [SCID]) and (2) identify factors associated with different diagnoses. As part of the multicenter study "Late-life depression in primary care: needs, health care utilization and costs (AgeMooDe)" a sample of 1113 primary care patients aged 75 years and older was assessed. The proportion of depression was 24.3% according to GPs, 21.8% for the GDS, 18.9% for the HADS and 8.2% for the SCID. Taking GDS, HADS and SCID as reference standards, recognition of GPs was 47%, 48% and 63%. Cohen's Kappa values indicate slight to moderate agreement between diagnoses. Multinomial logistic regression models showed that patient related factors of depression were anxiety, intake of antidepressants, female gender, a low state of health, intake of medication for chronic diseases and functional impairment. GPs performed better at ruling out depression than ruling in depression. High levels of disagreement between different perspectives on depression indicate that they may be sensitive to different aspects of depression.
