To identify clinical cues indicative of depression in medical records of cases in primary care with undetected depression.

Depressive disorders are common; the lifetime risk for men and women is 27% and 45%, respectively. Despite effective treatment methods such as antidepressants and cognitive behavioural therapy, depression often remains undiscovered in primary care, with great implications both on the individual and societal level.

Clinical cues indicating depression were sought in medical records the year before an opportunistic screening for depression in primary care. In a previous study of 221 patients in the waiting room of a primary care centre during 10 randomly selected days, 45 (20%) showed signs of depression (MADRS-S ⩾ 12) and 60% of these were verified as having depressive disorders (Prime-MD). These 45 patients constitute the cases in the present study. Age- and gender-matched controls were selected among those who scored below the chosen cut-off level.

Seventeen (38%) of the 45 cases compared with eight (18%) of the 45 controls had one or more cues [odds ratio (OR) 2.81; 95% confidence interval (CI): 1.06–7.43]. Sleep disturbance showed the greatest difference between cases and controls (OR 4.53; 95% CI: 1.17–17.55). A significant relationship was found between severity of depression, frequency of cues and lower functional level. Cues were twice as common in patients with undetected depression and their functional level was lower. A two-stage procedure, screening and a structured diagnostic interview, is recommended when sleep disturbances and lowered function are present.