Unassisted detection of depression by GPs: who is most likely to be misclassified?

Meta-analyses indicate 50% of cases of depression are not detected by GPs. It is important to examine patient and GP characteristics associated with misclassification so that systems can be improved to increase accurate detection and optimal management for groups at risk of depression.

To examine patient and GP characteristics associated with GP misclassification of depression for patients classified by the Patient Health Questionnaire-9 as depressed.

A cross-sectional study within general practices in two states of Australia. GPs completed a one-page paper and pencil survey indicating whether they thought each patient was clinically depressed. Patients completed a computer tablet survey while waiting for their appointment to provide demographic information and indicate depression status. Chi-square analyses were used to determine whether patient and GP characteristics were associated with a false-negative and false-positive result. The probability of misclassification was modelled using Generalized Estimating Equations to account for clustering of patients.

Fifty GPs from 12 practices participated. GPs completed surveys for 1880 patients. Younger patients aged 25-44, and those with a health care card were less likely to have a false-negative assessment. Patients with 0-3 GP visits in the past 12 months, and those with private health insurance were less likely to have a false-positive assessment. GPs who worked five sessions or fewer per week were more likely to make false-positive assessments.