Apart from nihilism and stigma: what influences general practitioners' accuracy in identifying incident dementia?

To assess the accuracy of the General Practitioner's (GP) judgment in the recognition of incident dementia cases and to explore factors associated with recognition.

Prospective observational cohort study, two follow-up assessments (FU 1 and FU 2) within 3 years after baseline.

One hundred thirty-eight general practice surgeries in the six study centers of a prospective German study.

Participants were between 75 and 89 years of age at baseline and were recruited from the GPs' patient lists. In FU 1, 2,402 patients and in FU 2, 2,177 patients were analyzed.

GPs' judgments on their patients' cognitive status as index test; at-home patient interviews and tests, consensus diagnosis as reference; validity of the GP judgment; associations between patient factors and GPs' dementia recognition.

One hundred eleven incident dementia cases with complete data were identified in FU 1 and FU 2. Overall sensitivity of the GP judgment was 51.4%, specificity 95.9%, positive predictive value 23.6%, and negative predictive value 98.8%. GPs missed dementia more frequently in patients living alone. GPs overrated the presence of dementia more frequently in patients with problems in mobility or hearing, in patients with memory complaints, and in patients with a GP-documented depression.

GPs miss nearly half of incident dementia cases. They should be alert not to miss dementia in patients living alone. Without seeking additional information, a positive GP judgment seems not sufficient for case finding. GPs should be aware of their tendency to overestimate dementia in depressed and frail patients.