Background: Dementia is said to be under-recognized and sub-optimally managed in primary care, but there is little information about actual processes of diagnosis and clinical care.

Aim: To determine general practitioners’ concordance with clinical guidelines on the diagnosis and management of patients with dementia.

Design: Unblinded, cluster randomized pre-test–post-test controlled trial involving 35 practices in the UK.

Methods: Patients with a diagnosis of probable or confirmed dementia were identified in practices, and permission sought from the older person and/or their carer to study the medical records of these patients. Medical records were reviewed using a data extraction tool designed for the study and based on published guidelines, and unweighted scores for diagnostic concordance and management concordance were calculated.

Results: We reviewed 450 records of patients aged 75 and over with a diagnosis of dementia and found that: only 4% of cases were identified first in secondary care; two-thirds of those identified in primary care were referred immediately; about one-third identified had informant history and blood tests documented at the Index consultation and one-fifth underwent cognitive function testing.

Discussion: The records analysed in this study came from a period before the Quality Outcomes Framework and show that the documentation in primary care of the diagnostic process in dementia syndromes is good, although there were significant gaps, particularly around depression case-finding. Information about management processes were less evident in the records.