OBJECTIVE: To investigate the rate of diagnostic evaluation of dementia for patients in whom a suspicion of dementia was raised, and to investigate reasons why a diagnostic evaluation was not always being performed. DESIGN: A prospective study among elderly patients aged 65+, and a follow-up study. SETTING: In all, 17 general practices in Copenhagen with 40 865 patients on their lists of whom 2934 were aged 65+. SUBJECTS: A total of 793 patients consulting their GP regardless of reason of encounter, in October and November 2002. MAIN OUTCOME MEASURES: MMSE score < or = 23, GP clinical impression of dementia, laboratory-screening tests prescribed by the GPs and referral status after 6 months, and follow-up questionnaire. RESULTS: Of 793 patients a total of 138 patients were identified with possible dementia. Among the identified patients 26 (20%) were referred for further evaluation within 6 months, and 4 (3%) were treated for depression or referred for another condition. A total of 6 patients were lost to follow-up. In the remaining 102 undiagnosed patients the main reasons for not performing a diagnostic evaluation of dementia were patient/relative hesitation (34%), the GP thought that it would not have any consequences for the patient, or the GP estimated that the patient was too fragile (21%). CONCLUSION: In 17% of elderly patients in general practice a suspicion of dementia could be raised based on the clinical impression of the GP or MMSE score. However, only 23% of this group were evaluated by their GP or referred to a memory clinic within a subsequent period of 6 months.