Results from a general practice study of cognitive impairment are used to construct guidelines for the detection of mild cognitive impairment (MCI) in general practice. The role of the general practitioner in the diagnosis of MCI is discussed and the potential feasibility of general practice screening is evaluated with data from a two-year prospective study conducted in the Montpellier region of France. Neuropsychological tests with the highest predictive value for dementia conversion and suitable for use in general practice, were determined by logistic regression modelling. An examination comprising three tests (delayed auditory verbal recall, verbal fluency and visuospatial construction), were found to give a specificity of 99% and sensitivity of 73%. Reports from other research centres suggest that MCI detection should not be limited to cognitive performance alone. It is suggested that proxy observations of behavioural change and information relating to loss of ability to perform activities of daily living should also be used to improve sensitivity and also to provide information needed in patient management. It may be potentially feasible for the family practitioner to verify cognitive complaints and to screen for MCI with a high degree of accuracy using a brief test battery derived from empirical observations in population studies.