To investigate the feasibility of early assessment of preventable disabilities in primary care, we developed a geriatric preventive screening examination with various indicators of physical, emotional, and social functions as well as laboratory exams. Cognitive impairment was measured by the modified MMSE. Severe cases of dementia, who would deserve home visits were excluded. Results of the assessment procedure in 446 patients aged 70 and over (71.5% females) were compared to ratings of general practitioners (n = 67). In these patients we found 4250 medical, 374 psychiatric, and 528 social problems. 45.4% of medical, 61.8% of psychiatric, and 56.8% of social problems where hitherto unknown to the GPs. The prevalence of cognitive impairment was 4.6% according to GPs diagnosis and 21% according to the MMSE. The sensitivity of GPs diagnosis was 14%, the specificity 98%, and the overall agreement measured by kappa was 0.17. There were significant (p < 0.05) associations of cognitive impairment with poor health, vascular disease, syncope, weight loss, previous hospitalization, depression, and ADL and IADL-items. Hypertension, or pathological thyroid function, occurred more frequently in the cognitively impaired (p > 0.05). Only 19.5% of dementia cases had severe functional loss, which substantiates our hypothesis that mild dementia was studied. Of all cases with newly identified cognitive impairment (n = 83 of 446 patients), three (3.6%) had reversible disorder such as depression (n = 1), drug toxicity (n = 2) 3 (3.6%) received counseling, and 5 (6%) further diagnostic assessment or treatment. One (1.2%) patient did not accept any treatment. In the remainder of 71 patients (85.5%), the GPs adopted a wait and see strategy with no intervention. In conclusion, memory deficits seem to be underdiagnosed in general practice despite much treatable comorbidity or social problems, and some reversible conditions such as depression and drug adverse effects.