Background: The purpose of this study was to evaluate the utility (i.e., positive and negative predictive value) of the 7 Minute Screen in identifying patients with probable Alzheimer's disease (AD) in a primary care practice. A second objective was to estimate the number of undiagnosed AD patients in a typical primary care practice. Methods: One hundred thirty-seven successive admissions (96%) of patients over the age of 60 to a primary care practice over a 53-day period who completed informed consent documents were administered the 7 Minute Screen. All patients who screened positive (n = 13) and a random sample of those who screened negative (n = 26) returned for full diagnostic evaluation. Positive predictive value (PPV) and negative predictive value (NPV) of the 7 Minute Screen were determined using the criterion standard of clinical diagnosis established by examination, history, and laboratory studies. Test-retest reliability and time for administration were also determined. Results: Of the 137 patients evaluated, 13 screened positive and 124 screened negative. Eleven of the 13 patients who screened positive were willing to return to the primary care practice for follow-up evaluation. A random sample of 26 patients who screened negative all agreed to return for follow-up evaluation. Of the 11 patients who screened positive who returned for evaluation, 10 were subsequently diagnosed with probable AD. The remaining patient was diagnosed with mixed dementia. The caregivers of the two patients who refused to return were contacted and both indicated that the patients were having significant cognitive problems as verified by an activities of daily living scale. Of the 26 patients who screened negative, 25 were judged to be cognitively normal and the 26th was judged to have mild cognitive impairment. Discussion: In successive admissions of patients over the age of 60 in a primary care practice, the 7 Minute Screen showed a PPV of 91% and an NPV of 96% in identifying patients who were subsequently identified with AD or other dementing disorder. These data suggest that this may be a useful instrument in identifying patients who should undergo diagnostic evaluation for AD and other dementing disorders. Additionally, extrapolation from the data in this practice suggests that there may be between 75 and 100 AD patients in the typical primary care practice, many of whom may not be diagnosed.