The Mini Mental State Examination (MMSE), a widely used tool for first-line assessment of cognitive function, cannot be fully administered to persons with severe visual or upper extremity impairments. This cross-sectional study, which was performed in a sample of patients admitted to the outpatient clinic of the INRCA Geriatric Department "I Fraticini" (Firenze, Italy) and their relatives, evaluated whether the Telephone Interview for Cognitive Status (TICS), a test originally created for telephone screening of cognitive impairment, is a valid alternative for assessment of cognition in persons who cannot provide valid responses to all the MMSE items. Fifteen subjects in each of seven MMSE strata (9-11, 12-14, 15-17, 18-20, 21-23, 24-26, 27-30) were consecutively selected. Inclusion criteria were: age > or = 60 years; Italian as the main language; education > or = 3 years; medical stability; and ability to complete the MMSE. At admission, participants were examined for dementia by an expert clinician according to the DSM IV criteria. Italian versions of the TICS and the MMSE were administered face-to-face by two trained interviewers. The TICS was readministered face-to-face one week later by the same interviewer that had previously administered the test. The TICS showed good test-retest reproducibility. Assessments taken one week apart differed on average by 0.52 +/- 1.9 points (p < 0.01), indicating a systematic improvement. The TICS explained 96% of the variance in the MMSE, and this association was independent of age, gender and education. More than 90% of the cases "screening positive" (i.e., scoring below a certain cut-off) by the TICS, were also identified as "screening positive" when comparable MMSE cut-off scores were used. Using comparable cut-off scores, the sensitivity and specificity of TICS and MMSE in detecting a standard clinical diagnosis of dementia largely overlapped. In persons who cannot be evaluated with the MMSE in its full form, the face-to-face administration of the TICS is a reliable and valid alternative.