The Italian version of cognitive function instrument (CFI): reliability and validity in a cohort of healthy elderly

• Published in: Neurological sciences Vol. 39; no. 1; pp. 111 - 118
• Main Authors: Chipi, Elena, Frattini, Giulia, Eusebi, Paolo, Mollica, Anita, D’Andrea, Katia, Russo, Mirella, Bernardelli, Alice, Montanucci, Chiara, Luchetti, Elisa, Calabresi, Paolo, Parnetti, Lucilla
• Format: Journal Article
• Language: English
• Published: Milan Springer Milan 2018; Springer Nature B.V
• Subjects: Annual variations; Clinical trials; Cognitive ability; Geriatrics; Medicine; Medicine & Public Health; Neurology; Neuroradiology; Neurosciences; Neurosurgery; Older people; Original Article; Psychiatry; Subjective cognitive decline; Questionnaire; Alzheimer’s disease; Italian validation; CFI
• ISSN: 1590-1874; 1590-3478
• DOI: 10.1007/s10072-017-3150-z

The Alzheimer’s disease Cooperative Study (ADCS)-Cognitive Function Instrument (CFI) is a 14-item questionnaire administered to the subject and the referent, aimed at detecting early changes in cognitive and functional abilities in individuals without clinical impairment. It is used for monitoring annual variations in cognitive functioning in prevention trials. The aim of the present study was to validate the Italian version of the CFI. A consecutive series of 257 functionally independent subjects was recruited among relatives of patients or as volunteers. They were administered CFI and global cognition measurements: Mini-Mental Status Examination (MMSE) and Repeatable Battery for the Assessment of Neuropsychological Status (RBANS). The reliability and criterion validity were comparable to the original in both self- and partner-report. Similarly to what reported in the original version, we found a corrected item-total correlation ranging between 0.38 and 0.54 in self-report and between 0.33 and 0.64 in partner-report. Cronbach’s
α
was 0.77 (95% CI 0.72–0.83) in self-report and 0.78 (95% CI 0.73–0.84) in partner-report. Total partner- and self-report scores were significantly correlated (rS = 0.31,
p
 < 0.001). CFI self-report and CFI total-score (partner + self-report) were negatively correlated with MMSE (rS = − 0.15,
p
 = 0.022; rS = − 0.17,
p
 = 0.008) and RBANS (rS = − 0.22,
p
 < 0.001; rS = − 0.25,
p
 < 0.001). Analogous trends were found in the partner score, with a correlation with RBANS (rS = − 0.17,
p
 = 0.014) and MMSE (rS = − 0.11,
p
 = 0.071). Our results support the reliability and validity of the Italian version of CFI. In order to definitely propose the use of CFI for tracking longitudinal changes of cognitive and functional abilities in subjects without clinical impairment, data from the follow-up of this cohort are needed.