Objective: To investigate the associations between linguistic parameters in spontaneous speech at baseline and cognitive impairment and frailty nine years later.
Methods: A prospective analysis was carried out on data of the Frailty in Brazilian Older People Study (FIBRA) Study, a population-based study on frailty. From a probabilistic sample of 384 individuals aged 65 and older at baseline (2008-2009), 124 aged 73 years and older at follow-up were selected, as they had scored above the cutoff values of cognitive screening for dementia adjusted by years of schooling at baseline and had answered to the question What is healthy aging and had no frailty at baseline. Verbal responses were submitted to content analysis and had its ideas and words counted. Number of ideas corresponded to the frequency of meaning categories and number of words to all identified significant textual elements in the text constituted by the sample answers to that question.
Results: Multivariate logistic regression analyses, controlling for the effects of age, sex, and education, showed that individuals with a high number of ideas at baseline had lower chance of having cognitive impairment (OR = 0.39; 95% CI 0.22 - 0.69) and frailty (OR 0.66; 95% CI 0.44 - 0.99) nine years later than those with low number of ideas.
Conclusions: Higher number of ideas, but not number of words, in spontaneous speech seems to be associated to a more positive prognosis in mental and physical health nine years later. Linguistic markers may be used to predict cognitive impairment and frailty in older individuals.
Keywords: Verbal behavior; cognition; frailty; longitudinal studies; older adults.