Predictive validity of the Brazilian version of the Tilburg Frailty Indicator for adverse health outcomes in older adults.

PURPOSE: This study aimed to determine the predictive value of the Brazilian Tilburg Frailty Indicator (TFI) for adverse health outcomes (falls, hospitalization, disability and death), in a follow-up period of twelve months. METHODS: This longitudinal study was carried out with a sample of people using primary health care services in Rio de Janeiro, Brazil. At baseline the sample consisted of 963 people aged 60 years and older. A subset of all respondents participated again one year later (n = 640, 66.6% response rate). We used the TFI, the Katz's scale for assessing ADL disability and the Lawton Scale for assessing IADL disability. Falls, hospitalization and death were also assessed using a questionnaire. RESULTS: The prevalence of frailty was 44.2% and the mean score of the TFI was 4.4 (SD = 3.0). There was a higher risk of loss in functional capacity in ADL (OR = 3.03, CI95% 1.45-6.29) and in IADL (OR = 1.51, CI95% 1.05-2.17), falls (OR = 2.08, CI95% 1.21-3.58), hospitalization (OR = 1.83, CI95% 1.10-3.06), and death (HR = 2.73, CI95% 1.04-7.19) for frail when compared to non-frail elderly, in the bivariate analyses. Controlling for the sociodemographic variables, the frailty domains together improved the prediction of hospitalization, falls and loss in functional capacity in ADL, but not loss in functional capacity in IADL. CONCLUSION: The TFI is a good predictor of adverse health outcomes among elderly users of primary care services in Brazil and appears an adequate and easy to administer tool for monitoring their health conditions.

Psychometric properties of the Brazilian version of the Tilburg frailty indicator (TFI).

This study aims to assess the psychometric properties of the Brazilian version of the TFI, an instrument that identifies frailty in elderly individuals. We interviewed 219 individuals aged 60 or older, living in the community. Individuals were predominantly female (52.5%) and mean age was 70.5 (±7.9) years. In order to assess test-retest reliability, 101 individuals were re-interviewed by the same observer within seven to ten days after the first interview. The internal consistency of the instrument was assessed using Cronbach's alpha. To assess construct validity, we used established alternative measures for the items that constitute the TFI, such as: body mass index (BMI), timed up and go (TUG) test, whisper test, Snellen test, upper extremity strength clinical test and mini-mental state examination (MMSE). The test-retest reliability showed high percent agreement for all the items of the instrument, with values ranging from 63% to 100%. Test-retest reliabilities were good (total TFI score r=0.88; physical domain r=0.88; psychological domain r=0.88; and social domain r=0.67). Internal consistency reliability of the Brazilian version was satisfactory (Cronbach's alpha=0.78). The correlations between TFI items and their corresponding measures were consistent except for one item (related to "ability to deal with problems"), demonstrating both convergent and divergent construct validity of the TFI and its items. After the completion of all stages of transcultural adaptation, the Brazilian version of the TFI proved to be well suited for assessing frailty in the elderly population of Brazil.