Frailty and its prediction of disability and health care utilization: the added value of interviews and physical measures following a self-report questionnaire.

AIMS: To establish whether the prediction of the adverse outcomes disability and six indicators of health care utilization one and two years later by the three frailty domains (physical, psychological, social) of the Tilburg Frailty Indicator (TFI) is improved by adding interview and physical measures of frailty.

MATERIALS AND METHODS: A representative sample of 245 Dutch community-dwelling persons aged 75 years and older (response rate 53%) participated in 2008, one year later in 2009 (n=179, 73%) and again two years later in 2010 (n=141, 58%). Frailty was assessed with the TFI, an easy to administer self-report measure. Disability was measured using the Groningen Activity Restriction Scale (GARS). Indicators of health care utilization were: visit to a general practitioner (gp), contacts with health care professionals (hcps), hospital admission, receiving personal care, receiving nursing care, and receiving informal care.

RESULTS: After controlling for background characteristics, the TFI predicted disability and the indicators of health care utilization. Interviews and physical measures of frailty improved the prediction of disability. The Hospital Anxiety and Depression Scale (HADS-A) improved the prediction of contacts with hcps, but the interview and physical measures of frailty did not improve the predictions of the other indicators of health care utilization.

CONCLUSIONS: Assessment by the self-report TFI is sufficient for predicting six indicators of health care utilization, but for predicting disability the use of both the TFI and the Timed Up & Go (TUG) test is recommended. It is advisable assessing all three frailty domains when examining frailty and its prediction of adverse outcomes.