Short-term frailty index fluctuations in older adults: Noise or signal?

Abstract

Reversible short-term fluctuations in the frailty index (FI) are often thought of as representing only noise or error. Here, we assess (1) size and source of short-term FI fluctuations, (2) variation across socio-demographics, (3) association with chronic diseases, (4) correlation with age, frailty level, frailty change, and mortality, and (5) whether fluctuations reflect discrete health transitions. Nationwide, biweekly longitudinal data from 426 community-dwelling older adults (70+) were collected in the FRequent health Assessment In Later life (FRAIL70+) study using a measurement burst design (5,122 repeated observations, median of 13 repeated observations per person). We calculated the intraindividual standard deviation (iSD) of the FI and used location-scale mixed regression models. Mean iSD was 0.04 (SD=0.03). Fluctuations were driven foremost by cognitive problems, somatic symptoms, and limitations in instrumental and mobility-related activities of daily living. Short-term fluctuations correlated with higher FI levels (r=0.62), one-year FI change (r=0.26), and older age (+3% per year). Older adults who took to bed due to a health problem (+50%), those who had an overnight hospital stay (+50%), and those who died during follow-up (+44%) exhibited more FI fluctuations. Short-term FI fluctuations were neither small nor random. Instead, as older adults become frailer, their measured health also becomes more unstable; hence short-term fluctuations in overall health status can be seen as a concomitant phenomenon of the aging process. Researchers and clinicians should be aware of existence of reversible fluctuations in the FI over weeks and months and its consequences for frailty monitoring.