A frailty index derived from a standardized comprehensive geriatric assessment predicts mortality and aged residential care admission

Abstract

BackgroundFrailty in older adults is a condition characterised by a loss or reduction in physiological reserve resulting in increased clinical vulnerability. However, evidence suggests that frailty may be modifiable, and identifying frail older people could help better target specific health care interventions and services.MethodsThis was a regional longitudinal study to develop a frailty index for older adults living in Canterbury New Zealand. Participants included 5586 community dwelling older people that had an interRAI Minimum Data Set (MDS-HC) Home Care assessment completed between 2008 and 2012. The outcome measures were mortality and entry into aged residential care (ARC), after five years.ResultsParticipants were aged between 65 and 101 (mean age was 82 years). The five-year mortality rate, including those who entered ARC, for this cohort was 67.1% (n = 3747). The relationship between the frailty index and both mortality and entry into ARC was significant (P < 0.001). At five years, 25.1% (n = 98) of people with a baseline frailty of < 0.1 had died compared with 28.2% (n = 22) of those with a frailty index of ≥0.5 (FS 5). Furthermore, 43.7% (n = 171) of people with a frailty index of < 0.1 were still living at home compared to 2.6% (n = 2) of those with a frailty index of ≥0.5.ConclusionA frailty index was created that predicts mortality, and admission into ARC. This index could help healthcare professionals and clinicians identify older people at risk of health decline and mortality, so that appropriate services and interventions may be put in place.Electronic supplementary materialThe online version of this article (10.1186/s12877-018-1016-8) contains supplementary material, which is available to authorized users.

Derivation of a Frailty Index from the interRAI-HC to Assess Frailty among Older Adults Receiving Home Care and Assistance (the “fraXity” Study)

“…Aside from the participants' characteristics, with individuals with few care needs deliberately being targeted, FI values are probably further underestimated due to methodological biases related to the voluntary participation in this study.Indeed, voluntary participation has long been known as a source of bias[69], either because people who volunteer are in condition than their peers (volunteer bias) and/or because people who do not participate are in worse condition than their peers (non-response bias). We cannot exclude that the convenient sampling method further contributed to exacerbate this bias.Notwithstanding, when looking at the FI mean value reported for the care group of 0.27 (SD = 0.10), it is not significantly different from those reported in homecare samples from Switzerland (N = 3714, m = 0.24, SD = 0.13[44], p > 0.05) and New Zealand (N = 5586, m = 0.27, SD = 0.12[43], p > 0.05). This observation suggests that the proposed computation algorithms lead to comparable results to those reported for larger clinical samples, which are usually routinely assessed, thus drastically reducing the risk of this aforementioned bias.…”

“…The FI-MDS derivation methodology benefits from convincing evidence stemming from studies that used the interRAI-Acute Care [38,39], interRAI ED-Contact Assessment [40], interRAI-Home Care [29,[41][42][43], and interRAI-Nursing Home [45]. [38,40,43,44], large-scale health studies [23,46,68], or even survey data [25]. We report a small but significant positive age effect (B = 0.003, 95% CI = (0.001-0.005)) on the FI, as previously documented in community [46] and homecare samples [43,44].…”

“…[38,40,43,44], large-scale health studies [23,46,68], or even survey data [25]. We report a small but significant positive age effect (B = 0.003, 95% CI = (0.001-0.005)) on the FI, as previously documented in community [46] and homecare samples [43,44]. work remains needed to refine the FI's psychometric properties, notably to assess its predictive validity with respect to adverse health outcomes and its reliability across measurement occasions.…”

“…A cut point of FI > 0.21 is preferred for studies conducted in community-based population [27][28][29][30][31][32]. A FI can be computed from routinely collected health data, whether from CGAs [26,33,34], primary care electronic health records [35] or minimum datasets [MDS] gathered with Resident Assessment Instruments (RAI) [36,37] designed for acute care [38][39][40], home care [29,[41][42][43][44] and nursing homes [45]. The rationale underlying deriving the FI from clinical datasets is to provide a measure of frailty that demonstrates substantial agreement across studies [25,46] and high predictive validity of undesirable health outcomes [47,48].…”

<p>Prediction of chemotherapy adverse reactions and mortality in older patients with primary lung cancer through frailty index based on routine laboratory data</p>

“…Finally, the choice of the FI-LAB cutoff used to define frailty refers to the study from Harvey et al,10 in which the frailty index is constructed based on comprehensive geriatric assessments. Nonetheless, the cutoffs for the frailty index are controversial now10,31,32 and regardless of the FI-LAB cutoffs, the risk of mortality also increased by per 0.01 or per SD in FI-LAB, demonstrating its ability to predict long-term mortality.…”