Comparison of the modified Frailty-Index based on laboratory tests and the Clinical Frailty Scale in predicting mortality among geriatric rehabilitation inpatients: RESORT.

Objective: To compare the associations of the FI-lab, modified (m)FI-lab and Clinical Frailty Scale (CFS) with one-year mortality.

Study Design: An observational longitudinal inception cohort of inpatients admitted to the geriatric rehabilitation wards in the Royal Melbourne Hospital, Victoria, Australia.

Main Outcome Measures: The measured ratio was defined as the proportion of measured laboratory tests to the total number of tests (n = 77). The FI-lab is the proportion of abnormal results to the total measured laboratory tests. The mFI-lab was calculated by dividing the FI-lab by the measured ratio. The measured ratio of laboratory tests, FI-lab, mFI-lab and CFS were assessed at admission to geriatric rehabilitation. Patients' mortality data were obtained from the Registry of Births, Deaths and Marriages Victoria and medical records.

Results: The total of 1819 inpatients had a median age of 83.3 [77.5-88.3] years and 56.5% were female. The median measured ratio, FI-lab, mFI-lab and CFS scores were 0.58 [0.47-0.70], 0.31 [0.23-0.38], 0.51 [0.38-0.69] and 6 (Abbasi et al., 2018Gill, Gahbauer, Allore & Han, 2006; Howlett et al., 2014;) respectively. The one-year mortality rate was 17.1%. The measured ratio was not associated with one-year mortality. Higher FI-lab (hazard ratio (HR)=1.180, 95%CI: 1.037-1.343), mFI-lab (HR=1.074, 95%CI: 1.030-1.119) and CFS scores (HR=1.350, 95%CI: 1.191-1.530) were associated with higher risk of one-year mortality. The area under the curve (AUC) of FI-lab, mFI-lab and CFS with one-year mortality were 0.581, 0.587 and 0.612 respectively.

Conclusion: The FI-lab, mFI-lab and CFS poorly predict mortality in geriatric rehabilitation inpatients despite the statistically significant associations shown.