Frailty index of deficit accumulation and falls: data from the Global Longitudinal Study of Osteoporosis in Women (GLOW) Hamilton cohort

Table 4 Sensitivity analysis of relationship between frailty index and falls for the whole group and age subgroups

Sensitivity analysis	n (%)	Age-adjusted models			Fully-adjusted models
Sensitivity analysis	n (%)	OR (95% CI)^a	P-value	AUC^b	OR (95% CI)	P-value	AUC
Any falls at baseline and during the 3-year follow-up ^c
Baseline falls	1483 (37.51)	1.03 (1.03-1.04)	<0.001	0.61	1.04 (1.03-1.04)	<0.001	0.62
Women < 65 years	540 (39.16)	1.04 (1.03-1.05)	<0.001	0.64	1.04 (1.03-1.05)	<0.001	0.64
Women ≥ 65 years	943 (36.62)	1.03 (1.02-1.04)	<0.001	0.61	1.03 (1.03-1.04)	<0.001	0.62
New incident falls ^d	1204 (51.43)	1.02 (1.02-1.03)	<0.001	0.59	1.02 (1.02-1.03)	<0.001	0.60
Women < 65 years	384 (47.64)	1.02 (1.00-1.03)	0.015	0.56	1.02 (1.00-1.03)	0.018	0.58
Women ≥ 65 years	820 (53.42)	1.02 (1.02-1.03)	<0.001	0.61	1.03 (1.02-1.04)	<0.001	0.61
Recurrence of falls ^e	1135 (81.54)	1.04 (1.03-1.05)	<0.001	0.64	1.04 (1.03-1.06)	<0.001	0.64
Women < 65 years	419 (80.58)	1.05 (1.03-1.07)	<0.001	0.66	1.06 (1.03-1.08)	<0.001	0.68
Women ≥ 65 years	716 (82.11)	1.04 (1.02-1.05)	<0.001	0.63	1.04 (1.02-1.05)	<0.001	0.66
Analysis with multiple imputation for missing data ^f	1317 (33.05)^g	1.03 (1.03-1.04)	<0.001	0.61	1.02 (1.02-1.03)	<0.001	0.69
Women < 65 years	448 (32.35)^g	1.03 (1.02-1.04)	<0.001	0.60	1.02 (1.01-1.03)	0.004	0.72
Women ≥ 65 years	869 (33.42)^g	1.03 (1.03-1.04)	<0.001	0.62	1.02 (1.02-1.03)	<0.001	0.68

^aOR: odds ratio; CI: confidence interval; All statistics measured a change of 0.01 on the frailty index.
^bAUC: areas under the receiver operating characteristic curves.
^cAnalyses adjusted for age, BMI, smoking, drinking, and education.
^dDefined as women had any falls during the follow-up but without baseline falls.
^eDefined as women had any falls during the follow-up and with baseline falls.
^fFor any falls during the third year of follow-up; Analyses adjusted for age, BMI, smoking, drinking, education, and baseline falls.
^gNumbers obtained after 10 multiple imputation.

ISSN: 1471-2474