Table 2.

Results of Cox proportional hazards regression analysis for mortality during 5‐year follow‐up

	Univariable			Age‐ and Sex‐Adjusted Modela			Fully Adjusted Model (n = 110)
Variable and coding	HR	95% CI	P value	HR	95% CI	P value	HR	95% CI	P value
Age (y)	1.10	1.06‐1.14	<.01				1.10	1.05‐1.16	.00
Sex (men vs women)	2.24	1.34‐3.74	<.01				2.29	1.17‐4.45	.02
Education (low vs middle and high)	1.66	0.95‐2.91	.08	1.84	1.01‐3.33	.05	0.96	0.50‐1.85	.91
FI (per 0.1)	1.87	1.48‐2.36	<.01	1.81	1.41‐2.31	<.01	1.78	1.06‐2.98	.03
Barthel (per point)	0.92	0.88‐0.96	<.01	0.94	0.89‐0.98	<.01	1.06	0.98‐1.13	.14
CIRS‐G (per point)	1.10	1.05‐1.15	<.01	1.09	1.04‐1.14	<.01	1.06	0.98‐1.15	.16
Diagnosis (NPS vs other diagnoses)	2.33	1.39‐3.89	<.01	1.87	1.11‐3.13	.02	1.98	1.00‐3.95	.05
CGI‐SA (per point)	1.49	1.02‐2.17	.04	1.38	0.95‐1.99	.09	1.51	0.99‐2.30	.06

Abbreviations: CGI‐SA: Clinical Global Impressions Scale of Severity on admission; CIRS‐G: Cumulative Illness Rating Scale Geriatrics; FI, frailty index; NPS, neuropsychiatric symptoms.

^aResults from partially adjusted models in which covariates were included separately, adjusted only for age and sex.