Frailty status changes are associated with healthcare utilization and subsequent mortality in the elderly population

Table 3 Medical utilization in 2 months at 1-year follow-up among the elderly with different baseline frailty status and change of frailty status

	Total n	Number of outpatient visit			Hospitalization admission	Emergency room utilization
		Total visits	Non-rehabilitation	Rehabilitation	Hospitalization admission	Emergency room utilization
		Geometric mean ± SD	Geometric mean ± SD	Geometric mean ± SD	n (%)	n (%)
Baseline frailty status*
Frail	37	1.1 ± 7.6 ^a	0.9 ± 6.5 ^a	0.0 ± 8.6	1 (2.7%)	1 (2.7%)
Pre-frail	232	0.6 ± 8.4^a	0.6 ± 8.3 ^a	0.0 ± 2.2	5 (2.2%)	1 (0.4%)
Robust	279	0.4 ± 9.4	0.4 ± 9.3	0.0 ± 2.1	0 (0.0%)	3 (1.1%)
p-value⁺		< 0.001	< 0.001	0.281	0.977	0.427
Change of frailty status^#
Deterioration	73	0.7 ± 7.1	0.7 ± 7.1 ^c	0.0 ± 1.0	1 (1.4%)	0 (0.0%)
Unchanged pre-frail or frail	133	0.8 ± 7.5	0.8 ± 7.0 ^c	0.0 ± 3.6	3 (2.3%)	1 (0.8%)
Unchanged robustness	223	0.4 ± 9.8 ^b	0.3 ± 9.6	0.0 ± 2.3	0 (0.0%)	3 (1.4%)
Improvement	119	0.5 ± 9.6 ^b	0.5 ± 9.4	0.0 ± 2.6	2 (1.7%)	1 (0.8%)
p-value⁺		< 0.001	< 0.001	0.197	0.972	0.952

The numbers in bold indicate statistically significant p-values
* Fried et al. proposed the definition of frailty status with the following components: shrinking, weakness, poor endurance and energy, slowness, and low physical activity level
#Changes in frailty status during 1-year period with the following categories: improvement of frailty status, no change between baseline and follow-up, and deterioration for frailty status
⁺p-values were calculated using univariate negative binomial regression models for the number of outpatient visits, and univariate logistic regression models for the hospitalization and emergency use
^a Statistically significant compared with the robust group at baseline using pairwise comparisons with Bonferroni correction (p < 0.05/3 comparisons)
^b Statistically significant compared with unchanged unchanged pre-frail or frail group at 1-year follow-up using pairwise comparisons with Bonferroni correction (p < 0.05/6 comparisons)
^c Statistically significant compared with unchanged unchanged robustness group at 1-year follow-up using pairwise comparisons with Bonferroni correction (p < 0.05/6 comparisons)

ISSN: 1471-2458