Concordances and differences between a unidimensional and multidimensional assessment of frailty: a cross-sectional study

Table 2 The CFAI mean scores (total and per domain) according to the Fried Phenotype distribution

	N	Not-frail (mean ± SD)	Fried Phenotype	Frail (mean ± SD)	P-value
	N	Not-frail (mean ± SD)	Pre-frail (mean ± SD)	Frail (mean ± SD)	P-value
CFAI	193	19.35 ± 10.88¹	23.84 ± 11.61²	41.05 ± 14.28¹²	p < 0.000
CFAI Physical	195	6.25 ± 13.11³⁴	22.07 ± 30.61³⁵	61.84 ± 32.22⁴⁵	p < 0.000
CFAI Psychological	193	14.96 ± 16.32⁶	16.46 ± 15.52⁷	36.95 ± 22.92⁶⁷	p < 0.000
CFAI Social	195	46.61 ± 18.70	45.38 ± 18.04	51.21 ± 20.48	ns
CFAI Environmental	195	9.57 ± 13.16	10.99 ± 12.23	14.21 ± 14.64	ns

Anova test. According to the Levene’s Statistic, the variance of CFAI Physical and CFAI psychological were not equal instead the Welch test and Brown-Forsythe test used to determine the p-value. As post-hoc, the Tukey HSD test was conducted to find differences in mean between pairs (see the superscripts). Superscripts with the same number indicate a significant mean difference between two pairs of groups. The CFAI and its domains are a continuous scale (0–100). For the psychological domain, there were missing data for two participants and, for the Fried Phenotype one participant had missing data. ns = non-significant

ISSN: 1471-2318