Four different frailty models predict health outcomes in older patients with stable chronic obstructive pulmonary disease

Table 3 Results of receiver operating characteristic curve analysis showing the ability of Fried Frailty Phenotype, Short Physical Performance Battery, Clinical Frailty Scale and Frailty Index of Accumulative Deficits for predicting 1-year all-cause mortality in elderly COPD patients

	Frail	Non-frail	Sensitivity	Specificity	PPV	NPV	PLR	AUC (95% CI)	P-value	AUC Model^a (95% CI)	P-value
FFP	25/154	3/148	0.89	0.53	0.16	0.98	1.89	0.76 (0.69–0.83)	Ref	0.83 (0.76–0.89)	Ref
CFS	27/194	1/108	0.96	0.39	0.14	0.99	1.57	0.70 (0.61–0.78)	0.114	0.82 (0.76–0.88)	0.835
SPPB	26/177	2/125	0.93	0.45	0.15	0.98	1.69	0.80 (0.72–0.87)	0.371	0.82 (0.76–0.88)	0.779
FI-CD	27/180	1/122	0.96	0.44	0.15	0.99	1.71	0.76 (0.68–0.83)	0.843	0.82 (0.77–0.88)	0.908

AUC area under the curve, PLR positive likelihood ratio (how much more likely is a person who died during 1-year to be classified as frail than those who survived), NPV negative predictive value (probability that a person has no outcome when identified as non-frail), PPV positive predictive value (probability that a person has the outcome when identified as frail)
^a Adjusted model: frailty instrument plus age, sex, CCI, medication, GOLD severity, moderate-to-severe exacerbation history, and CAT

ISSN: 1471-2318