Determination of functional prognosis in hospitalized patients following an intensive care admission

Table 3 Receiver-operating characteristic curve analysis for significant factors for predicting functional outcomes

Functional outcomes		ΔKATZ					ΔMRC
		AUC	Cut-off	Sensitivity	Specificity	P (Sig.)	AUC	Cut-off	Sensitivity	Specificity	P (Sig.)
ICU	Age, yr	0.664	78.5	41%	42%	0.001	1	1	1	1	1
	SAPS3, %	1	1	1	1	1	0.661	50.5	68%	52%	0.014
	KATZ at ICU admission, score	0.515	4.5	47%	58%	0.753	1	1	1	1	1
	MRC at ICU admission, score	0.592	45.5	57%	56%	0.053	0.653	37	44%	35%	0.001
Ward	Age, yr	0.712	81.5	32%	47%	< 0.001	1	1	1	1	1
	KATZ at ICU admission, score	0.590	4.5	35%	52%	0.035	1	1	1	1	1
	MRC at ICU admission, score	1	1	1	1	1	0.681	44.5	39%	42%	< 0.001
	KATZ at ward admission, score	0.746	2.5	63%	12%	< 0.001	1	1	1	1	1
	MRC at ward admission, score	0.721	47	69%	67%	< 0.001	0.553	43.3	54%	48%	0.200
Total	Age, yr	0.592	81.5	38%	52%	0.058	1	1	1	1	1
	Total length of stay, d	0.617	11.5	51%	72%	0.016	0.649	10.5	54%	73%	0.001
	KATZ at ICU admission, score	0.621	3.5	53%	28%	0.012	1	1	1	1	1
	KATZ at ward admission, score	0.515	4.5	38%	58%	0.750	1	1	1	1	1
	MRC at ward admission, score	1	1	1	1	1	0.696	37	51%	35%	< 0.001

¹Variables with no significant effect on the generalized linear model were not analyzed using ROC curves. KATZ: KATZ Index of Independence in Activities of Daily Living; MRC: Medical Research Council; SAPS3: Simplified acute physiology score; AUC: Area under the curve.