Table 4.

Logistic models predicting the likelihood of home discharge.

		Model 1			Model 2			Model 3			Model 4
		pseudo R2 = .15 p < .001			pseudo R2 = .39 p < .001			pseudo R2 = .51 p < .001			pseudo R2 = .32 p < .001
		OR	95% CI	p	OR	95% CI	p	OR	95% CI	p	OR	95% CI	p
KF-NAP at admission		.89	[.84, .95]	<.001	1.01	[.92, 1.11]	.786	.96	[.85, 1.08]	.469
FIM at admission					1.04	[.96, 1.13]	.353	1.09	[.97, 1.22]	.131
FIM at discharge					1.00	[.92, 1.09]	.964	.97	[.86, 1.09]	.621
FIM improvement rate					4.83	[.95, 24.53]	.058	13.02	[1.23, 137.36]	.033	5.25	[2.35, 11.73]	<.001
Age					1.01	[.95, 1.08]	.723	1.03	[.93, 1.13]	.570
Marital status (ref = single)	Married				6.78	[.92, 49.75]	.060	4.69	[.34, 65.59]	.251
	Widowed				6.80	[.69, 67.37]	.101	1.63	[.08, 35.20]	.757
	Divorced				1.19	[.11, 12.90]	.888	.42	[.02, 9.52]	.585
Female								1.02	[.18, 5.75]	.986
White race								1.21	[.20, 7.21]	.833
Hispanic ethnicity								--	--	--
Years of education								1.27	[.87, 1.84]	.214
Employment (ref = unemployed)	Employed							.63	[.01, 37.71]	.823
	Retired							.90	[.02, 35.23]	.957
Annual income level								.40	[.17, .95]	.037	.72	[.46, 1.12]	.141
Conley Scale								1.39	[.91, 2.11]	.129
Number of falls								4.16	[.50, 34.55]	.187
Residual		7.61	[3.56, 16.26]	< .001	.002	[<.001, 4.89]	.120	<.001	[<.001, 7.65]	.105	.54	[.16, 1.80]	.313
Model comparison (Likelihood-ratio test)					Model 2 vs. 1: χ2 (7) = 25.25, p < .001			Model 3 vs. 1: χ2 (15) = 36.82, p = .001			Model 4 vs. 1: χ2 (1) = 17.82, p < .001
								Model 3 vs. 2: χ2 (8) = 10.97, p = .204			Model 4 vs. 2: χ2 (6) = 7.44, p = .282
											Model 4 vs. 3: χ2 (14) = 19.09, p = .162

Note: Hispanic ethnicity was omitted from Model 3 because all Hispanic participants returned home at IRF discharge.

Abbreviations: KF-NAP, Kessler Foundation Neglect Assessment Process; FIM, Functional Independence Measure; OR, odds ratio