Influencing factors of Barthel index scores among the community-dwelling elderly in Hong Kong: a random intercept model

Table 3 Model comparison results of evaluation criteria

	Formula	Random Effects	Multi.R²	Adj.R²	AIC	BIC	-Loglik
LM	Model 1: BI ~ Demo	None	0.46	0.40	2519	2631	1229
	Model 2: BI ~ Demo + Disease		0.56	0.49	2481	2626	1201
	Model 3: BI ~ Demo + Disease + Health History		0.76	0.69	2360	2577	1120
	Formula	Random Effects	Cond.R²	Marg.R²	AIC	BIC	-Loglik
LME	Model 4: BI ~ Demo	ID	0.79	0.49	2433	2549	1185
	Model 5: BI ~ Demo + Disease		0.77	0.58	2432	2580	1175
	Model 6: BI ~ Demo + Disease + Health History		0.84	0.75	2337	2557	1107

Note: Demo represents 17 socio-demographic variables; Disease represented 9 disease conditions, Health History represented 20 extracted features from health history in EHRs. Multi.R², multiple R square. Adj.R², adjusted R square. Cond.R², Conditional R square. Marg.R², Marginal R square,

ISSN: 1471-2318