Table 4.

Linear regression analysis of the natural logarithm of length of stay

	Coefficient	95% CI	p value
Intercept	2.41	1.89–2.93	<0.001
Fixed effect
Age at admission, year	0.01*	0.01–0.02	<0.001
Sex, male	0.11	−0.05–0.28	0.183
Education level (junior high school = 0)
High school	−0.02	−0.23–0.19	0.848
Vocational school, college	0.14	−0.16–0.44	0.361
University	0.04	−0.23–0.31	0.771
Unidentified	−0.16	−0.49–0.17	0.337
Working status before admission (not engaged in work = 0)
Engaged in work	−0.15	−0.33–0.02	0.083
Unidentified	−0.23	−0.66–0.19	0.285
Primary diagnosis (other = 0)
Schizophrenic	0.40*	0.20–0.59	<0.001
Affective disorder	0.18	−0.05–0.41	0.129
Compulsory admission	0.47*	0.27–0.67	<0.001
Total number of psychiatric admissions	−0.01	−0.03–0.01	0.267
Physical comorbidity	0.04	−0.17–0.24	0.714
Current place of residence (cohabitation with family = 0)
Lives alone at own home	−0.18	−0.42–0.05	0.130
Residential care facility	0.01	−0.35–0.37	0.921
Readmitted to hospital from home	−0.18	−0.59–0.22	0.372
Dead shortly after discharge	−0.49	−1.72–0.74	0.435
Place of residence unidentified	−0.21	−0.56–0.15	0.258
Social functioning at discharge	−0.03	−0.08–0.02	0.255
Follow-up outpatient service after discharge (follow-up outpatient service by the hospital = 0)
Follow-up outpatient service of other hospital	−0.35*	−0.54 to −0.17	<0.001
No follow-up/unidentified	−1.17*	−1.63 to −0.72	<0.001
Received support on community care coordination	0.61*	0.43–0.80	<0.001
Random effect
Residual	0.73	0.63–0.84
Intercept: hospital	0.003	10−11 × 0.28 − 23,291.62
Intra-class correlation coefficient (ICC)	<0.001
Fitness of model
χ2 (23)	178.61		<0.001
Log likelihood	−566.48
Akaike’s information criterion (AIC)	1,182.97

* Significant at p < 0.05.