Table 4.

Multivariate logistic regression analysis including sociodemographic and circumstantial variables associated with inpatient psychiatric admission

	OR	95% CI		p-Value
		LL	UL
Gender	1.06	0.90	1.26	0.480
Age	1.01	1.00	1.01	0.100
Aggressive towards people	2.20	1.40	3.43	< 0.001**
Mechanical restraint	2.29	1.12	4.70	0.023*
Previous psychiatric treatment	1.88	1.54	2.29	< 0.001**
Means of presentation (reference = walk-in)				< 0.001**
By ambulance	1.48	1.23	1.80	< 0.001**
By police escort	2.18	1.54	3.11	< 0.001**
Transfer	6.20	3.17	12.14	< 0.001**
Suicide attempt prior to presentation (reference = no suicide attempt)				< 0.001**
Via intoxication	9.74	2.96	32.05	< 0.001**
Via self-injury	13.66	1.80	103.75	0.011*
Via other methods (e.g. strangulation, combination of methods)	25.30	3.36	190.81	0.002*
Day of the week (reference = Sunday)				0.022*
Tuesday	1.48	1.06	2.07	0.021*
Wednesday	1.49	1.07	2.07	0.017*
Retired (reference = (self-)employed)	1.89	1.29	2.78	0.001*
Living alone (reference = with others)	1.54	1.20	1.97	< 0.001**

OR: odds ratio, 95% CI: 95% confidence interval, LL: lower Limit, UL: upper limit, transfer: transfer from another hospital or another medical discipline within Hannover Medical School; * statistically significant, ** highly statistically significant

Omnibus test: p < 0.001, Nagelkerkes R-Square = 0.162, correctly predicted: 62.6%. Variables included in multivariate analysis after univariate testing: aggressive prior to presentation, aggressive towards people, aggressive towards objects, mechanical restraint, previous psychiatric treatment, means of presentation, intoxication, suicide attempt prior to presentation, day of the week, work status, living arrangement