Table 3.

Multivariable mixed ordinal logistic regression model for the duration of stroke unit stay (≤ 24 h, > 24- ≤ 72 h, or > 72 h) in 3355 patients (random intercept models, random intercept for the 38 centres)

	Duration of stroke unit stay (≤ 24 h, > 24-72 h, > 72 h)
	Adjusted OR	95% CI	P-value
Sex
Male	1
Female	0.88	0.75–1.03	0.112
Index event
TIA	1
Stroke	1.72	1.45–2.05	< 0.001
Stroke severity
NIHSS < 5 points	1
NIHSS ≥ 5 points	1.60	1.30–1.96	< 0.001
Age category
≥ 75 years	1
60–74 years	1.04	0.87–1.26	0.654
< 60 years	0.84	0.68–1.02	0.078
Type of stroke centre
PCS	1
UCSC	1.95	0.63–6.07	0.248
nUCSC	0.72	0.24–2.14	0.555

The regression model was additionally adjusted for duration of the in-hospital stay

OR odds ratio, CI confidence interval, NIHSS National Institutes of Health Scale, TIA transient ischaemic attack, PCS primary stroke centre, UCSC university-based comprehensive stroke centre, nUCSC non university-based comprehensive stroke centre