Table 2. Contributory factors for choking in logistic regression analysis.

				Univariable model			Multivariable model
		n	Choking (%)	Crude OR	95% CI	p value	Adjusted OR	95% CI	p value
Patient age
	5 years ≦ age < 65	13931	20(0.14)	ref.			ref.
	< 5 years	3963	10(0.25)	1.76	[0.82–3.76]	0.145	2.73	[1.24–6.02]	0.013
	≧ 65 years	17546	38(0.22)	1.51	[0.88–2.60]	0.139	1.48	[0.84–2.60]	0.171
Patient Status
	Cognitive impairment (ref. = none)	1228	5(0.41)	2.22	[0.89–5.52]	0.087	1.42	[0.55–3.68]	0.465
	Mental illness (ref. = none)	1268	7(0.55)	3.10	[1.42–6.80]	0.005	3.14	[1.39–7.08]	0.006
Occurring situation
	Others			ref.			ref.
	Intensive care unit	17038	53(0.31)	0.51	[0.11–2.25]	0.374	0.39	[0.08–1.74]	0.213
	Hospitalised in the general wards	4272	2(0.05)	3.39	[1.85–6.22]	0.033	3.13	[1.70–5.76]	<0.001

ref. = reference; OR = odds ratio; 95% CI = 95% confidence interval.