. 2021 Sep;62(3):460–470. doi: 10.1016/j.jpainsymman.2021.01.138

Table 3.

Multivariate Logistic Regression: Factors Associated with Shortages

Multivariate logistic Regression: factors associated with PPE shortages
Variables	Odds Ratio	CI_lower	CI_upper	P value
Unit management type
Public	Ref
Charitable	3.07	1.81	5.20	P < .001
Other	0.97	0.46	2.05	P = .95
Inpatient palliative care unit
No	Ref
Yes	2.34	1.46	3.75	P < .001
Hospital palliative care team
No	Ref
Yes	0.51	0.32	0.82	P = .005
Multivariate logistic regression: factors associated with staff shortages
Variables	Odds ratio	CI_lower	CI_upper	P value
Country
UK	Ref
Rest of Europe	0.63	0.37	1.07	.09
Rest of the world	0.44	0.26	0.76	.003
Multivariate logistic regression: factors associated with other equipment shortages
Variables	Odds ratio	CI_lower	CI_upper	P value
Inpatient palliative care unit
No	Ref
Yes	0.35	0.18	0.65	.001
Country
UK	Ref
Rest of Europe	0.15	0.04	0.51	.002
Rest of the world	0.46	0.20	1.08	.07
Level of busyness
About the same	Ref
A lot more busy	10.81	3.10	37.71	<.001
Slightly more busy	5.41	1.48	19.76	.01
Slightly less busy	3.24	0.81	12.89	.10
Much less busy	1.32	0.21	8.34	.77

Note: No multivariate logistic regression analysis could be carried out for medicines shortages because only one variable (inpatient palliative care unit) was relevant from the univariable analysis.

Ref: Reference category, CI = confidence interval.