Mind the gap: covariate constrained randomisation can protect against substantial power loss in parallel cluster randomised trials

Table 1 Illustrative case study randomising ten emergency departments (clusters) with three cluster-level binary covariates

	Emergency department (cluster)
Characteristic	1	2	3	4	5	6	7	8	9	10
Large patient volume	Yes	No	Yes	Yes	No	Yes	No	No	No	No
Dedicated mental health team	Yes	No	Yes	Yes	Yes	No	No	No	No	Yes
Access to urgent mental health follow-up	Yes	No	No	No	Yes	Yes	Yes	No	No	No
	Strata
	1	2	3	4	5	6	7	8
Large patient volume	Yes	Yes	Yes	Yes	No	No	No	No
Dedicated mental health team	Yes	Yes	No	No	Yes	Yes	No	No
Access to urgent mental health follow-up	Yes	No	Yes	No	Yes	No	Yes	No
Number of clusters	1	2	1	0	1	1	1	3

ISSN: 1471-2288