Table 4.

Power and sample size calculations (based on two wards and four periods)

Range of period effects	Delta	VW	VC	VP	NP	Power
−1.5, 1.5	1.50	0.025	0.025	4.0	24	0.651
−1.0, 1.0	1.50	0.025	0.025	4.0	24	0.670
−0.75, 0.75	1.50	0.025	0.025	4.0	24	0.680
−1.5, 1.5	1.50	0.025	0.025	4.0	30	0.714
−1.0, 1.0	1.50	0.025	0.025	4.0	30	0.741
−0.75, 0.75	1.50	0.025	0.025	4.0	30	0.744
−1.0, 1.0	1.80	0.025	0.025	4.0	24	0.797
−1.0, 1.0	1.65	0.025	0.025	4.0	30	0.803

Footnote: Power and sample size calculations were calculated for the F test (α = 0.05) for a difference between the intervention and control groups in Day-3 worst pain. Delta is the absolute value of the difference in the mean pain scores. Components of variance in Day-3 pain scores are fixed as follows: between wards (VW = 0.025), between cohorts of patients within the same ward (VC = 0.025) and between patients within a cohort within a ward (VP = 4.0). NP is the number of patients in a cohort. The power (1-β) is the probability that the null hypothesis, of no difference in the mean Day-3 worst pain scores between the control and intervention groups, is rejected when the true, but unknown, difference is delta, the components of variance are as given, there are two wards and four cohorts per ward managed contemporaneously in four time periods, and the F-test is conducted at the 5 % significance level (α = 0.05). In these scenarios, three different ranges (in equally spaced steps) for the effects of the four periods on day-3 worst pain scores are investigated