Table 2.

Sample size assuming 5 % level of significance, 90 % power, 50 % relative risk reduction

Incidence of primary outcome in ‘no decision-support’ group	Incidence of primary outcome in ‘decision-support’ group	Relative risk	Total sample size required
3 per 1000	1.5 per 1000	0.5	46,000
4 per 1000	2 per 1000	0.5	34,000
5 per 1000	2.5 per 1000	0.5	27,000
6 per 1000	3 per 1000	0.5	22,000