Skip to main content
. 2019 Dec 2;20:665. doi: 10.1186/s13063-019-3724-6

Fig. 2.

Fig. 2

Participants receive each treatment A, B, C and D; they are randomised to a treatment sequence according to a four-period four-treatment Williams square design: ABDC, BCAD, CDBA and DACB, balanced by ethnic group