Table 3.

Power calculations assuming 50 clusters allocated in a 1:1 ratio between treatment and controls to achieve power of 80%

TIRS efficacy	No. events	Total effective sample size	No. per cluster, unadjusted	No. per cluster, adjusted*	Total sample size, adjusted*
70%	28	1038	74	92	4600
75%	22	870	43	54	2700
80%	18	734	30	37	1850
90%	12	534	17	21	1050

*Adjusted for a 20% loss to follow-up