First page Back Continue Last page Overview Graphics

# Correspondence between 95% CI for the mortality ratio, R, in a **cluster**-randomised trial & equivalent numbers of deaths (treated vs control) in a large, evenly balanced, **individually** randomised trial:

## Correspondence between 95% CI for the mortality ratio, R, in a **cluster**-randomised trial & equivalent numbers of deaths (treated vs control) in a large, evenly balanced, **individually** randomised trial:

## 95% CI of (0.89-1.03) for R=0.96 in DEVTA
**would be equivalent to** 1411 vs 1470 deaths*

## * 95% CI corresponds to (1+R)k vs (1+1/R)k deaths, where k is the square of 3.92/loge(upper/lower limit); each limit is first rounded to 2 dp in these calculations