For each composite component, the user is prompted for the anticipation of: the probability of having the event (or an interval in which this probability takes values); the effect relevant to detect (measured either by the odds ratio, risk ratio or risk difference). The correlation between the components is also required.

For the sample size calculation, the user should provide: the type of variance estimate to be used (unpooled or pooled variance); the type I error; and the desired power.

- P1, the probability of successes for the Endpoint1 in Control Group
- P2, the probability of successes for the Endpoint2 in Control Group
- D1, the effect on the Endpoint1
- D2, the effect on the Endpoint2
- pooled or unpooled variance estimate
- alpha, the significance level (value between 0 and 1)
- 1-b, the desired power for sample size estimation (value between 0 and 1)
- rho, correlation between Endpoint1 and Endpoint2. In case that this value is partially or totally unknown, a degree of strengh could be required.

- Sample Size for each composite component
- Sample Size for the composite endpoint given the correlation value
- Sample Size for the composite endpoint given the correlation category

- A new approach for sizing trials with composite binary endpoints using anticipated marginal values and accounting for the correlation between components. M. Bofill Roig, G. Gómez Melis.

Statistics in Medicine. doi: 10.1002/sim.8092.