Treatment of Middle East respiratory syndrome with a combination of lopinavir/ritonavir and interferon-ß1b (MIRACLE trial): statistical analysis plan for a recursive two-stage group sequential randomized controlled trial

Arabi, Yaseen M.; Asiri, Ayed Y.; Assiri, Abdullah M.; Aziz Jokhdar, Hani A.; Alothman, Adel; Balkhy, Hanan H.; AlJohani, Sameera; Al Harbi, Shmeylan; Kojan, Suleiman; Al Jeraisy, Majed; Deeb, Ahmad M.; Memish, Ziad A.; Ghazal, Sameeh; Al Faraj, Sarah; Al-Hameed, Fahad; AlSaedi, Asim; Mandourah, Yasser; Al Mekhlafi, Ghaleb A.; Sherbeeni, Nisreen Murad; Elzein, Fatehi Elnour; Almotairi, Abdullah; Al Bshabshe, Ali; Kharaba, Ayman; Jose, Jesna; Al Harthy, Abdulrahman; Al Sulaiman, Mohammed; Mady, Ahmed; Fowler, Robert A.; Hayden, Frederick G.; Al-Dawood, Abdulaziz; Abdelzaher, Mohamed; Bajhmom, Wail; Hussein, Mohamed A.

Here α k, 2 is the precalculated rejection boundary for efficacy at the second stage of the two-stage design at the k th stage, and p k, 1 is the raw table probability corresponding to the Z ki statistic. At the first interim analysis, should the data suggest that another stage of the twostage steps is required, we will recalculate the conditional error and new boundaries will be calculated for K = 2. Let β k + 1, 1 , α k + 1, 1 be the rejection boundaries for futility and efficacy for the first (i = 1) of the twostage step of the k th + 1 stage. Then the conditional error is: Efficacy stopping boundary (α 2 ) 0.2250 Stop trial for efficacy at the second stage or recalculate based on conditional power at first interim analysis α1 is the maximum probability threshold under which the trial will be terminated early for efficacy. β1 is the maximum probability threshold above which the trial will be terminated for futility. α2 is the maximum probability threshold (the sum of the stage-wise p-values), above which the study will be declared as met its endpoint