Fig. 3.
Trial sequential analysis (TSA) for mortality in randomized controlled trials: one-sided boundary, incidence of 8.59% in the control arm, incidence of 5.99% in the treatment arm, low-bias estimated relative-risk reduction of 80%, α of 5%, and power of 80% were set. There is an estimated required information sample size of 2739 randomized patients, which was not achieved. The boundaries for futility are crossed