How to do binary endpoint analyses for population pharmacokinetics and exposure-response reports

It is recommended to only include adverse events in the report when adverse events have an incidence of >10%. Furthermore, it is recommended to do log-rank p-value tests. If there are any log-rank tests with significance p-value (p-value < 0.05), then Cox proportional hazard models should be conducted. When Cox proportional hazard models are conducted it is recommended to do univariate and multivariate analyses.