Trials should assess overall survival to evaluate potential harm, with crossover limited to avoid confounding treatment ...