An approach to trial design and analysis in the era of non-proportional hazards of the treatment effect