Cancer vaccines as a class of anti-cancer interventions have properties that must be taken into account when designing cancer vaccine clinical trials and characterizing the outcomes from these trials. These properties, as contrasted to other classes of anti-cancer agents, include generally lower toxic impact, weak or absent ability to shrink tumors, and the possibility of delay of effect. Lower toxic impact changes the benefit to harm ratio which means that smaller effect sizes are more meaningful than for most other classes of anti-cancer interventions. Weak or absent ability to shrink tumors means that effects intermediate to survival may not be observable. The property with perhaps the greatest impact, however, is the delay of effect. There is building evidence of the reality of delay of effect from completed trials of cancer vaccines. Failure to take into account the possibility of delay of effect when computing trial size will result in an underpowered trial if the delay of effect is present and if standard methods are used to compute estimates of effect size for a trial with delayed effect then the effect size estimate will not correctly characterize the effect of the vaccine. Delay of effect is a deviation from the usual proportional hazards statistical model typically used in cancer trials. This presentation will discuss a statistical model where the delay of effect is taken into account as a delay of proportional hazards. The advantage of this model is that it provides for the estimation of the magnitude of ultimate effect that is analogous to the hazard ratio estimated under proportional hazards. This presentation will describe this model and then illustrate its use to compute adequately powered trials and correctly estimate effect magnitudes when there is delay of effect consistent with this model.