Synthesis of a damage-mitigating control law requires additional information on damage states beyond what is needed for the design of a conventional output feedback controller. In this context, this brief paper establishes the necessity of a fatigue damage model that must account for the impact of variable-amplitude stress excitation on crack growth rate (e.g., crack retardation and sequence effects). It is shown that predicted structural durability and the damage-mitigating controller design could be grossly inaccurate if the fatigue crack damage model does not represent the effects of variable-amplitude cyclic stress. A specific example is given based on the design of output-feedback damage-mitigating controllers for a reusable rocket engine that was reported in an earlier publication. Simulation results are presented in this paper to compare the predicted structural durability and closed-loop performance of the rocket engine under the same controllers for two different damage models, with and without consideration of the effects of variable-amplitude stress.