Is statistical power necessary to quantify the impact on study conclusions of moving from two to one assessments