ABSTRACT
In personalized medicine, continuous biomarker values are often dichotomized to classify patients into target and nontarget populations. For example, baseline hemoglobin A1c (A1C) and Positive and Negative Syndrome Scale (PANSS) measurements are potential biomarkers in diabetes and psychiatry, respectively. For Alzheimer's Disease (AD), currently Mild versus Moderate classification is used as a potential binary biomarker for treatment effect prediction. However, Normal, Mild, Moderate, and Severe staging of AD is a discretization of the mini-mental state examination (MMSE), which is measured on a scale of 0–30. Therefore, MMSE is potentially a continuously valued biomarker for AD treatment. For such scenarios, we provide simultaneous confidence intervals for efficacy corresponding to all candidate thresholds. Our method allows for rigorous and flexible decision making, taking into consideration medical impact based on both the size of the target population and efficacy in the target population.