Speaker: Anu Mishra, Graduate Student, UW Biostatistics
Abstract: Risk markers or scores are often used to prescribe an intervention. Estimating statistical measures of calibration and discrimination are necessary steps in risk score evaluation. Recently, it has been advocated that additionally reporting estimates of clinical utility is important in assessing novel biomarkers or risk prediction tools, when the tool will be used to prescribe intervention. In this talk, I will adopt decision analytic framework and address problems of individualized decision making, recalibration, and developing risk scores when the ultimate goal is to use the risk score for clinical decision-making. I will briefly review methods of estimating clinical utility from Bayesian and frequentist standpoints and draw connections between the two methods. When existing risk models are applied to new populations, issues of miscalibration can arise. I will present two methods for recalibration that account for the clinical context in which the risk model will be used. Finally, I’ll address the problem of developing risk score. I will present a non-parametric method for developing linear combinations of risk markers that maximize net benefit. In this work, we consider applications of prostate cancer, cardiac disease, and diabetes.