A0448
Title: Uncertainty quantification of treatment regime in precision medicine by confidence distributions
Authors: Min-ge Xie - Rutgers University (United States) [presenting]
Abstract: Personalized decision rule in precision medicine can be viewed as a discrete parameter, for which theoretical development for statistical inference is lagged behind. A new way to quantify the estimation uncertainty in a personalized decision using confidence distribution (CD) is proposed. Specifically, in a regression setup, suppose the decision for treatment vs control for an individual $x_a$ is determined by a linear decision rule $D_a = I(x_a b> x_a d)$, where $b$ and $d$ are unknown regression coefficients in models for potential outcomes of treatment and control, respectively. The data-driven decision $\hat D_a$ relies on the estimates of $b$ and $d$ and has uncertainty. We propose to find a CD for $c_a =x_a b - x_a d$ and compute a confidence measure of $\{D_a = 1\} = \{c_a > 0\}$. This measure is in $(0,1)$ and provides a frequency-based assessment on how reliable our decision is. For example, if the confidence measure of the decision ${D_a = 1}$ is 63\%, then we know, out of 100 patients who are the same as patient $x_a$, 63 will benefit to have the treatment and 37 will be better off to be in the control group. Numerical study suggests that this new measurement is inline with classical assessments (i.e., sensitivity, specificity), but different from the classical assessments, this measurement can be directly computed from the observed data. Utility of this new measure will also be illustrated in an application of an adaptive clinical trial.
