Abstract:
|
In clinical research it is often desirable to have a scalar outcome measure, according to which treatments are compared. For example, in precision medicine, a primary focus has been on determining treatment decisions that maximize the value of a decision when the outcome of interest is a scalar variable. For longitudinal studies where the individual outcomes are trajectories over time, several options exist for extracting a useful scalar outcome from the trajectory. The scalar quantity considered in this article is the average rate of change of the trajectory and the focus is on quadratic trajectories although more complex trajectories are also discussed. Additionally, in studies where there are high placebo response rates, it is desirable to extract information from the trajectories that can distinguish placebo effect from specific drug effect. This motivates a weighting approach to modify the scalar quantity extracted from the curve.
|