|
Abstract:
|
In the emergency setting, the time it takes to receive the treatment (medical device or drug) is often one of the most crucial factors for predicting clinical outcome. However, the clinical trial is not typically designed to study the effect of the time of treatment on the final clinical outcome. The most appropriate time point to deliver the emergency treatment is of great interest to both clinicians seeking to improve clinical practice and researchers seeking to gain knowledge for future clinical trials. We refer to the critical time point at which treatment effects before and after differ drastically from each other as the changing point. With ordinal outcomes, such as modified Rankin Scale, which evaluates the functional recovery for stroke patients, we propose a two-stage iterative algorithm to find the changing time point. We also develop a likelihood-based sequential hypothesis test to evaluate whether the identified changing time points are meaningful. The performance of our method is evaluated by simulation and psuedo-real data simulated from a real study. We expect that this agile tool may assist researchers in finding the crucial time interval to perform the treatment.
|
