|
Activity Number:
|
187
- Bayesian Analysis of Spatial and Time Series Data
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 4, 2020 : 10:00 AM to 2:00 PM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
|
Abstract #312964
|
|
|
Title:
|
Analysis of Personalized Trials Using Bayesian Distributed Lag Model with AR(P) Error
|
|
Author(s):
|
Ziwei Liao* and Ken Cheung and Ian Kronish and Karina Davidson
|
|
Companies:
|
Columbia University and Columbia University and Columbia University, Department of medicine and Feinstein Institute for Medical Research
|
|
Keywords:
|
Bayesian;
distributed lag;
autoregressive error;
personalized trial;
carryover effect;
time series
|
|
Abstract:
|
The ultimate goal of personalized or precision medicine is to identify the best treatment at patient level considering individual variability. In order to maximize the benefit and minimize potential risk to individual patient, an innovative type of trial is expected, which focuses on individual outcome instead of population or group mean responses to a given intervention. Existing statistical methods in personalized trial include nonparametric test, mixed effect model and autoregressive model. Personalized trial are usually multiple-period crossover trials performed within a single individual. These methods may fail to handle measurements autocorrelation both from consecutive interventions and errors. Also, it is expected to adjust for potential carryover effects. Distributed lag model is a regression model that uses lagged predictors. We propose a Bayesian distributed lag model with autocorrelated errors (BDLAR) that integrate prior knowledge on the shape of lagged coefficients and explicitly model the magnitude and length of carryover effect. We give real data examples to illustrate our method and simulation study was conducted to compare the performance of different models.
|
Authors who are presenting talks have a * after their name.
