Activity Number:
|
214
- Contributed Poster Presentations: Quality and Productivity Section
|
Type:
|
Contributed
|
Date/Time:
|
Tuesday, August 4, 2020 : 10:00 AM to 2:00 PM
|
Sponsor:
|
Quality and Productivity Section
|
Abstract #313776
|
|
Title:
|
A Bayesian Approach for Post-Market Safety Surveillance of New Products
|
Author(s):
|
Wei Zhou* and Danielle Boree and Jiali Lin
|
Companies:
|
Johnson & Johnson Vision and Johnson & Johnson Vision and Johnson & Johnson Vision
|
Keywords:
|
Bayesian ;
informative prior ;
applied ;
safety surveillance;
medical device;
generalized linear model
|
Abstract:
|
The FDA requires all medical device manufacturers conduct post-market safety surveillance. When a new product is launched, only limited historical data are available which poses unique safety risks for patients and challenges to the safety monitoring teams. For established products, abundant historical data are available and the maximum likelihood estimation (MLE) method can be used for inference. However, for a new product, the MLE method may be unreliable due to the insufficient data. This paper presents a Bayesian approach for new product signal detection for its flexibility in handling limited data. Poisson regression is fit to the complaint counts and informative Gaussian priors for key product attributes as predictors are constructed by utilizing historical data from established products during their respective launch periods. The prior means and variances of the regression parameters are given by the MLE method in fitting a Poisson model to established products. A prediction interval for the new product complaint count is produced from the posterior predictive distribution. An observed count above the prediction interval is considered as a safety signal.
|
Authors who are presenting talks have a * after their name.