|
Activity Number:
|
335
- Spatial Smoothing and Bayesian Uncertainty Quantification
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 5, 2020 : 10:00 AM to 2:00 PM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
|
Abstract #312341
|
|
|
Title:
|
Computational Methods in Bayesian Disease Mapping: An Application to Chronic Lower Respiratory Disease (CLRD) Mortality in the U.S., 2000-2017
|
|
Author(s):
|
Diba Khan* and John Pleis and Elizabeth Arias and Maria Dolores Ugarte and Aritz Adin
|
|
Companies:
|
CDC and National Center for Health Statistics and CDC/NCHS and Public University of Navarre Pamplona, Spain and Public University of Navarre Pamplona, Spain
|
|
Keywords:
|
Bayesian;
Small area;
Hierarchichal;
software;
computation
|
|
Abstract:
|
Hierarchical spatio-temporal Bayesian models including covariates are often used to describe spatio-temporal patterns in small areas. Due to the methodological and computational complexity in dealing with large datasets, several methods have been proposed in the literature. Markov Chain Monte Carlo (MCMC) methods have been traditionally used in Bayesian disease mapping but they are computationally intensive and time consuming. A viable alternative to MCMC methods to reduce computational cost is the INLA (Integrated nested Laplace approximations) technique. Appropriate selection of method and software in analyzing spatio-temporal variations in health outcomes results in less computation time, flexibility in implementation of appropriate models and precise estimates. In this study we propose to fit and compare small area models often used in Bayesian disease mapping using the software NIMBLE (Numerical Inference for Statistical Models using Bayesian and Likelihood Estimation) and R-INLA to estimate county level Chronic Lower Respiratory Disease (CLRD) mortality risks in the U.S., 2000-2017. A simulation study will be also conducted to compare both estimation techniques.
|
Authors who are presenting talks have a * after their name.
