Abstract
Considering only the spatial component of diseases can identify areas with reduced or elevated risk, but not capture anything about temporal variation of risk which could be more or equally crucial. Hence, both spatial and temporal components of diseases need to be considered. Bayesian methods are useful due to the ease of specifying additional information, including temporal or spatial structure, through prior distributions. Here, we examine a range of different Bayesian spatio-temporal models available using CARBayes. Combinations of model formulations and climatic covariates were compared using goodness-of-fit measures, such as Watanabe Akaike Information Criterion (WAIC). Comparisons were made in the context of a substantive case study, namely monthly dengue fever incidence from January 2013 to December 2017 and climatic covariates in 14 geographic areas of Makassar, Indonesia. A spatio-temporal conditional autoregressive adaptive model combining rainfall and average humidity provided the most suitable model.
Highlights
Spatial or spatio-temporal models using Bayesian methods are useful in modelling dengue fever
The six models implemented in CARBayesST were compared: Spatio-temporal conditional Autoregressive (ST conditional autoregressive (CAR)) linear, ST CAR Autoregressive (AR), ST CAR adaptive, ST CAR separate spatial and ST CAR localised models, but no climatic variables were included
Since the localised model was not recommended for monthly dengue fever data, this paper aims to examine the most suitable Bayesian ST CAR models in modelling monthly dengue fever with and without climatic factors
Summary
Spatial or spatio-temporal models using Bayesian methods are useful in modelling dengue fever. Aswi et al [1] found through a systematic review that only a limited number of studies included Bayesian spatio-temporal random effects when modelling dengue fever. When they were used, the spatial random effects were commonly assigned a conditional autoregressive (CAR) prior, while the temporal effects were commonly used the first order autoregressive AR(1). Aswi et al [2] compared six different Bayesian spatio-temporal CAR models using annual dengue data across Makassar, Indonesia from 2002 to 2015. Another study used only Bayesian ST CAR localised models and some climatic covariates in examining monthly and annual dengue data [3].
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
