Bayesian Zero-Inflated Joint Modelling on Ischaemic Heart Disease Mortality in California

Activity: SupervisionMaster Dissertation Supervision


The mortality of ischemic heart disease is influenced by many factors, one of which is the level of environmental air pollution, which greatly impacts this disease. Our study mainly focuses on a spatio-temporal analysis using a zero-inflated Poisson model on the monthly average mortality rate of ischemic heart disease and the monthly average PM2.5 concentration data collected from air quality monitoring stations set by various states in California,
USA, from 2018 to 2020. In Chapter 2, we conducted descriptive statistics on the collected data and other explanatory variables, identifying an excess of zeros in the disease data and missing values in the air pollution data. In Chapter 3, we used the INLA’s SPDE method to address the missing values and decided to conduct a spatio-temporal analysis of the standardized mortality rate of ischemic heart disease using a joint model of the Bernoulli distribution and zero-inflated Poisson distribution. This involved using RW1 and BYM2
models to explore the temporal and spatial random effects, respectively, and sharing temporal effects in the model. In Chapter 4, we analyzed the model’s results and found a significant relationship between the number of ischemic heart disease mortality events and the average concentration of PM2.5. The impact of other variables on the mortality rate of ischemic heart disease, apart from temperature and precipitation, was similar over time.