Date of Award
Open Access Thesis
Generalized linear models which include spatially varying coefficient terms allow researchers to determine if the association between predictor and outcome variables vary across geographic space. Such models are particularly applicable to research with public health data where interventions and limited health care resources must be allocated carefully. The integrated nested Laplace approximation (INLA) methodology available in the R INLA package is a popular tool to estimate spatially varying coefficients. To assess the performance of the estimation procedure, patient emergency department (ED) visits were simulated from data sourced from a pilot study at Prisma Health. The INLA technique was used to assess whether the association between the patient response to a social determinant of health (SDoH) screening question and the number of ED visits varied across census block groups. The power of the estimation procedure for increasing numbers of positive screening rates of the SDoH question and for varying values of the variance parameter governing the distribution of the spatially varying coefficients was of interest. Furthermore, the type I error rate of the INLA estimation was also investigated. It was found that the power in detecting spatial variation increases as both the number of positive screens and variance parameter increases. The type I error rate remained below 0.1% for all simulations. The INLA estimation procedure was subsequently applied to the Prisma Health pilot study data, and no spatial variation for the association between screening positive for violence/abuse SDoH and ED visits was found.
DeMass, R. J.(2023). Detecting Spatially Varying Coefficient Effects With Conditional Autoregressive Models: A Simulation Study Using Social Determinants of Health Screening Data. (Master's thesis). Retrieved from https://scholarcommons.sc.edu/etd/7150
Available for download on Wednesday, May 15, 2024