Bayesian conditional autoregressive model for mapping human immunodeficiency virus incidence in the National Capital Region

Human Immunodeficiency Virus (HIV) infection continues to exhibit public health threat in the Philippines and causes several numbers of deaths globally. While the prevalence of HIV infection in the Philippines is still low, new HIV cases have been increasing by over 25% from 2001 to 2009. The mappin...

Full description

Saved in:
Bibliographic Details
Main Authors: Natividad, Jochelle Maan C., Necesito, Reinna Mae R.
Format: text
Language:English
Published: Animo Repository 2016
Subjects:
Online Access:https://animorepository.dlsu.edu.ph/etd_bachelors/13990
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: De La Salle University
Language: English
Description
Summary:Human Immunodeficiency Virus (HIV) infection continues to exhibit public health threat in the Philippines and causes several numbers of deaths globally. While the prevalence of HIV infection in the Philippines is still low, new HIV cases have been increasing by over 25% from 2001 to 2009. The mapping of disease incidence at smaller administrative districts or regions plays a big role in public health and epidemiology. Sixteen cities and one municipality of the National Capital Region (NCR) were treated as the spatial units in this study. Using Moran's I, results showed that sexually active male population of NCR exhibits significant spatial autocorrelation. It was observed that percentage of sexually active male individuals (15-85 years of age) are found to be correlated to the incidence of HIV in NCR. Bayesian Conditional Autoregressive (CAR) model that takes spatial autocorrelation into account was then used. The aim of this study is to provide a Bayesian spatial CAR that gives shrinkage and spatial smoothing of the raw relative risk estimates of HIV in the said region. In line with this, Poisson-Gamma models were also used, together with the Bayesian CAR, to identify the best fitted model for the estimated relative risk. Results revealed that Bayesian CAR was the best model as it involves spatial autocorrelation and has the lowest value of DIC and MSPE compared to the Poisson-Gamma models. It was found out that in Bayesian CAR, eight areas in NCR have high relative risk estimates namely Mandaluyong, Makati, Manila, Marikina, Pasay, Pasig, Pateros, and San Juan. It would be best for health public officials to provide health programs and to allocate funds in the said areas to reduce the incidence of HIV cases in the country.