Analysis of Income Inequality Measures on Human Immunodeficiency Virus Mortality: a Spatiotemporal Bayesian Perspective

Abstract

Summary Social, economic, environmental and behavioural factors impacting health are well recognized in the literature. We consider the use of various income inequality measures in addition to a poverty measure and investigate their effects on human immunodeficiency virus (HIV) mortality. In doing so, we make use of models that can capture zero inflation and spatiotemporal effects. The research is motivated by the lack of studies from an inference and modelling perspectives in explaining HIV mortality by using measures that take into account socio‐economic status as well as time and location. Such a study can help policy makers to identify cases of environmental injustice and areas of outstanding health risk to assist in resource allocation problems. In our numerical example, we make use of mortality data obtained for the state of New York, estimate model parameters from a Bayesian inference perspective and discuss the implications and interpretations of various income inequality measures. The methodological novelty of our study is the introduction of a zero‐inflated Poisson model that can account for both spatial and temporal effects across 5 years (2000–2004). The practical novelty of our study is its attempt to find inequality measures which can improve our understanding of HIV mortality risk. Our results indicate that, for the data at hand, if inequality is calculated on the basis of county‐specific income shares rather than the whole state, HIV mortality can be better explained. In addition, accounting for temporal and spatial effects was found to contribute to our understanding of HIV mortality risk.

A Bayesian approach for analyzing zero‐inflated clustered count data with dispersion

“…Another possibility might be to consider a spatial structure to define the dependence across teeth, perhaps along the lines of some previous work. () Intuitively, we may expect that adjacent teeth are highly correlated and that the dependence decays farther away the teeth are located. In principle, our model can handle this approach by defining a γ i j and δ i j term for every tooth, and the full vector is drawn from MVN with covariance matrix Σ i that provides a (low‐dimension) spatial structure on the basis of tooth location.…”

“…() There are also mixed effects models for zero‐inflated clustered data that use random effects to introduce dependence. () However, these methods are typically limited to the equidispersed or overdispersed cases because of the properties of Poisson and negative binomial distributions.…”

Assessing the determinants of healthcare expenditures in single-person households

“…There is also a clear connection between a person’s socioeconomic situation and his or her state of health [ 7 ]. Abundant professional literature and copious findings on this issue demonstrate a relation between income inequality and disparities and mortality and other health indicators [ 1 , 30 , 36 ]. A study encompassing data from twenty-two European countries, for example, found that in almost all countries surveyed, low-socioeconomic-status groups had higher mortality rates and lower health self-assessments than those of high socioeconomic status [ 31 ].…”

A Hierarchical Mixed Effect Hurdle Model for Spatiotemporal Count Data and its Application to Identifying Factors Impacting Health Professional Shortages

“…The data exhibit a complex dependence structure. Counties are nested within states, and adjacent counties are considered to be geographically correlated (Aktekin and Musal, ; Musal and Aktekin, ). The observations from different counties within a state are likely to be impacted by state level policies; thus, they are possibly correlated.…”