Bayesian hierarchical modeling of joint spatiotemporal risk patterns for Human Immunodeficiency Virus (HIV) and Tuberculosis (TB) in Kenya

Abstract

The simultaneous spatiotemporal modeling of multiple related diseases strengthens inferences by borrowing information between related diseases. Numerous research contributions to spatiotemporal modeling approaches exhibit their strengths differently with increasing complexity. However, contributions that combine spatiotemporal approaches to modeling of multiple diseases simultaneously are not so common. We present a full Bayesian hierarchical spatio-temporal approach to the joint modeling of Human Immunodeficiency Virus and Tuberculosis incidences in Kenya. Using case notification data for the period 2012-2017, we estimated the model parameters and determined the joint spatial patterns and temporal variations. Our model included specific and shared spatial and temporal effects. The specific random effects allowed for departures from the shared patterns for the different diseases. The space-time interaction term characterized the underlying spatial patterns with every temporal fluctuation. We assumed the shared random effects to be the structured effects and the disease-specific random effects to be unstructured effects. We detected the spatial similarity in the distribution of Tuberculosis and Human Immunodeficiency Virus in approximately 29 counties around the western, central and southern regions of Kenya. The distribution of the shared relative risks had minimal difference with the Human Immunodeficiency Virus disease-specific relative risk whereas that of Tuberculosis presented many more counties as high-risk areas. The flexibility and informative outputs of Bayesian Hierarchical Models enabled us to identify the similarities and differences in the distribution of the relative risks associated with each disease. Estimating the Human Immunodeficiency Virus and Tuberculosis shared relative risks provide additional insights towards collaborative monitoring of the diseases and control efforts.

A Systematic Review of Joint Spatial and Spatiotemporal Models in Health Research

“…The Bayesian estimation approach was used in more than three-quarters (n = 35, 81.40%) of the included studies [ 32 , 33 , 35 , 36 , 37 , 38 , 39 , 40 , 41 , 42 , 43 , 44 , 45 , 46 , 47 , 48 , 49 , 50 , 51 , 52 , 53 , 54 , 55 , 56 , 57 , 58 , 59 , 60 , 61 , 62 , 63 , 64 , 65 , 73 , 79 ]. For Bayesian inference, the MCMC estimation approach was utilised in 27 studies [ 32 , 33 , 35 , 36 , 37 , 38 , 39 , 40 , 41 , 43 , 44 , 45 , 46 , 47 , 48 , 49 , 50 , 51 , 52 ,…”

“…To analyse two or more health outcomes simultaneously, several joint spatial and spatiotemporal models were used. To investigate the relative risk of the study variables and their risk factors, joint spatial models were used in 24 studies [ 36 , 38 , 39 , 40 , 42 , 43 , 45 , 46 , 47 , 48 , 49 , 50 , 51 , 52 , 54 , 56 , 57 , 59 , 60 , 61 , 62 , 64 , 73 , 75 ], joint spatiotemporal models in 12 studies [ 32 , 33 , 35 , 37 , 41 , 44 , 55 , 58 , 63 , 65 , 80 ], and joint spatial and spatiotemporal autocorrelation methods such as Moran Index statistics, Local Indicator Spatial Analysis (LISA), Getis Ord Gi statistics, or Kulldroff spatial and Spatio-temporal scan statistical tests used in 7 studies [ 34 , 72 , 74 , 76 , 77 , 78 , 79 ].…”

“…In joint spatial temporal models, the structured and unstructured spatial random effects were considered to account for the spatial dependence and independent effects, respectively. In more than two-thirds (n = 26, 72.2%) of the studies, prior CAR was considered for the spatially structured random effect [ 32 , 33 , 35 , 36 , 37 , 38 , 39 , 40 , 41 , 43 , 44 , 46 , 47 , 48 , 49 , 50 , 51 , 52 , 53 , 55 , 58 , 60 , 61 , 63 , 64 , 65 ], and all assigned Identical and Independent Distribution (IID) for the unstructured spatial random effects. Out of 12 joint spatiotemporal studies, seven (58.23%) considered prior first-order random walk to account for the temporal dependence in the model [ 32 , 33 , 35 , 41 , 53 , 58 , 65 ].…”

“…In more than two-thirds (n = 26, 72.2%) of the studies, prior CAR was considered for the spatially structured random effect [ 32 , 33 , 35 , 36 , 37 , 38 , 39 , 40 , 41 , 43 , 44 , 46 , 47 , 48 , 49 , 50 , 51 , 52 , 53 , 55 , 58 , 60 , 61 , 63 , 64 , 65 ], and all assigned Identical and Independent Distribution (IID) for the unstructured spatial random effects. Out of 12 joint spatiotemporal studies, seven (58.23%) considered prior first-order random walk to account for the temporal dependence in the model [ 32 , 33 , 35 , 41 , 53 , 58 , 65 ]. The simple exchangeable hierarchical structure was taken into consideration for the spatiotemporal interaction terms in five of the joint spatial and temporal studies [ 32 , 33 , 41 , 53 , 65 ].…”

“…The development of joint spatial and spatiotemporal methodologies has coincided with a huge advancement in statistical approaches addressing technical and statistical issues related to advanced spatial statistics. For example, authors employed different Bayesian inference techniques such as Markov chain Monte Carlo (MCMC) [ 32 , 33 , 35 , 36 , 37 , 38 , 39 , 40 , 41 , 43 , 44 , 45 , 46 , 47 , 48 , 49 , 50 , 51 , 52 , 53 , 54 , 55 , 56 , 57 , 58 , 59 , 60 ] or Integrated Nested Laplace Approximation (INLA) methods [ 42 , 43 , 61 , 62 , 63 , 64 , 65 ] in Bayesian shared component spatial and spatiotemporal models.…”