Bayesian Spatial Modeling of Diabetes and Hypertension: Results from the South Africa General Household Survey

Abstract

Determining spatial links between disease risk and socio-demographic characteristics is vital in disease management and policymaking. However, data are subject to complexities caused by heterogeneity across host classes and space epidemic processes. This study aims to implement a spatially varying coefficient (SVC) model to account for non-stationarity in the effect of covariates. Using the South Africa general household survey, we study the provincial variation of people living with diabetes and hypertension risk through the SVC model. The people living with diabetes and hypertension risk are modeled using a logistic model that includes spatially unstructured and spatially structured random effects. Spatial smoothness priors for the spatially structured component are employed in modeling, namely, a Gaussian Markov random field (GMRF), a second-order random walk (RW2), and a conditional autoregressive (CAR) model. The SVC model is used to relax the stationarity assumption in which non-linear effects of age are captured through the RW2 and allow the mean effect to vary spatially using a CAR model. Results highlight a non-linear relationship between age and people living with diabetes and hypertension. The SVC models outperform the stationary models. The results suggest significant provincial differences, and the maps provided can guide policymakers in carefully exploiting the available resources for more cost-effective interventions.

A Systematic Review of Areal Units and Adjacency Used in Bayesian Spatial and Spatio-Temporal Conditional Autoregressive Models in Health Research

“…Of the total of 52 studies, 26 (50.0%) were on infectious diseases, 11 (21.1%) on malaria [ 30 , 31 , 32 , 33 , 34 , 35 , 36 , 37 , 38 , 39 ], 8 (15.3%) on Dengue fever [ 40 , 41 , 42 , 43 , 44 , 45 ], 4 (7.6%) on HIV [ 46 , 47 , 48 , 49 ], and 3 (5.7%) on TB [ 50 , 51 , 52 ]. Overall, 10 (19.2%) studies were on chronic diseases [ 53 , 54 , 55 , 56 , 57 , 58 , 59 , 60 , 61 , 62 ], 8 (15.5%) on maternal and child health [ 63 , 64 , 65 , 66 ], 3 (5.7%) on nutrition [ 67 , 68 , 69 ] and 5 (9.6%) were applied to other health-related outcomes [ 58 , 70 , 71 , 72 , 73 ]. Most of the studies (42, 80.7%) used an ecological study design [ 30 , 31 , 33 , 34 , 37 , 38 , ...…”

“…Overall, 10 (19.2%) studies were on chronic diseases [ 53 , 54 , 55 , 56 , 57 , 58 , 59 , 60 , 61 , 62 ], 8 (15.5%) on maternal and child health [ 63 , 64 , 65 , 66 ], 3 (5.7%) on nutrition [ 67 , 68 , 69 ] and 5 (9.6%) were applied to other health-related outcomes [ 58 , 70 , 71 , 72 , 73 ]. Most of the studies (42, 80.7%) used an ecological study design [ 30 , 31 , 33 , 34 , 37 , 38 , 39 , 40 , 41 , 42 , 43 , 44 , 45 , 46 , 47 , 51 , 52 , 53 , 54 , 55 , 56 , 58 , 59 , 60 , 61 , 62 , 64 , 65 , 66 , 67 , 69 ,…”

“…The majority of the studies (80.7%) used years as a temporal unit measure and 9 (17.3%) studies used months as a temporal unit measure. A total of 46 (88.5%) studies used the MCMC estimation technique for their Bayesian modelling [ 31 , 32 , 33 , 35 , 36 , 37 , 38 , 39 , 40 , 41 , 42 , 43 , 44 , 45 , 46 , 47 , 48 , 49 , 50 , 51 , 52 , 54 , 55 , 57 , 58 , 59 , 60 , 61 , 62 , 63 , 64 , 65 , 68 , 69 , 70 , 71 , 72 , 73 , 74 , 75 , 76 , 77 , 78 , 79 , 80 , 81 ] and 6 (11.5%) studies used INLA as their estimation technique [ 30 , 34 , 53 , 56 , 66 , 67 ].…”

“…Of the 52 studies, 24 (46.1%) reported model diagnostics in their study. The majority of the studies (44, 84.6%) reported their data in a choropleth map [ 30 , 31 , 32 , 33 , 34 , 35 , 36 , 37 , 38 , 39 , 40 , 41 , 42 , 44 , 45 , 46 , 47 , 48 , 49 , 50 , 51 , 52 , 53 , 55 , 56 , 57 , 58 , 59 , 62 , 63 , 64 , 65 , 67 , 68 , 69 , 70 , 71 , 73 , 74 , 75 , 76 , 77 , 78 , 79 , 80 ], and only 5 (9.6%) studies provided script codes as a Supplementary File [ 41 , 50 , 51 , 71 , 72 ]. A total of 28 (53.8%) studies reported that burn-in, thinning, and iteration were used in their Bayesian modelling.…”

Spatial patterns in sociodemographic factors explain to a large extent the prevalence of hypertension and diabetes in Aragon (Spain)

“…Previous studies modeling the relationships between these diseases and socioeconomic risk factors through GWR techniques have also found regional variations. Ogunsakin ( 43 ) revealed that educational status had different effects across provinces of South Africa when investigating diabetes and hypertension. Park et al ( 44 ) addressed the hypertensive disease mortality in South Korea and found that the effects of socioeconomic status were only consistent in some regions of the country.…”