Analysing malaria incidence at the small area level for developing a spatial decision support system: A case study in Kalaburagi, Karnataka, India

Spatial decision support systems have already proved their value in helping to reduce infectious diseases but to be effective they need to be designed to reflect local circumstances and local data availability. We report the first stage of a project to develop a spatial decision support system for infectious diseases for Karnataka State in India. The focus of this paper is on malaria incidence and we draw on small area data on new cases of malaria analysed in two-monthly time intervals over the period February 2012 to January 2016 for Kalaburagi taluk, a small area in Karnataka. We report the results of data mapping and cluster detection (identifying areas of excess risk) including evaluating the temporal persistence of excess risk and the local conditions with which high counts are statistically associated. We comment on how this work might feed into a practical spatial decision support system.

Multi-pollutant exposure profiles associated with term low birth weight in Los Angeles County

Research indicates that multiple outdoor air pollutants and adverse neighborhood conditions are spatially correlated. Yet health risks associated with concurrent exposure to air pollution mixtures and clustered neighborhood factors remain underexplored. Statistical models to assess the health effects from pollutant mixtures remain limited, due to problems of collinearity between pollutants and area-level covariates, and increases in covariate dimensionality. Here we identify pollutant exposure profiles and neighborhood contextual profiles within Los Angeles (LA) County. We then relate these profiles with term low birth weight (TLBW). We used land use regression to estimate NO2, NO, and PM2.5 concentrations averaged over census block groups to generate pollutant exposure profile clusters and census block group-level contextual profile clusters, using a Bayesian profile regression method. Pollutant profile cluster risk estimation was implemented using a multilevel hierarchical model, adjusting for individual-level covariates, contextual profile cluster random effects, and modeling of spatially structured and unstructured residual error. Our analysis found 13 clusters of pollutant exposure profiles. Correlations between study pollutants varied widely across the 13 pollutant clusters. Pollutant clusters with elevated NO2, NO, and PM2.5 concentrations exhibited increased log odds of TLBW, and those with low PM2.5, NO2, and NO concentrations showed lower log odds of TLBW. The spatial patterning of pollutant cluster effects on TLBW, combined with between-pollutant correlations within pollutant clusters, imply that traffic-related primary pollutants influence pollutant cluster TLBW risks. Furthermore, contextual clusters with the greatest log odds of TLBW had more adverse neighborhood socioeconomic, demographic, and housing conditions. Our data indicate that, while the spatial patterning of high-risk multiple pollutant clusters largely overlaps with adverse contextual neighborhood cluster, both contribute to TLBW while controlling for the other.

Modelling collinear and spatially correlated data

In this work we present a statistical approach to distinguish and interpret the complex relationship between several predictors and a response variable at the small area level, in the presence of (i) high correlation between the predictors and (ii) spatial correlation for the response. Covariates which are highly correlated create collinearity problems when used in a standard multiple regression model. Many methods have been proposed in the literature to address this issue. A very common approach is to create an index which aggregates all the highly correlated variables of interest. For example, it is well known that there is a relationship between social deprivation measured through the Multiple Deprivation Index (IMD) and air pollution; this index is then used as a confounder in assessing the effect of air pollution on health outcomes (e.g. respiratory hospital admissions or mortality). However it would be more informative to look specifically at each domain of the IMD and at its relationship with air pollution to better understand its role as a confounder in the epidemiological analyses. In this paper we illustrate how the complex relationships between the domains of IMD and air pollution can be deconstructed and analysed using profile regression, a Bayesian non-parametric model for clustering responses and covariates simultaneously. Moreover, we include an intrinsic spatial conditional autoregressive (ICAR) term to account for the spatial correlation of the response variable.

Sampling from Dirichlet process mixture models with unknown concentration parameter: mixing issues in large data implementations

We consider the question of Markov chain Monte Carlo sampling from a general stick-breaking Dirichlet process mixture model, with concentration parameter [Formula: see text]. This paper introduces a Gibbs sampling algorithm that combines the slice sampling approach of Walker (Communications in Statistics - Simulation and Computation 36:45-54, 2007) and the retrospective sampling approach of Papaspiliopoulos and Roberts (Biometrika 95(1):169-186, 2008). Our general algorithm is implemented as efficient open source C++ software, available as an R package, and is based on a blocking strategy similar to that suggested by Papaspiliopoulos (A note on posterior sampling from Dirichlet mixture models, 2008) and implemented by Yau et al. (Journal of the Royal Statistical Society, Series B (Statistical Methodology) 73:37-57, 2011). We discuss the difficulties of achieving good mixing in MCMC samplers of this nature in large data sets and investigate sensitivity to initialisation. We additionally consider the challenges when an additional layer of hierarchy is added such that joint inference is to be made on [Formula: see text]. We introduce a new label-switching move and compute the marginal partition posterior to help to surmount these difficulties. Our work is illustrated using a profile regression (Molitor et al. Biostatistics 11(3):484-498, 2010) application, where we demonstrate good mixing behaviour for both synthetic and real examples.