A statistical downscaling approach for generating high spatial resolution health risk maps: a case study of road noise and ischemic heart disease mortality in Melbourne, Australia

Inferring the regional distribution of Visceral Leishmaniasis incidence from data at different spatial scales

BACKGROUND

As cases of visceral leishmaniasis (VL) in India dwindle, there is motivation to monitor elimination progress on a finer geographic scale than sub-district (block). Low-incidence projections across geographically- and demographically- heterogeneous communities are difficult to act upon, and equitable elimination cannot be achieved if local pockets of incidence are overlooked. However, maintaining consistent surveillance at this scale is resource-intensive and not sustainable in the long-term.

METHODS

We analysed VL incidence across 45,000 villages in Bihar state, exploring spatial autocorrelation and associations with local environmental conditions in order to assess the feasibility of inference at this scale. We evaluated a statistical disaggregation approach to infer finer spatial variation from routinely-collected, block-level data, validating against observed village-level incidence.

RESULTS

This disaggregation approach does not estimate village-level incidence more accurately than a baseline assumption of block-homogeneity. Spatial auto-correlation is evident on a block-level but weak between neighbouring villages within the same block, possibly suggesting that longer-range transmission (e.g., due to population movement) may be an important contributor to village-level heterogeneity.

CONCLUSIONS

Increasing the range of reactive interventions to neighbouring villages may not improve their efficacy in suppressing transmission, but maintaining surveillance and diagnostic capacity in areas distant from recently observed cases - particularly along routes of population movement from endemic regions - could reduce reintroduction risk in currently unaffected villages. The reactive, spatially-targeted approach to VL surveillance limits interpretability of data observed at the village level, and hence the feasibility of routinely drawing and validating inference at this scale.

Effects of ambient heat exposure on risk of all-cause mortality in children younger than 5 years in Africa: a pooled time-series analysis

BACKGROUND

Reducing child mortality is a Sustainable Development Goal, and climate change constitutes numerous challenges for Africa. Previous research has shown an association between leading causes of child mortality and climate change. However, few studies have examined these effects in detail. We aimed to explore the effects of ambient heat on neonate, post-neonate, and child mortality rates.

METHODS

For this pooled time-series analysis, health data were obtained from the International Network for the Demographic Evaluation of Populations and Their Health (INDEPTH) Health and Demographic Surveillance System. We included data from 29 settlements from 13 countries across Africa, collected via monthly surveys from Jan 1, 1993, to Dec 31, 2016. Climate data were obtained from ERA5, collected from Jan 1, 1991, to Dec 31, 2019. We pooled these data for monthly mean daily maximum wet bulb globe temperature (WBGT) and downscaled to geolocations. Due to data heaping, we pooled our health data on a monthly temporal scale and a spatial scale into six different climate regions (ie, Sahel [ie, Burkina Faso and northern Ghana], Guinea [ie, southern Ghana, Côte d'Ivoire, and Nigeria], Senegal and The Gambia, eastern Africa [ie, Kenya, Malawi, Tanzania, Mozambique, and Uganda], South Africa, and Ethiopia). Our outcomes were neonate (ie, younger than 28 days), post-neonate (ie, aged 28 days to 1 year), and child (ie, older than 1 year and younger than 5 years) mortality. To assess the association between WBGT and monthly all-cause mortality, we used a time-series regression with a quasi-Poisson, polynomial-distributed lag model.

FINDINGS

Between Jan 1, 1993, and Dec 31, 2016, there were 44 909 deaths in children younger than 5 years across the 29 sites in the 13 African countries: 10 078 neonates, 14 141 post-neonates, and 20 690 children. We observed differences in the association of heat with neonate, post-neonate, and child mortality by study region. For example, for Ethiopia, the relative risk ratio of mortality at the 95th percentile compared with median heat exposure during the study period was 1·14 (95% CI 1·06-1·23) for neonates, 0·99 (0·90-1·07) for post-neonates, and 0·79 (0·73-0·87) for children. Across the whole year, there was a significant increase in the relative risk of increased mortality for children in eastern Africa (relative risk 1·27, 95% CI 1·19-1·36) and Senegal and The Gambia (1·11, 1·04-1·18).

INTERPRETATION

Our results show that the influence of extreme heat on mortality risk in children younger than 5 years varies by age group, region, and season. Future research should explore potentially informative ways to measure subtleties of heat stress and the factors contributing to vulnerability.

FUNDING

EU Horizons as part of the Heat Indicators for Global Health (HIGH) Horizons project.

Spatiotemporal high-resolution prediction and mapping: methodology and application to dengue disease

Dengue disease has become a major public health problem. Accurate and precise identification, prediction and mapping of high-risk areas are crucial elements of an effective and efficient early warning system in countering the spread of dengue disease. In this paper, we present the fusion area-cell spatiotemporal generalized geoadditive-Gaussian Markov random field (FGG-GMRF) framework for joint estimation of an area-cell model, involving temporally varying coefficients, spatially and temporally structured and unstructured random effects, and spatiotemporal interaction of the random effects. The spatiotemporal Gaussian field is applied to determine the unobserved relative risk at cell level. It is transformed to a Gaussian Markov random field using the finite element method and the linear stochastic partial differential equation approach to solve the "big n" problem. Sub-area relative risk estimates are obtained as block averages of the cell outcomes within each sub-area boundary. The FGG-GMRF model is estimated by applying Bayesian Integrated Nested Laplace Approximation. In the application to Bandung city, Indonesia, we combine low-resolution area level (district) spatiotemporal data on population at risk and incidence and high-resolution cell level data on weather variables to obtain predictions of relative risk at subdistrict level. The predicted dengue relative risk at subdistrict level suggests significant fine-scale heterogeneities which are not apparent when examining the area level. The relative risk varies considerably across subdistricts and time, with the latter showing an increase in the period January-July and a decrease in the period August-December.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10109-021-00368-0.

Dynamic stiffness of road pavements: Construction characteristics-based model and influence on tire/road noise

Tire/pavement noise is produced by the interaction between a tire and the road surface. Complex noise generation mechanisms are involved in this process. Dynamic stiffness (or mechanical impedance) is considered as a related reduction/amplification mechanism. Despite its potential importance, in conventional hot bituminous mixtures this mechanism has little influence on tire/pavement noise. This is because the dynamic stiffness of tire treads is significantly lower than that of the hot bituminous mixes (conventional road surfaces). This paper presents a field assessment of the dynamic stiffness of five experimental sections with different characteristics: from the environmentally friendly mixes such as the Cold-In place-Recycled (CIR) pavement to the gap graded hot bituminous mixtures with crumb rubber (CR) from end of life tires (ELTs). The obtained dynamic stiffness spectra are fitted using a Boltzmann Sigmoid function. The relationships between the fitting coefficients and a few pavement construction characteristics are assessed in order to model the pavement dynamic stiffness as a function of these characteristics. In addition to dynamic stiffness, several surface characteristics such as the Close ProXimity (CPX) sound levels from the tire/pavement interaction, macrotexture in terms of the Mean Profile Depth (MPD), unevenness in terms of the International Roughness Index (IRI), and acoustic absorption are studied. According to our results, lower dynamic stiffness values are related to lower noise levels at high frequencies. No relationship was observed between the low dynamic stiffness values and noise attenuation at medium frequencies in this research work.