Selected work

Representative Papers

Suicide Mortality in Spain (2010-2022): Temporal Trends, Spatial Patterns, and Risk Factors

Sep 01, 2025

This study investigates the spatiotemporal heterogeneity and socioeconomic drivers of age- and sex-specific suicide mortality across Spanish provinces from 2010 to 2022. Method: Leveraging provincial-level panel data, we estimate a multilevel mixed Poisson regression model that controls for spatiotemporal confounding effects to systematically assess how structural factors—including rurality rate and unemployment rate—differentially influence male and female suicide risk. Contribution/Results: (1) Female suicide mortality increased significantly over time, whereas male rates remained relatively stable; (2) a 10-percentage-point increase in rurality was associated with a 5.3% rise in male suicide mortality; (3) a one-percentage-point rise in unemployment corresponded to a 2.4% increase in female suicide mortality. This is the first national-scale study to identify asymmetric effects of urban–rural structure and labor market conditions on gendered suicide risk, providing empirical foundations for targeted, regionally tailored mental health interventions.

0 citationsRead paper

Multivariate Spatio-temporal Modelling for Completing Cancer Registries and Forecasting Incidence

Jul 29, 2025

Cancer incidence data often suffer from temporal lags, incompleteness, and asynchronous registration initiation across regions and countries, resulting in spatiotemporal misalignment. To address this, we propose a multivariate spatiotemporal shared-component Bayesian hierarchical model that jointly integrates mortality data with sparse, asynchronous incidence observations. Using Markov Chain Monte Carlo (MCMC) inference, the model imputes missing registrations and forecasts future trends. Our key contributions are: (i) the first integration of a spatiotemporal shared structure into a multivariate Bayesian framework, explicitly accommodating heterogeneous registration start times and uneven geographic coverage; and (ii) rigorous model selection via cross-validation and predictive error assessment. Evaluated on English lung cancer data (2001–2019), the model substantially improves the timeliness and spatial resolution of incidence estimates, enabling more accurate, high-resolution cancer burden assessment. The approach is generalizable to other cancers and jurisdictions with fragmented surveillance systems.

0 citationsRead paper

A proposal for homoscedastic modelling with conditional auto-regressive distributions

Jul 11, 2025

Conditional autoregressive (CAR) models, widely used in spatial modeling, suffer from inherent conditional construction that induces underestimation of marginal variances at boundary regions, severe heteroscedasticity, and geometry-dependent artifacts—compromising posterior inference reliability in applications such as disease mapping. To address these limitations, we propose the homoscedastic CAR (HCAR) distribution, which integrates a global variance constraint and an adjacency-structure-aware weight rescaling mechanism into the classical CAR framework, explicitly correcting edge effects and geometric bias. HCAR preserves spatial dependence modeling capability while ensuring uniform marginal variances across all regions, thereby enhancing posterior stability and interpretability. Empirical evaluation on multiple real-world disease mapping datasets demonstrates that HCAR effectively mitigates boundary anomalies, yields more plausible marginal distributions, and remains computationally compatible with standard Bayesian inference pipelines.

0 citationsRead paper

Recent publications

Latest Papers

Suicide Mortality in Spain (2010-2022): Temporal Trends, Spatial Patterns, and Risk Factors

Sep 01, 2025

This study investigates the spatiotemporal heterogeneity and socioeconomic drivers of age- and sex-specific suicide mortality across Spanish provinces from 2010 to 2022. Method: Leveraging provincial-level panel data, we estimate a multilevel mixed Poisson regression model that controls for spatiotemporal confounding effects to systematically assess how structural factors—including rurality rate and unemployment rate—differentially influence male and female suicide risk. Contribution/Results: (1) Female suicide mortality increased significantly over time, whereas male rates remained relatively stable; (2) a 10-percentage-point increase in rurality was associated with a 5.3% rise in male suicide mortality; (3) a one-percentage-point rise in unemployment corresponded to a 2.4% increase in female suicide mortality. This is the first national-scale study to identify asymmetric effects of urban–rural structure and labor market conditions on gendered suicide risk, providing empirical foundations for targeted, regionally tailored mental health interventions.

0 citationsRead paper

Multivariate Spatio-temporal Modelling for Completing Cancer Registries and Forecasting Incidence

Jul 29, 2025

Cancer incidence data often suffer from temporal lags, incompleteness, and asynchronous registration initiation across regions and countries, resulting in spatiotemporal misalignment. To address this, we propose a multivariate spatiotemporal shared-component Bayesian hierarchical model that jointly integrates mortality data with sparse, asynchronous incidence observations. Using Markov Chain Monte Carlo (MCMC) inference, the model imputes missing registrations and forecasts future trends. Our key contributions are: (i) the first integration of a spatiotemporal shared structure into a multivariate Bayesian framework, explicitly accommodating heterogeneous registration start times and uneven geographic coverage; and (ii) rigorous model selection via cross-validation and predictive error assessment. Evaluated on English lung cancer data (2001–2019), the model substantially improves the timeliness and spatial resolution of incidence estimates, enabling more accurate, high-resolution cancer burden assessment. The approach is generalizable to other cancers and jurisdictions with fragmented surveillance systems.

0 citationsRead paper

A proposal for homoscedastic modelling with conditional auto-regressive distributions

Jul 11, 2025

Conditional autoregressive (CAR) models, widely used in spatial modeling, suffer from inherent conditional construction that induces underestimation of marginal variances at boundary regions, severe heteroscedasticity, and geometry-dependent artifacts—compromising posterior inference reliability in applications such as disease mapping. To address these limitations, we propose the homoscedastic CAR (HCAR) distribution, which integrates a global variance constraint and an adjacency-structure-aware weight rescaling mechanism into the classical CAR framework, explicitly correcting edge effects and geometric bias. HCAR preserves spatial dependence modeling capability while ensuring uniform marginal variances across all regions, thereby enhancing posterior stability and interpretability. Empirical evaluation on multiple real-world disease mapping datasets demonstrates that HCAR effectively mitigates boundary anomalies, yields more plausible marginal distributions, and remains computationally compatible with standard Bayesian inference pipelines.

0 citationsRead paper