International Journal of Health Geographics最新文献

{"title":"Prevalence, spatial distribution, and determinants of anemia among under-five children in the Democratic Republic of Congo: evidence from the 2023-24 DHS.","authors":"Thomas Kidanemariam Yewodiaw, Helen Lamesgin Endalew, Mihret Getnet, Amare Belete Getahun, Tiget Ayelgn Mengstie, Engidaw Fentahun Enyew, Hiwot Tezera Endale, Tseganesh Asefa, Mequanent Dessie Bitewa","doi":"10.1186/s12942-026-00484-4","DOIUrl":"10.1186/s12942-026-00484-4","url":null,"abstract":"<p><strong>Background: </strong>Anemia among children under five remains a major public health problem in the Democratic Republic of Congo (DRC), affecting growth, cognitive development, and overall child health. We estimated the prevalence, spatial distribution, and determinants of anemia and identified high-risk geographic areas using multilevel and spatial analyses.</p><p><strong>Methods: </strong>We analyzed data from 11,393 children aged 6-59 months from the 2023-24 Democratic Republic of Congo Demographic and Health Survey (DHS). Weighted prevalence estimates were calculated, and multilevel mixed-effects logistic regression was used to assess individual- and community-level determinants of anemia. We applied Moran's I, hotspot detection, Kriging interpolation, and SaTScan analyses to identify high-risk areas. Model performance was evaluated using AIC, BIC, and log-likelihood ratio tests to ensure the results were reliable.</p><p><strong>Results: </strong>Overall, 51.7% of children (95% CI: 49.5%-53.9%) were anemic, most cases were mild (25.6%) or moderate (25.0%), while severe anemia was rare (1.1%). Prevalence was highest among infants 0-5 months (56.2%) and boys (52.5%) than girls (50.8%). Key risk factors for anemia included wasting (AOR = 1.29), malaria infection (AOR = 1.21), high community malaria risk (AOR = 1.42), and low altitude (AOR = 1.39). Children were less likely to be anemic if their mothers were overweight (AOR = 0.66), if they came from wealthier households (AOR = 0.77), or if they lived in moderately sized families (AOR = 0.84). Spatial analysis revealed high-prevalence areas in Haut-Lomami, Maniema, and Tshuapa, whereas lower prevalence was seen in Haut-Uele and Nord Ubangi.</p><p><strong>Conclusion: </strong>Anemia among under-five children in the DRC is geographically clustered and influenced by both individual- and community-level factors. Targeted and integrated interventions focusing on malaria prevention, nutrition, and maternal health are needed in high-burden areas.</p>","PeriodicalId":48739,"journal":{"name":"International Journal of Health Geographics","volume":"25 1","pages":""},"PeriodicalIF":4.1,"publicationDate":"2026-07-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13428445/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148649511","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Geospatial-based surveillance of malaria risk in dar es salaam using a hybrid 3DCNN + LSTM and CA model.","authors":"Edmund Steven Kanjagaile, Dorothea Deus, Anastazia Daniel Msusa","doi":"10.1186/s12942-026-00482-6","DOIUrl":"https://doi.org/10.1186/s12942-026-00482-6","url":null,"abstract":"<p><p>Malaria remains a significant public health burden in tropical and subtropical regions, where the efficient identification and prediction of risk areas remain challenging. Conventional field surveys used to map Anopheles breeding sites are costly, time consuming and often incomplete. Therefore, there is a pressing need for a geospatially integrated surveillance framework that can accurately map malaria risk and forecast future risk dynamics to support targeted control efforts. A geospatial hybrid modelling framework was developed by integrating multi-source remote sensing, malaria, and climate datasets. A Random Forest model was employed to determine the relative importance of the input variables, which were then weighted and selected for inclusion in a deep-learning architecture. The predictive model combines a 3D Convolutional Neural Network (3DCNN) to capture spatial patterns with a Long Short-Term Memory (LSTM) network to learn temporal dynamics. The model was trained against a baseline mean squared error (MSE) of 0.1 representing a naïve mean predictor. To improve spatial realism of the final risk maps, a Cellular Automata (CA) model was incorporated using a 3 × 3 Moore neighborhood structure, with parameters calibrated at γ = 0.293 and β = 43.9 to enhance the spatial propagation of risk across neighboring cells. The framework successfully mapped malaria risk in Dar es Salaam with a Spearman rank correlation of 0.92, identifying Kigamboni South, Tundwi, and Msongola as high-risk areas. The hybrid 3DCNN-LSTM model reduced the training by 97.3% and validation losses by 90.2% respectively from the baseline. Projections to 2060 indicate a steadily spatial increase in malaria risk, with a cumulative slope of 0.077 risk units over the 35-year horizon (an annual slope of 0.002), demonstrating the model's ability to estimate future risk.</p>","PeriodicalId":48739,"journal":{"name":"International Journal of Health Geographics","volume":" ","pages":""},"PeriodicalIF":4.1,"publicationDate":"2026-07-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148431306","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Characterizing spatiotemporal spread of infectious diseases using ellipse-shaped transmission hotspots: application to dengue virus outbreaks.","authors":"Pei-Sheng Lin, Chun-Hong Chen, Wei-Liang Liu, Tzai-Hung Wen, Yu-Chun Lu, Li-Wei Chen, Hsiang-Yu Yuan, Yi-Hung Kung","doi":"10.1186/s12942-026-00481-7","DOIUrl":"https://doi.org/10.1186/s12942-026-00481-7","url":null,"abstract":"<p><strong>Background: </strong>Identifying transmission hotspots associated with micro-clustering patterns at the early stages of epidemics is helpful to characterize spatiotemporal spread of infectious diseases. However, standard methods with statistical validation to establish a dynamic warning system for emerging infectious diseases are lacking. We therefore aimed to integrate a geographic information system-based surveillance system and data-driven methods to identify transmission hotspots, thereby assisting decision-makers to implement appropriate policies at the early stages of epidemics.</p><p><strong>Methods: </strong>We propose a method to identify micro-clusters and transform them into ellipse-shaped transmission hotspots to characterize disease propagation. The ellipse-shaped transmission hotspots built by the machine-learning method and mathematical models agree with 100(1 - α)% confidence regions in multivariate analysis.</p><p><strong>Results: </strong>We provided a flowchart for the construction of ellipse-shaped transmission hotspots. Import parameters, such as cluster range and orientation, were determined by statistical models as validation. The elliptical transmission hotspots reveal the expansion pattern for the dengue virus infection. Compared with density-based spatial clustering of applications with noise, the ellipse-shaped transmission hotspots more effectively characterize the orientations of the spread patterns at the early stage of epidemics.</p><p><strong>Conclusions: </strong>The geographic information system-based surveillance system used here to characterize ellipse-shaped transmission hotspots of dengue fever can also be applied to other infectious diseases to visualize the evolution of their dispersion areas and conduct early impact assessments. The mapped information could be aggregated to continuously update the propagation patterns, thereby helping public health departments to implement appropriate control measures for infectious diseases.</p>","PeriodicalId":48739,"journal":{"name":"International Journal of Health Geographics","volume":" ","pages":""},"PeriodicalIF":4.1,"publicationDate":"2026-07-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148425441","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"EpiGIS pro: an AI-powered geospatial intelligence platform for integrated disease surveillance and predictive analytics.","authors":"Chenxi Guo, Peter Scott","doi":"10.1186/s12942-026-00479-1","DOIUrl":"10.1186/s12942-026-00479-1","url":null,"abstract":"<p><strong>Background: </strong>Global infectious disease surveillance requires timely integration of heterogeneous data sources. Existing platforms typically address only one dimension, leaving cross-domain correlations unexplored. This study presents EpiGIS Pro, an AI-powered geospatial platform that consolidates automated disease event extraction, environmental monitoring, mobility analysis, and predictive analytics within a single web-based framework, providing a shared data infrastructure for future cross-domain modeling.</p><p><strong>Methods: </strong>EpiGIS Pro employs a modular multi-application architecture built on Django REST Framework with PostGIS spatial extensions. Disease intelligence is automated from six source categories using Claude AI (Anthropic) for structured extraction with standardized epidemiological metadata. Environmental data, traffic congestion indicators, and international flight route data are ingested via scheduled Celery Beat tasks. Machine learning modules include Prophet-based time-series forecasting with multiplicative seasonality for WHO FluNet data, seasonal Z-score anomaly detection, and effective reproduction number (R_t) estimation. Semantic search is enabled through Voyage AI embeddings stored in pgvector-indexed PostgreSQL. A systematic evaluation framework covers six analytical dimensions across 9 countries spanning 6 years of data.</p><p><strong>Results: </strong>The platform integrates 327,000 + records across six data domains. Claude AI extraction processed 353 disease event articles from 56 countries with 92.4% success and 100% completeness for disease type, severity, and priority fields. Using 2,868 WHO FluNet records across 9 countries (2019-2026), seasonality analysis correctly identified hemisphere-concordant peak timing in all 7 evaluated countries. Prophet multiplicative forecasting reduced RMSE by 21-29% compared to the seasonal naive baseline for the United States, and consistently outperformed log-transformed Prophet across all countries and horizons. R_t estimation demonstrated epidemiologically consistent patterns, with onset-period R_t exceeding trough-period R_t in Australia (1.27 vs. 1.09) and Japan (1.80 vs. 1.08). Cross-border risk assessment computed 4,828 flight corridor risk scores by linking disease activity with 633 international aviation routes.</p><p><strong>Conclusions: </strong>EpiGIS Pro demonstrates that consolidating AI-driven event extraction, multi-source environmental and mobility data, and geospatial analytics within a unified platform enhances situational awareness for global disease surveillance. The modular architecture and reproducible evaluation framework position EpiGIS Pro as both a practical surveillance tool and a research testbed for computational epidemiology.</p>","PeriodicalId":48739,"journal":{"name":"International Journal of Health Geographics","volume":" ","pages":""},"PeriodicalIF":4.1,"publicationDate":"2026-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13531970/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148266719","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Disparities of children's obesogenic environments in Louisiana.","authors":"Yutian Zeng, Fahui Wang, Senlin Chen","doi":"10.1186/s12942-026-00477-3","DOIUrl":"10.1186/s12942-026-00477-3","url":null,"abstract":"<p><strong>Background: </strong>Childhood obesity remains a pressing public health concern, with significant long-term implications for children's growth and development, as well as an increased risk for various diseases. In recent years, the obesity prevalence among children has shown a concerning rise, particularly in the Deep South (e.g., Louisiana) exhibiting higher rates compared to other regions of the United States. Meanwhile, environmental factors play a critical role in shaping children's health behaviors and outcomes. There has been an increasing body of research on the relation between obesogenic environment and health, but far less focused on child health.</p><p><strong>Methods: </strong>This study provides a comprehensive measurement of obesogenic environments for children in Louisiana covering park accessibility, walkability, and food environment at the ZIP Code area level. Geographic information system (GIS) analyses revealed significant urban-rural disparities and heterogeneity by race and socioeconomic deprivation. A three-way analysis of variance (ANOVA) further revealed the main and interaction effects of urbanicity, Black population percentage, and area deprivation index (ADI) on obesogenic environment indicators.</p><p><strong>Results: </strong>Regarding the disparities of walkability, not only the individual factors including urbanicity, Black population percentage, and ADI significantly affect walkability, but also the interaction terms among these factors, including both two-way and three-way interactions, are significant. Urban-rural differences and ADI levels contribute to disparities in the food environment. But no significant disparities are observed in park accessibility across different subgroups.</p><p><strong>Conclusion: </strong>While park accessibility disparities are limited, the food environment varies significantly by urbanicity and ADI level, walkability shows complex socio-spatial interactions. The study sheds light on targeted public health interventions and urban planning to close the gaps in obesogenic environments for children's health.</p>","PeriodicalId":48739,"journal":{"name":"International Journal of Health Geographics","volume":" ","pages":""},"PeriodicalIF":4.1,"publicationDate":"2026-06-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13330262/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148220320","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Spatial non-stationarity in son preference: a district-level geographically weighted regression analysis of NFHS-5 in India.","authors":"Soumen Barik, Dewaram A Nagdeve, Anuj Singh, Mayank Singh","doi":"10.1186/s12942-026-00475-5","DOIUrl":"10.1186/s12942-026-00475-5","url":null,"abstract":"<p><strong>Background: </strong>Son preference remains a key driver of gender inequality in India, yet most studies treat its determinants as uniform across space, obscuring critical subnational variation. Addressing this gap, this study investigates the geographic heterogeneity of son preference and examines how its predictors vary spatially. Data and methods Using district-level data on 102,045 ever married women aged 15-49 from the National Family Health Survey-5 (2019-2021), we applied a spatially explicit analytical framework, including choropleth mapping, Global Moran's I, hotspot analysis, Local Indicators of Spatial Association (LISA) cluster mapping, kriging interpolation, and Geographically Weighted Regression (GWR). The use of GWR was justified by significant spatial non-stationarity (Koenker BP = 32.78, p < 0.001) detected in OLS diagnostics.</p><p><strong>Results: </strong>Son preference prevalence ranged 8.2%-42.3%. Global Moran's I (0.397, z = 57.86, p < 0.001) confirmed significant clustering. OLS identified five significant predictors: parity 2, no mass media exposure, household size 5-8, illiterate mothers, and younger maternal age, explaining 64% variance. GWR demonstrated superior fit (AICc: 4459.98 vs 4478.46; adjusted R²: 0.65 vs 0.64). Local R² ranged 0.45-0.71, highest in northern/central districts. Women's illiteracy (β: 0.40-0.62), large household size (β: 0.35-0.83), and younger mothers showed strongest associations in northern/central India, but negligible effects in southern regions, confirming spatial heterogeneity.</p><p><strong>Conclusion: </strong>Son preference is a spatially embedded social process, shaped by localized patriarchy, economy, and institutions. Geographically targeted, gender-transformative policies such as conditional incentives for girls' schooling and localized media campaigns effectively address the structural devaluation of daughters.</p>","PeriodicalId":48739,"journal":{"name":"International Journal of Health Geographics","volume":" ","pages":""},"PeriodicalIF":4.1,"publicationDate":"2026-06-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13536809/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148213055","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Modeling hospital catchment areas in pediatric oncology using an empirically parameterized extended Huff-model.","authors":"Jonas Kapitza, Thomas Wieland, Markus Metzler","doi":"10.1186/s12942-026-00478-2","DOIUrl":"10.1186/s12942-026-00478-2","url":null,"abstract":"<p><strong>Background: </strong>Specialized pediatric oncology is typically concentrated in a few high-volume centers, creating tensions between the need for centralization and equitable spatial access. For regional health planning, robust methods are required to delineate hospital catchment areas and understand how structural site characteristics and accessibility shape patient-to-hospital travel flows. This study uses pediatric oncology in Bavaria, Germany, as a case to develop and test an extended, empirically calibrated Huff model for modeling hospital catchment areas.</p><p><strong>Methods: </strong>We analyzed 3,320 incident cases of pediatric oncology recorded in the German Childhood Cancer Registry between 2014 and 2023, which were treated at the seven specialized hospitals in Bavaria. An extended Huff model was specified that integrates structural indicators of hospital capacity and quality (bed capacity, staffing, cancer center accreditation), a spatial clustering variable that captures proximity-related interactions among nearby hospital sites, and a logistic distance-decay function based on travel times. Model parameters were estimated using maximum likelihood, and competing specifications were compared primarily using mean absolute percentage error (MAPE). A scenario analysis was conducted to assess how a reduction of nurse staffing ratios at two Munich hospitals would affect patient-to-hospital travel flows and catchment areas.</p><p><strong>Results: </strong>Our final baseline model, comprising four structural indicators, a clustering variable, and a logistic travel-time function, achieved a MAPE of 5.85% and an R² of 0.89. Capacity and quality indicators displayed positive effects on hospital choice, whereas the clustering parameter was negative, indicating proximity-related interaction effects among nearby hospitals. In the case scenario, a 20% reduction in the nursing staff ratio at the Munich sites led to declining modeled patient shares at both hospitals (- 2.0 and - 2.5% points, respectively) and corresponding gains primarily at Augsburg (+ 3.5% points) and Regensburg (+ 1.3% points), particularly in overlapping and transitional catchment zones.</p><p><strong>Conclusions: </strong>Our extended Huff model, which combines multidimensional structural indicators, spatial clustering, and realistic travel-time effects, can accurately represent hospital catchment areas and patient-to-hospital travel flows in specialized pediatric oncology. The approach provides a transparent, empirically grounded framework for assessing accessibility, identifying spatial interdependencies between hospital sites, and conducting scenario-based simulations to inform regional health planning and workforce policy in specialized care settings.</p>","PeriodicalId":48739,"journal":{"name":"International Journal of Health Geographics","volume":" ","pages":""},"PeriodicalIF":4.1,"publicationDate":"2026-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13273989/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148195222","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Spatial and telehealth accessibility to eating disorder treatment in the United States: evidence from registry and LLM-augmented data.","authors":"Lingbo Liu, Chuying Huo, Ariel L Beccia, Tracy K Richmond, S Bryn Austin","doi":"10.1186/s12942-026-00474-6","DOIUrl":"10.1186/s12942-026-00474-6","url":null,"abstract":"&lt;p&gt;&lt;strong&gt;Background: &lt;/strong&gt;Adequate geographic access to eating disorder treatment is essential for timely care. Yet the distribution of in-person and telehealth programs, and their accessibility across urban and rural settings and across communities with different socioeconomic conditions, remains unclear.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Methods: &lt;/strong&gt;We combined an official registry of eating disorder programs from the National Alliance for Eating Disorders with an AI-augmented web corpus built from U.S. domains using ISTARI.AI and large language models. After cleaning and linkage, we analyzed 328 registry centers for in-person care and an augmented set of 2,045 physical sites for proximity checks; telehealth programs were mainly limited to intensive outpatient and partial hospitalization. In-person accessibility was estimated at the census tract level with the two-step floating catchment area method (2SFCA) using program-based capacity. Telehealth accessibility used an unbounded two-step virtual catchment area (u2SVCA) that discounts demand by internet subscription. We also derived tract-level proximity indicators within 30 miles, namely nearest distance and the best available facility size. Population and covariates came from the American Community Survey and the 2020 Census. Multivariable ordinary least squares models related accessibility to median household income, rurality, education, insurance, and race or ethnicity. Correlation-based sensitivity analyses compared metrics across data sources and service regimes that either permit cross-state care or restrict care to within-state.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Results: &lt;/strong&gt;Large language models classified employee size from web summaries with an accuracy of 0.641 and Cohen's kappa of 0.176. In-person access concentrates in metropolitan corridors, with longer distances and lower accessibility in rural tracts; allowing cross-state care improves proximity near many borders, while within-state constraints reduce reachable capacity across interior states. Telehealth per capita availability varies by state, and effective telehealth access declines after discounting by subscription, with lower values across parts of the South and interior West and the longest mean distances in Alaska. Regression models show strong rural and income gradients. Higher income and urban residence are associated with shorter distance and higher accessibility, while rurality is associated with poorer access for both in-person and telehealth measures. Conditional associations for Black and Latine population shares are small once socioeconomic factors are included. Proximity metrics from the AI-augmented set are moderately correlated with registry-based 2SFCA scores, and telehealth and in-person accessibility show lower but nonzero correlation, suggesting that telehealth and in-person measures capture partially distinct dimensions of potential access, while cross-source differences may also reflect coverage and measurement differences.&lt;/p&gt;&lt;","PeriodicalId":48739,"journal":{"name":"International Journal of Health Geographics","volume":" ","pages":""},"PeriodicalIF":4.1,"publicationDate":"2026-05-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13483615/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148139560","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Spatio-temporal epidemic forecasting with graph-based transformer.","authors":"Mahmoud Ezzat, Youssef Mohamed Malek, Tamer AbdelKader, Nagwa Badr","doi":"10.1186/s12942-026-00476-4","DOIUrl":"10.1186/s12942-026-00476-4","url":null,"abstract":"<p><p>Epidemic forecasting plays a vital role in modern public health. The COVID-19 outbreak underscored the critical need for accurate and responsive models. Recent spatio-temporal Graph Neural Network (GNN) models that integrate human mobility networks face challenges in fully capturing complex, non-linear temporal dynamics and long-range spatial dependencies. To bridge this gap, we introduce two novel spatio-temporal architectures that combine GNNs with Transformer-based temporal modeling: a local-attention-model, which restricts self-attention to temporally adjacent windows of the same node, and a global-attention-model, which leverages full sequence-wide attention across all nodes to capture long-range dependencies. We benchmark our approaches against Persistence, Graph Convolutional Recurrent Network (GCRN) and Graph WaveNet baselines using two real-world datasets from Spain and Brazil. Our models show competitive and superior performance across most metrics compared to recurrent and temporal convolution baselines. The Linear Temporal Graph Convolutional Network (LinearTGCN) variant achieves the best Symmetric Mean Absolute Percentage Error (SMAPE) 24.74% and Mean Directional Accuracy (MDA) 72.52% on Spain dataset, outperforming the full attention-models. While, in Top-40 cities subset of Brazil dataset, local-attention-model slightly matches or outperforms the compared baselines with RMSE (3873.63) and SMAPE (83.47%). Our experiments demonstrate that simple linear models can match or exceed Transformers on structured time series, while Transformers show a great performance on noise or unstructured datasets like Brazil dataset. We found that SMAPE values varies across models by only a few percentage points, while the models are significantly better at directional prediction than the Persistence baseline.</p>","PeriodicalId":48739,"journal":{"name":"International Journal of Health Geographics","volume":" ","pages":""},"PeriodicalIF":4.1,"publicationDate":"2026-05-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13255379/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148144944","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Spatial pattern evolution and regional interaction mechanisms of sports and recreation facilities in jiangsu province, China: a gis-based spatial econometric analysis.","authors":"Xin Lyu, Yanling Li, Yuming Tai","doi":"10.1186/s12942-026-00472-8","DOIUrl":"10.1186/s12942-026-00472-8","url":null,"abstract":"<p><p>The equitable allocation of public service facilities is a key issue in regional coordinated development; however, the long-term spatiotemporal dynamics of their spatial distribution and interregional interaction mechanisms remain underexplored. Based on point-of-interest (POI) data from 13 prefecture-level cities in Jiangsu Province, this study employs exploratory spatial data analysis (ESDA) methods and the spatial Durbin model to investigate the evolutionary patterns and driving mechanisms of sports and recreation facility allocation from 2014 to 2023. The findings reveal that the distribution of sports and recreation facilities in Jiangsu exhibits a stable core-periphery structure, yet limited diffusion is evident, characterized by the sprawling of core areas, the formation of corridors along the Yangtze River, and the emergence of growth poles in peripheral regions. The spatial pattern is undergoing a balanced reconstruction from linear agglomeration to a \"from-line-to-area\" configuration. Regarding driving mechanisms, the spatial externalities of different factors exhibit significant differentiation: economic growth generates positive synergistic spillovers, while population density, industrial agglomeration, and road density manifest as negative spillovers through siphon effects, crowding-out effects, and consumption substitution, respectively. This dynamic structure helps explain the underlying mechanisms sustaining the highly stable macro pattern. This study is an attempt to reveal the duality of spatial spillover effects of public service facilities with a progressive analytical paradigm. The empirical evidence from Jiangsu Province may provide useful references for the optimization of facility layout and spatial governance in other regions.</p>","PeriodicalId":48739,"journal":{"name":"International Journal of Health Geographics","volume":" ","pages":""},"PeriodicalIF":4.1,"publicationDate":"2026-05-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13479553/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148007065","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}