Vito Telesca

Acute otitis media (AOM) is a leading cause of pediatric Emergency Department visits, particularly among children under five years of age. Although its seasonal pattern is well established, the role of air pollution and meteorological factors remains unclear. This study aims to investigate their impact on daily AOM visits by integrating machine learning and epidemiological approaches. We conducted a retrospective analysis of pediatric AOM diagnoses (2007–2015) at S. Croce and Carle Hospital (Cuneo, Italy). Predictors included PM10, NO₂, O₃, and eleven meteorological variables. Ensemble machine learning models (Random Forest, XGBoost, and AdaBoost) were trained and validated using 10-fold cross-validation. Model interpretability was assessed through SHAP values. Distributed Lag Nonlinear Models (DLNM) were applied to estimate delayed exposure–response relationships over lag periods of 0–1, 0–3, 0–5, and 0–10 days, with results expressed as Relative Risks (RRs) and 95% Confidence Intervals (CIs). AdaBoost showed the best performance (R² = 0.974; MAE = 0.019 cases/day; cross-validated R² = 0.987). SHAP analysis identified mean temperature as the most influential predictor (44%), while PM10, NO₂, and O₃ each contributed approximately 10%. DLNM analysis confirmed a strong and consistent effect of temperature across all lag periods (RR > 1.20, CI > 1). Moderate associations were observed for NO₂ and PM10 (RR: 1.02–1.04). O₃ exhibited smaller but significant effects at shorter lags (RR = 1.01 at 0–1 days; RR = 1.02 at 0–3 days; CI > 1). Environmental factors, particularly temperature, play a significant role in pediatric AOM incidence. The integration of machine learning and DLNM enhances predictive accuracy and improves the understanding of exposure timing. These findings support the development of early warning systems and targeted preventive strategies under adverse environmental conditions. Further validation in larger urban settings is needed.

Flood damage assessment remains challenging, as conventional flood risk management mainly relies on hydraulic hazard maps that do not explicitly reproduce observed damage patterns. Recent advances in remote sensing and machine learning (ML) enable the integration of environmental and socio-economic data with historical impact information to improve flood damage modeling. This study proposes an explainable machine learning framework for flood damage susceptibility mapping, using observed institutional damage records from the 2011 and 2013 flood events combined with 17 geospatial flood risk factors (FRFs) representing hazard, exposure, and vulnerability. This approach enables the capture of non-linear relationships between flood damage and FRFs. For comparison purposes, the same framework was also applied using hydraulically modeled flood extents corresponding to return periods of 30, 200, and 500 years. The framework was tested along the Basilicata Ionian coast in southern Italy, a Mediterranean region characterized by complex geomorphology, intense rainfall events, and recurrent flood impacts. An eXtreme Gradient Boosting (XGBoost) model was trained using 17 FRFs related to hazard, exposure, and vulnerability at a spatial resolution of 20 m. The model achieved high performance with an accuracy of 0.988, an F1-score for the minority class of 0.860, and an ROC-AUC (test) of 0.996. High to very high flood damage probability was predicted in approximately 4.1% of the study area, mainly in low-lying floodplains near river corridors and infrastructure. SHAP-based explainability analysis revealed that damage susceptibility was predominantly driven by hazard and exposure factors: Drainage density (17.10%), Railway distance (16.33%), and Elevation (15.42%), extreme precipitation (Max rainfall, 10.66%) and Street distance (7.51%), with socio-economic vulnerability contributing less than 4%. The observed damage target exhibited clear threshold-like patterns (e.g., sharp risk increases below ~25/35 m elevation or within ~150/200 m of road infrastructure), contrasting with the smoother, continuous gradients produced by hydraulic scenarios. This analysis identified the most influential predictors and their response ranges. The proposed framework complements hydraulic hazard mapping by explicitly modeling observed flood damage, supporting flood risk assessment in flood-prone coastal regions.

This study presents an approach based on machine learning (ML) techniques to analyze the relationship between emergency room (ER) admissions for cardiorespiratory diseases (CRDs) and environmental factors. The aim of this study is the development and verification of an interpretable machine learning framework applied to environmental and health data to assess the relationship between environmental factors and daily emergency room admissions for cardiorespiratory diseases. The model’s predictive accuracy was evaluated by comparing simulated values with observed historical data, thereby identifying the most influential environmental variables and critical exposure thresholds. This approach supports public health surveillance and healthcare resource management optimization. The health and environmental data, collected through meteorological sensors and air quality monitoring stations, cover eleven years (2013–2023), including meteorological conditions and atmospheric pollutants. Four ML models were compared, with XGBoost showing the best predictive performance (R² = 0.901; MAE = 0.047). A 10-fold cross-validation was applied to improve reliability. Global model interpretability was assessed using SHAP, which highlighted that high levels of carbon monoxide and relative humidity, low atmospheric pressure, and mild temperatures are associated with an increase in CRD cases. The local analysis was further refined using LIME, whose application—followed by experimental verification—allowed for the identification of the critical thresholds beyond which a significant increase in the risk of hospital admission (above the 95th percentile) was observed: CO > 0.84 mg/m³, P_atm ≤ 1006.81 hPa, Tavg ≤ 17.19 °C, and RH > 70.33%. The findings emphasize the potential of interpretable ML models as tools for both epidemiological analysis and prevention support, offering a valuable framework for integrating environmental surveillance with healthcare planning.

Floods and landslides are two distinct natural phenomena influenced by different conditioning factors, though some environmental triggers may overlap. This study applied eXtreme Gradient Boosting (XGBoost) to develop susceptibility maps for both phenomena, using a unified approach based on the same geospatial predictors. The approach integrated topographical, geological, and remote sensing datasets. Flood event data were collected from institutional sources using multi-source and high-resolution remotely sensed data. The landslide inventory was compiled based on historical records and geomorphological analysis. Key conditioning factors such as elevation, slope, lithology, and land cover were analyzed to identify areas prone to floods and landslides. The methodology was applied to the Basento River basin in Southern Italy, a region frequently impacted by both hazards, to assess its vulnerability and inform risk management strategies. While flood susceptibility is primarily associated with low-lying areas near river networks, landslides are more influenced by steep slopes and geological instability. The XGBoost model achieved a classification accuracy close to 1 for flood-prone areas and 0.92 for landslide-prone areas. Results showed that flood susceptibility was primarily associated with low Elevation and Relative Elevation, and high Drainage Density, whereas landslide susceptibility was more influenced by a broader and balanced set of factors, including Elevation, Drainage Density, Relative Elevation, Distance and Lithology. The resulting susceptibility maps offered critical approaches for land use planning, emergency management, and risk mitigation. Overall, the results demonstrated the effectiveness of XGBoost in multi-hazard assessments, offering a scalable and transferable approach for similar at-risk regions worldwide.

Background: Cardiovascular diseases (CVD) remain the predominant global cause of mortality, with both low and high temperatures increasing CVD-related mortalities. Climate change impacts human health directly through temperature fluctuations and indirectly via factors like disease vectors. Elevated and reduced temperatures have been linked to increases in CVD-related hospitalizations and mortality, with various studies worldwide confirming the significant health implications of temperature variations and air pollution on cardiovascular outcomes. Methods: A database of daily Emergency Room admissions at the Giovanni XIII Polyclinic in Bari (Southern Italy) was developed, spanning from 2013 to 2019, including weather and air quality data. A Random Forest (RF) supervised machine learning model was used to simulate the trend of hospital admissions for CVD. The Seasonal and Trend decomposition using Loess (STL) decomposition model separated the trend component, while cross-validation techniques were employed to prevent overfitting. Model performance was assessed using specific metrics and error analysis. Additionally, the SHapley Additive exPlanations (SHAP) method, a feature importance technique within the eXplainable Artificial Intelligence (XAI) framework, was used to identify the feature importance. Results: An R2 of 0.97 and a Mean Absolute Error of 0.36 admissions were achieved by the model. Atmospheric pressure, minimum temperature, and carbon monoxide were found to collectively contribute about 74% to the model’s predictive power, with atmospheric pressure being the dominant factor at 37%. Conclusions: This research underscores the significant influence of weather-climate variables on cardiovascular diseases. The identified key climate factors provide a practical framework for policymakers and healthcare professionals to mitigate the adverse effects of climate change on CVD and devise preventive strategies.

The objective of this study was to determine the relationship between weather conditions and hospital admissions for cardiovascular diseases (CVD). The analysed data of CVD hospital admissions were part of the database of the Policlinico Giovanni XXIII of Bari (southern Italy) within a reference period of 4 years (2013–2016). CVD hospital admissions have been aggregated with daily meteorological recordings for the reference time interval. The decomposition of the time series allowed us to filter trend components; consequently, the non-linear exposure–response relationship between hospitalizations and meteo-climatic parameters was modelled with the application of a Distributed Lag Non-linear model (DLNM) without smoothing functions. The relevance of each meteorological variable in the simulation process was determined by means of machine learning feature importance technique. The study employed a Random Forest algorithm to identify the most representative features and their respective importance in predicting the phenomenon. As a result of the process, the mean temperature, maximum temperature, apparent temperature, and relative humidity have been determined to be the most suitable meteorological variables as the best variables for the process simulation. The study examined daily admissions to emergency rooms for cardiovascular diseases. Using a predictive analysis of the time series, an increase in the relative risk associated with colder temperatures was found between 8.3 °C and 10.3 °C. This increase occurred instantly and significantly 0–1 days after the event. The increase in hospitalizations for CVD has been shown to be correlated to high temperatures above 28.6 °C for lag day 5.

Meteorological conditions play a crucial role in air pollution by affecting both directly and indirectly the emissions, transport, formation, and deposition of air pollutants. Extreme weather events can strongly affect surface air quality. Understanding relations between air pollutant concentrations and extreme weather events is a fundamental step toward improving the knowledge of how excessive heat impacts on air quality. In this work, we developed a statistical procedure for investigating the variations in the correlation structure of four air pollutants (NO_x, O₃, PM₁₀, PM_2.5) during extreme temperature events measured in monitoring sites located of Emilia Romagna region, Northern Italy, in summer (June–August) from 2015 to 2017. For the selected stations, Hot Days (HDs) and Heat Waves (HWs) were identified with respect to historical series of maximum temperature measured for a 30-year period (1971–2000). This method, based on multivariate techniques, allowed us to highlight the variations in air quality of study area due to the occurrence of HWs. The examined data, including PM concentrations, show higher values, whereas NO_x and O₃ concentrations seem to be not influenced by HWs. This operative procedure can be easily exported in other geographical areas for studying effects of climate change on a local scale.