Chaos, Solitons & Fractals

With significant population fractions in many societies who refuse vaccines, it is important to reconsider how vaccination is incorporated into compartmental epidemiology models. It is still most common to apply the vaccination rate to the entire class of susceptibles, rather than to use the more realistic assumption that the vaccination rate function should depend only on the population of susceptibles who are willing and able to receive a vaccination. This study uses a simple generic disease model to address two questions: (1) How much error is introduced in key model outcomes by neglecting vaccine unwillingness?, and (2) Can the error be reduced by incorporating vaccine unwillingness into the vaccination rate constant rather than the rate diagram? The answers depend greatly on the time scale of interest. For the endemic time scale, where longterm behavior is studied with equilibrium point analysis, the error in neglecting unwillingess is large and cannot be improved upon by decreasing the vaccination rate constant. For the epidemic time scale, where the first big epidemic wave is studied with numerical simulations, the error can still be significant, particularly for diseases that are relatively less infectious and vaccination programs that are relatively slow.

Neurological health score (NHS), indicating the health of brain and nervous system, helps in identifying high risk individuals, and in recommending lifestyle modifications. In the present study, we developed NHS based on genetic, lifestyle and biochemical variables associated with eight neurological disorders - dementia, stroke, Parkinsons disease, amyotrophic lateral sclerosis, schizophrenia, bipolar disorder, multiple sclerosis and migraine. UK Biobank data from Caucasian individuals was used to develop the model, and the data from individuals of Indian ethnicity was used to validate the model. Logistic regression and XGBoost algorithms were used in selecting the significant variables for the disorders. NHS developed from the selected variables was found to be very significant after adjusting for age and sex (AUC:0.6, OR: 0.95). Higher NHS was associated with a lower risk of neurological disorders and better social well-being. Highest NHS group (top 25%) showed 1.3 times lower risk compared to the rest of the individuals. Results of our study help in developing a framework for quantifying the neurological health in clinical setting.

The COVID-19 pandemic has presented severe challenges in understanding and predicting the spread of infectious diseases, necessitating innovative approaches beyond traditional epidemiological models. This study introduces an advanced method for automated model discovery using the Sparse Identification of Nonlinear Dynamics (SINDy) algorithm, leveraging a dataset from the COVID-19 outbreak in Thuringia, Germany, encompassing over 400,000 patient records and vaccination data. By analysing this dataset, we develop a flexible, data-driven model that captures many aspects of the complex dynamics of the pandemics spread. Our approach incorporates external factors and interventions into the mathematical framework, leading to more accurate modelling of the pandemics behaviour. The fixed coefficient values of the differential equation as globally determined by the SINDy were not found to be accurate for locally modelling the measured data. We therefore refined our technique based on the differential equations as found by SINDy, by investigating three modifications that account for recent local data. In a first approach, we re-optimized the coefficient values using seven days of past data, without changing the globally determined differential equation. In a second approach, we allowed a temporal dependence of the coefficient values fitted using all previous data in combination with regularization. As a last method, we kept the coefficients fixed to the original values but augmented the differential equation with a small neural network, locally optimized to the data of the past week. Our findings reveal the critical role of vaccination and public health measures in the pandemics trajectory. The proposed model offers a robust tool for policymakers and health professionals to mitigate future outbreaks, providing insights into the efficacy of intervention strategies and vaccination campaigns. This study advances the understanding of COVID-19 dynamics and lays the groundwork for future research in epidemic modelling, emphasising the importance of adaptive, data-informed approaches in public health planning.

Current clinical risk stratification for thoracic aortic aneurysms (TAA) relies primarily on maximum diameter, which is a poor predictor of rupture. Recent fluid-structure interaction studies have identified a dimensionless "flutter instability parameter" (N{omega} ) that accurately classifies abnormal aortic growth. However, this parameter currently serves as a static diagnostic snapshot. In this work, we propose a proof-of-concept computational framework that links flutter instability to microstructural tissue damage via a coupled system of ordinary differential equations (ODEs). We model a feedback loop where flutter-induced energy dissipation drives elastin degradation and collagen remodeling, which in turn reduces wall stiffness and amplifies the instability. To address the challenge of unobservable tissue properties, we implement a Bayesian inference engine to infer model parameters. We demonstrate feasibility on a synthetic patient cohort calibrated to published clinical growth rates and diameters. Our results show that this approach can infer hidden damage parameters and capture the qualitative bifurcation between stabilizing remodeling and runaway aneurysm expansion. While validation on real patient data remains essential, this work establishes the mathematical foundation for transforming a static physiomarker into a personalized prognostic trajectory.

Visceral leishmaniasis (VL) is considerably more severe among individuals infected with human immunodeficiency virus (HIV), leading to higher parasite loads, frequent relapse, and increased mortality. To examine the epidemiological interaction between the two diseases, we develop a comprehensive VL-HIV co-infection model that incorporates transmission pathways, treatment effects, and relapse dynamics. The model is parameterized using real-time data from Bihar, India, including monthly VL-only and VL-HIV co-infected cases and annual HIV prevalence data. Our analysis shows that HIV infection drives the resurgence and persistence of VL even in settings where VL alone would not sustain transmission, underscoring the amplifying effect of HIV-induced immunosuppression on VL dynamics. We further demonstrate that increasing HIV treatment coverage substantially reduces co-infection prevalence and lowers VL relapse rates. Numerical simulations and optimal control analysis highlight the effectiveness of integrated intervention strategies that combine awareness, treatment enhancement, and vector control. Overall, this study emphasizes the need for coordinated VL and HIV control programs and provides data-driven guidance for designing sustainable intervention strategies in endemic regions.

ObjectiveThis study investigates whether incorporating physiological coupling concepts into neural network design can support stable and interpretable feature learning for histopathological image classification under limited data conditions. MethodsA physiologically inspired architecture, termed CardioPulmoNet, is introduced to model interacting feature streams analogous to pulmonary ventilation and cardiac perfusion. Local and global tissue features are integrated through bidirectional multi-head attention, while a homeostatic regularization term encourages balanced information exchange between streams. The model was evaluated on three histopathological datasets involving oral squamous cell carcinoma, oral submucous fibrosis, and heart failure. In addition to end-to-end training, learned representations were assessed using linear support vector machines to examine feature separability. ResultsCardioPulmoNet achieved performance comparable to several pretrained convolutional neural networks across the evaluated datasets. When combined with a linear classifier, improved classification performance and higher area under the receiver operating characteristic curve were observed, suggesting that the learned feature embeddings are well structured for downstream discrimination. ConclusionThese results indicate that physiologically motivated architectural constraints may contribute to stable and discriminative representation learning in computational pathology, particularly when training data are limited. The proposed framework provides a step toward integrating physiological modeling principles into medical image analysis and may support future development of transferable and interpretable learning systems for histopathological diagnosis.

AbstractsO_ST_ABSBackgroundC_ST_ABSAccurate dengue forecasting is vital for public health preparedness. Despite a surge in forecasting approaches, a quantitative ranking of the relative performance and practical utility of dengue forecasting is lacking. MethodsA systematic review and Network Meta-Analysis (NMA) of studies comparing dengue forecasting methods (2014-2024) was conducted. Models were categorised into five groups: Time Series, Deep Learning (DL), Machine Learning (excluding DL), Hybrid, and Ensembles. NMA was applied to the logarithm of the most common forecast error metric to rank relative performance--an "Implementability Score" quantified analyst and data requirements, and computational costs. Results59 studies were included. NMA of Root Mean Squared Error identified k-Nearest Neighbour (k-NN) models as achieving the highest predictive accuracy, followed closely by Vector Autoregression, Kalman Filtering, Generalised Linear Model and Autoregressive Neural Network (ARNN). While DL models showed high potential, they scored lowest in implementability due to poor interpretability and high data requirements. Most studies utilised meteorological covariates, with significant gaps in the use of socio-economic and entomological predictors. ConclusionsAlthough there was some trade-off between accuracy and implementability, traditional statistical models were often comparable in accuracy to machine learning approaches, with advantages in interpretability and data needs. Under-explored areas for future research include the use of ensemble models and the use of socio-economic and entomological data. RegistrationPROSPERO CRD420251016662. Author SummaryDengue is a critical global health threat affecting the worlds population. While many forecasting models exist to help officials prepare for outbreaks, there has been no standardised way to compare their performance. This leaves health experts in resource-limited areas uncertain about which tools are truly reliable or easy to use under their specific local conditions. We conducted a network meta-analysis of studies comparing dengue forecasting methods accuracy, grouping them into five categories: Machine Learning, Deep Learning, Time Series, Ensemble, and Hybrid. Beyond ranking their accuracy, we developed an "Implementability Score" to evaluate the practical feasibility of each model, accounting for technical complexity, data requirements, and software accessibility. Our analysis identified the top-performing models. Notably, traditional statistical models often performed as well as complex Deep Learning algorithms. While advanced models show potential, they are often difficult to implement or explain to decision-makers. There is no "one-size-fits-all" solution; the best model depends on capacity and data in each setting. This study provides a roadmap for public health officials to select tools that are both accurate and feasible.

Schistosomiasis is a neglected parasitic disease caused by various trematode species of the genus Schistosoma for which 251 million people needed treatment in 2021. Many mathematical models of Schistosoma mansoni transmission incorporate the effect of chemoprophylaxis on parasite burden within the human host. While praziquantel is the most commonly implemented pharmaceutical used to control schistosomiasis, due to its applicability over several species and its negligible side effects, it is not very effective against juvenile schistosomes in humans. This limited efficacy on the juvenile life-stage of the parasite may be an important factor in the persistence of the disease. The demographic consequences of praziquantel use on schistosome population age and sex composition within the human host may obfuscate the effectiveness of these chemoprophylactic control strategies. Furthermore, the effectiveness of this treatment is heavily dependent on the force of infection to humans and the frequency at which these pharmaceuticals are administered. Using a stochastic mechanistic model, we investigated the effects of inconsistent drug efficacy among parasite life stages, varying parasite population structure within the human host, and alternative treatment regimes to the prevailing once-yearly strategy. This allowed us to identify the reduction in infection prevalence under differing infection risk scenarios, parasite population structures at the time of treatment, and treatment schedules. Our results indicate that if elimination is the goal, then widespread (>75% of the population) treatment should be the target and that more frequent treatment schedules are useful up to several treatments a year.

BackgroundTuberculosis (TB) remains a major public health challenge in Nepal, with incidence rates substantially higher than global estimates. Accurate forecasting of TB incidence is essential for early warning systems, resource allocation, and targeted interventions. This study aimed to develop and validate a hybrid Seasonal Autoregressive Integrated Moving Average (SARIMA) and Convolutional Neural Network Auto-Regressive (CNNAR) model for TB incidence forecasting in Nepal. MethodsMonthly TB incidence data (January 2015 to December 2024) were obtained from the National Tuberculosis Control Center (NTCC), Nepal. A hybrid SARIMA-CNNAR model was developed, where SARIMA modeled linear seasonal trends and CNNAR captured nonlinear patterns in the residuals. Hyperparameters were optimized using grid search with 5-fold cross-validation. Model performance was evaluated using Mean Absolute Error (MAE), Root Mean Square Error (RMSE), Mean Absolute Percentage Error (MAPE), and R2 on the 2024 test set. Structural break analysis and sensitivity analysis assessed model robustness. The hybrid model was compared against standalone SARIMA, CNNAR, and three state-of-the-art benchmarks: Long Short-Term Memory (LSTM), Facebook Prophet, and XGBoost. ResultsTB incidence in Nepal increased from a monthly average of 2,048 cases in 2015 to 3,447 in 2024 (68.4% increase). The hybrid SARIMA-CNNAR model demonstrated strong performance with test set metrics of MAE=248.35, RMSE=294.31, MAPE=7.2%, and R2=0.79. Comparative performance: CNNAR (MAE=251.08, RMSE=336.55, MAPE=7.7%, R2=0.73); LSTM (MAE=267.91, RMSE=324.55, MAPE=7.5%, R2=0.75); XGBoost (MAE=314.74, RMSE=373.99, MAPE=8.5%, R2=0.66); Prophet (MAE=371.15, RMSE=478.40, MAPE=10.4%, R2=0.45); SARIMA (MAE=401.11, RMSE=503.93, MAPE=10.99%, R2=0.39). All models captured seasonal peaks in March-May and July-August, with forecasts for 2025 indicating continued seasonal patterns. Sensitivity analysis confirmed robustness with <5% metric variation across parameter configurations. ConclusionsThis first validated hybrid model for TB prediction in Nepal demonstrates high forecasting accuracy by integrating linear seasonal modeling with nonlinear pattern detection. The approach offers a robust tool for evidence-based public health planning in resource-limited settings and it is suitable for integration into national surveillance systems. Author SummaryTuberculosis remains a major public health challenge in Nepal, with cases increasing substantially over the past-decade. In this study, we developed a computer model that combines two different forecasting ap proaches: one that captures regular seasonal patterns and another that learns complex trends from data to predict monthly TB cases. Using ten years of national surveillance data, our hybrid model achieved high accuracy in forecasting TB incidence, outperforming standard approaches including SARIMA, PROPHET, CNNAR, LSTM neural networks, and XGBoost. The model successfully predicted seasonal peaks in March-May and July-August, with forecasts for 2025 suggesting continued high case numbers. These predictions can help Nepals health authorities prepare by pre-positioning diagnostic supplies, scheduling additional staffs during peak months, and targeting awareness campaigns. The modeling approach is desig ned to be adaptable for other diseases and countries with similar health data.

BackgroundThe growing interest in applying artificial intelligence in personalized nutrition is challenged by the complex nature of dietary advice that must balance health, economic, and personal factors. Though automated solutions using either Linear Programming (LP) or Large Language Models (LLMs) already exist, they have significant drawbacks. LP often lacks personalization, whereas LLMs can be unreliable for precise calculations. ObjectivesTo develop and assess a model that integrates a Mixed Integer Linear Programming (MILP) solver with an LLM to generate personalized meal plans and compare it with standalone LLM and MILP models. MethodsThe proposed hybrid MILP+LLM model first uses an LLM (GPT-4o) to filter a unified food dataset (n=297), which combines regional Central Asian and global food items, according to the users profile. The filtered list of food items is then received by a MILP solver which identifies the set of top 10 optimal solutions. Finally, given this set of solutions, LLM chooses the most appropriate meal plan. The model was evaluated using five synthesized, clinically complex patient profiles sourced from Adilmetova et al. [4]. The performance of this hybrid model was compared against standalone MILP and LLM using 5-point Likert scale with Kruskal-Wallis and post hoc Dunns tests for Nutrient Accuracy, Personalization, Practicality, and Variety. ResultsFindings demonstrated that the proposed MILP+LLM model reached balanced performance achieving scores of more than 3.6 points in all criteria, with high scores in Nutrient Accuracy (3.96), Personalization (3.81), and Practicality (3.99). The standalone LLM model performed the weakest in all criteria, with statistically significant lower scores compared to the other two methods. The standalone MILP model performed best in Nutrient Accuracy (4.93) and in Variety (4.10) but lagged behind the MILP+LLM model in Practicality and Personalization. Kruskal-Wallis and Dunns tests showed MILP and MILP+LLM outperformed LLM across all criteria. MILP was more accurate (p<0.0001), while MILP+LLM model was more practical (p=0.021). ConclusionsThe findings suggest that integrating the LLM with the MILP solver creates a model that combines qualitative personalization with quantitative precision. This model produces comprehensive, reliable meal plans, addressing the limitations of using either model alone.

The COVID-19 pandemic exposed major vulnerabilities of hospital capacity and management worldwide, particularly in intensive care units (ICUs) and emergency rooms (ER), imposing prompt adaptation and resource reallocation. Although SARS-CoV-2 is no longer endangering healthcare systems, winter seasons continue to bring recurrent overload of critical care services, primarily due to respiratory infections. In France e.g., this pattern led to the reactivation of the national emergency response plan during the 2024-2025 seasonal influenza peak, highlighting the continuous need for improved predictive tools. However, forecasting hospitalization surges at a local scale remains a methodological challenge because the (very) low incidence numbers are subject to strong stochasticity and therefore require additional input of information and dedicated approaches. This study investigates the potential for early forecasting of respiratory infection peaks by analyzing ER visit trends. By clustering all-cause ER visits during the 2023-2025 winter seasons from the Nimes University Hospital (France), we identified a strong temporal correlation between early pediatric hospitalizations ([&le;]5 years old) and the following weeks adult hospitalization incidence for respiratory infections. The results suggest that tracking hospital admissions of pediatric ER visits, even without hospital care needs, can serve as a valuable early warning signal for upcoming peaks in respiratory-related hospitalizations. This predictive approach could improve hospital preparedness and resource management during seasonal influenza outbreaks. Author summaryThe epidemics of respiratory viruses present a significant challenge to hospitals in the temperate zone on an annual basis. Frequently, the hospital overload is mitigated by the late reactive allocation of human and material resources that are, hence, suboptimal. This study proposes a statistical framework to assist hospitals in anticipating bed requirements during seasonal influenza waves, despite high noise at the local level, by enhancing hospitalization forecasting with emergency room (ER) visit data. The prediction of the adult epidemic peak is possible through the analysis of the respiratory pediatric ER visits, which facilitates hospital management.

Thyroid carcinoma is one of the most prevalent endocrine malignancies worldwide, and accurate preoperative differentiation between benign and malignant thyroid nodules remains clinically challenging. Diagnostic methods that medical practitioners use at present depend on their personal judgment to evaluate both imaging results and separate clinical tests, which creates inconsistency that leads to incorrect medical evaluations. The combination of radiological imaging with clinical information systems enables healthcare providers to enhance their capacity to make reliable predictions about patient outcomes while improving their decision-making abilities. The study introduces a deep learning framework that utilizes multiple data sources by combining magnetic resonance imaging (MRI) data with clinical text to predict thyroid cancer. The system uses a Vision Transformer (ViT) to obtain advanced MRI scan features, while a domain-adapted language model processes clinical documents that contain patient medical history and symptoms and laboratory results. The cross-modal attention system enables the system to merge imaging data with textual information from different sources, which helps to identify how the two types of data are interconnected. The system uses a classification layer to classify the fused features, which allows it to determine the probability of cancerous tumors. The experimental results show that the proposed multimodal system achieves better results than the unimodal base systems because it has higher accuracy, sensitivity, specificity, and AUC values, which help medical personnel to make better preoperative decisions.

Sodium (23Na) magnetic resonance imaging (MRI) provides valuable metabolic information, but it is limited by a low signal-to-noise ratio (SNR) and long acquisition times. To overcome these challenges, we present a Deep Image Prior (DIP)-based framework that combines anatomically guided proton (1H) MRI and metabolically guided 23Na MRI denoising via a fused proton-sodium prior within a directional total variation (dTV) regularization scheme. The DIP-Fusion approach minimizes a variational loss function combining data fidelity, fused dTV regularization, gradient consistency, and bias-field correction to reconstruct sodium images. MRI data were acquired from healthy volunteers and breast cancer patients. Healthy datasets were retrospectively undersampled at multiple factors, and fully sampled scans served as the ground truth. Patient datasets acquired for clinical purposes were reconstructed using the baseline DIP and the proposed DIP-Fusion methods. Sodium images were reconstructed using sum-of-squares (SoS) and adaptive combined (ADC) coil combination methods. We evaluated reconstruction performance using quantitative image quality metrics, including peak signal-to-noise ratio (PSNR), structural similarity index measure (SSIM), mean squared error (MSE), learned perceptual image patch similarity (LPIPS), feature similarity index (FSIM), and Laplacian focus. In healthy volunteers, DIP-Fusion outperformed state-of-the-art reconstruction techniques across all undersampling factors. In patient datasets, DIP-Fusion demonstrated superior performance compared with baseline DIP, achieving improved structural fidelity and sodium-specific signal preservation. These results demonstrate the potential for robust, highquality sodium MRI reconstruction under accelerated acquisition, which could lead to reduced scan times and enhanced clinical feasibility.

Loneliness and psychological well-being are increasingly recognized as critical public health concerns, yet their multi factorial determinants remain poorly understood. Traditional research often examines demographic, lifestyle, or social variables in isolation, yielding fragmented insights that overlook complex psychosocial interactions. In this study, we leverage a rich behavioral and psychological dataset from the Human Phenotype Project (HPP) to examine how lifestyle behaviors, social health indicators, and demographic characteristics collectively influence mental health outcomes. Employing advanced machine learning (ML) methods, including feature engineered representations, classical predictive models, and Large Language Model (LLM) classifiers, we identify latent psychosocial patterns associated with loneliness and psychological symptoms. Our approach combines predictive performance with interpretability, enabling the identification of key drivers of well-being across heterogeneous populations. Results indicate that certain lifestyle and social engagement factors consistently correlate with lower loneliness and improved psychological health, while other influences are context-dependent. This work demonstrates the potential of integrating computational modeling with psychological theory to reveal complex, multidimensional determinants of mental health, offering insights for targeted interventions, digital health applications, and evidence-based public health strategies.

Data scarcity and stylistic heterogeneity pose major challenges for emotion intensity classification. This paper presents a cross-dataset augmentation framework that leverages prompt-conditioned generative models alongside deterministic and heuristic transformations to synthesize target-style examples for improved transfer learning. We introduce a unified taxonomy of augmentation strategies--Heuristic Lexical Perturbation (HLA), Prompt-Conditioned Generative Augmentation (CGA), Sequential Hybrid Pipeline (SHA), Rule-Guided Style Adaptation (DSGA), and Enhanced Hybrid Augmentation (EHA)--and detail an interpretability-oriented prompt engineering approach that conditions LLMs on authentic target exemplars and stylistic features extracted from the target dataset. Augmented datasets were evaluated using multi-dimensional quality metrics (transformation quality, stylistic consistency, BLEU/CHRF, Self-BLEU, uniqueness) and downstream classification via a two-phase BERT-LSTM training with rigorous statistical testing. During source dataset pretraining and subsequent target dataset fine-tuning, CGA achieved the highest single-method gains in F1 and accuracy (F1 = 0.8816; accuracy = 0.8819, 95% CI recalculated). HLA and SHA exhibited improved cross-domain stability, suggesting stronger domain-generalizable features. We observe systematic trade-offs between fluency, lexical diversity, and emotion fidelity: high surface similarity often correlates with classifier performance but does not fully capture affective authenticity. We discuss methodological pitfalls, propose best practices for emotion-aware augmentation, and provide reproducible artifacts (prompts, example transformations, evaluation scripts) to facilitate further research in affective NLP.

Nipah virus (NiV) is a sporadic yet extremely deadly zoonotic pathogen, with reported case fatality rates of 40%-75% in impacted areas. Prolonged incubation, documented relapse, and delayed-onset encephalitis following apparent recovery indicate that NiV dynamics are influenced by intricate temporal processes. However, mechanistic contributions of these processes to epidemic persistence remain poorly understood. In this study, we develop and analyze a delay differential equation model for NiV transmission that explicitly incorporates incubation delay, relapse, and post-recovery delay effects. We compute a primary-transmission reproduction threshold (R0), characterize the disease-free and endemic equilibria, and analyze their stability, including delay-induced Hopf bifurcations. We show that relapse modifies the endemic-equilibrium existence condition, so an endemic equilibrium is not determined solely by the classical threshold criterion R0 = 1. We calibrate the model to NiV incidence data from Bangladesh (2001-2024) and perform simulations and sensitivity analyses to evaluate the effects of relapse and delays across epidemiological scenarios. Results indicate that sustained oscillations occur only under hypothetical parameter regimes, suggesting that delay-induced periodic outbreaks are unlikely under empirically informed conditions. Scenario analyses demonstrate that relapse and encephalitis-related delays predominantly influence post-peak dynamics, while incubation delay alters the time and intensity of the epidemic peak. We also introduce a relapse-driven replenishment fraction to quantify contribution of relapse to continued transmission, demonstrating its growing significance following the first outbreak peak. Overall, our results identify relapse as a key mechanism for epidemic persistence and underscore the importance of incorporating relapse and biological time delays into epidemiological modeling and public health strategies.

Non-alcoholic fatty liver disease (NAFLD) is a globally prevalent hepatic condition caused by the buildup of fat in the liver. It is frequently associated with metabolic comorbidities such as hypertension, cardiovascular disease (CVD), and prediabetes. However, early detection remains challenging due to the asymptomatic progression, and existing primary diagnostic methods, such as imaging or liver biopsy, are often expensive and inaccessible in rural areas. This study proposes a two-stage, interpretable machine learning pipeline for the non-invasive and cost-effective prediction of NAFLD and its key comorbidities using routine clinical parameters. The NAFLD prediction model was developed using the XGBoost algorithm, trained on a hybrid dataset that combines real patient data with rule-based synthetic data generated by simulating clinically plausible cases. Upon NAFLD-positive prediction, three separate XGB models, trained on data labelled based on thresholds, assess individual risks for hypertension, cardiovascular disease, and prediabetes. Explainability is obtained using SHAP (SHapley Additive exPlanations), which provides insight into feature relevance, while biomarker radar plots help in the visual interpretation of comorbidities. A user-friendly Streamlit interface enables real-time interaction with the tool for potential clinical application. The NAFLD model demonstrated robust performance, while the models used for predicting comorbidities achieved perfect performance, which may be a reflection of the limited dataset size used in the second stage. This work underscores the potential of AI-driven tools in NAFLD diagnosis, particularly when combined with explainable AI methods.

Wastewater is increasingly being recognized as an important data stream that can contribute to infectious disease surveillance and forecasting. With this recognition, a growing number of statistical inference approaches are being developed to use wastewater data to provide quantitative insights into epidemiological dynamics. However, few existing approaches have allowed for systematic integration of data streams for inference, for example by combining case incidence data and/or serological data with wastewater data. Furthermore, only a subset of existing approaches have been able to handle missing data without imputation and to handle datasets with different sampling times or intervals. Here, we develop a statistically rigorous, yet lightweight, approach to infer and forecast time-varying effective reproduction numbers (Rt values) using longitudinal wastewater virus concentrations either alone or jointly with additional data streams including case incidence data and serological data. Our approach relies on a state-space modeling approach for inference and forecasting, within the context of a simple bootstrap particle filter. We first describe the structure of our underlying disease transmission process model as well as our observation models. Using a mock dataset, we then show that Rt can be accurately estimated by interfacing this model with case incidence data, wastewater data, or a combination of these two data streams using the bootstrap particle filter. Of note, we show that these data streams alone do not allow for reconstruction of underlying infection dynamics due to structural parameter unidentifiability. We then apply our particle filter to a previously analyzed SARS-CoV-2 dataset from Zurich that includes case data and wastewater data. Our analyses of these real-world datasets indicate that incorporation of process noise (in the form of environmental stochasticity) into the state space model greatly improves our ability to reconstruct the latent variables of the model. We further show that underlying infection dynamics can be made identifiable through the incorporation of serological data and that the bootstrap particle filter can be used to make forecasts of Rt, case incidence, and wastewater virus concentrations. We hope that the inference approach presented here will lead to greater reliance on wastewater data for disease surveillance and forecasting that will aid public health practitioners in responding to infectious disease threats.

We describe a fast, noninvasive, low-cost survey method designed to understand the mode of transmission of an emerging pathogen. It is inspired from the standard household prevalence survey consisting in sampling households and counting the total number of people infected in each household, but refines it with the aim of improving diagnosis and estimating more parameters of the model of intra-household transmission. The survey was carried out in May-June 2020, during part of the first national French lockdown and received responses from more than 6,000 households involving a total of 20,000 people. We explain how we conceived the questionnaire, how we disseminated it, to the public through an open website hosted by CNRS, marketed through media and social media, and to a socially representative panel hosted by two survey institutes (BVA, Bilendi). We used the data obtained from the representative panel to correct for sampling biases in the CNRS survey using a classical raking procedure. Our results indicate that raking correctly canceled statistical biases between the two populations. We obtain the empirical distribution in households of the number and nature of symptoms. The main factors affecting the presence of symptoms are age, gender, body mass index (BMI), household size, but not necessarily in the expected direction. Our study shows that combining self-reporting and representative surveys allows investigators to obtain information on prevalence and household transmission mechanisms on emerging diseases at low cost.

Malaria remains a major public health challenge in Nigeria, and increasing climate variability poses substantial threats to recent gains in control. However, malaria transmission does not respond uniformly to climate drivers across epidemiological settings, highlighting the need to explore climate-malaria dynamics within heterogeneous contexts. This study examined the non-stationary temporal dynamics of malaria incidence and two key climatic drivers--rainfall and temperature--in Lagos and Zamfara states. These states were selected to represent heterogeneous transmission intensities, urbanisation and climatic regimes. Monthly malaria incidence and corresponding climate data (2015-2024) were analysed using wavelet-based model to characterise the non-stationary periodicities, quantify time-varying climate-malaria associations and identify time-dependent lead-lag relationships. Malaria incidence exhibited transient semi-annual, annual, and multi-annual cycles that were weak and temporally localized, despite persistent annual cycles in rainfall and temperature in Lagos. Cross-wavelet spectra revealed intermittent associations within the 8-16-month band, while phase analysis indicated short-lived alignment in which malaria incidence lagged rainfall by approximately one month, particularly between 2019 and 2022. The relationship with temperature was unstable, suggesting rainfall exerted more consistent influence on malaria incidence. In contrast, Zamfara displayed strong and dominant annual cycles of malaria incidence throughout the study period, with rainfall and temperature showing stable, statistically significant annual co-variability. Phase analysis revealed malaria incidence lagged rainfall by approximately one month and temperature by approximately three to four months, consistent with climate-modulated transmission processes. These findings highlight the heterogeneity of climate-malaria dynamics across transmission settings with contrasting epidemiological implications within Nigeria. The observed lag structures provide a basis for climate-informed early warning systems and intervention timing. While non-climatic drivers were not explicitly modelled, the analysis focuses on isolating climate-driven temporal signals. Consequently, to sustain control and elimination progress, climate-adaptive surveillance and region-specific interventions that anticipate rainfall- and temperature-driven transmission cycles must be integrated into Nigerias malaria control framework to ensure timely, targeted, and climate-resilient public health responses. Author summaryMalaria transmission does not respond uniformly to climate drivers across epidemiological settings, highlighting the need to explore climate-malaria dynamics within heterogeneous contexts. Identical climatic forcing can produce qualitatively different outcomes depending on the underlying epidemiological setting, indicating the limitations of generalising control efforts from a single context. Motivated by the need to understand these differences, in this study, we examined the cross-epidemiological scale-dependent and lag-specific climatic forcing of malaria transmission at the sub-national context, providing support for malaria control and elimination strategies. We addressed the following questions to understand the hidden patterns of the temporal cycles and the corresponding associations between the climate variables and malaria incidence in the two states: O_LIWhat are the dominant temporal cycles in malaria incidence in the study region? C_LIO_LIHow do the periodicities of climate variables compare with those of malaria incidence? C_LIO_LIAre there significant time-dependent associations between climate variability and malaria incidence? C_LIO_LIHow do these association vary across different time scales (intra-annual vs interannual) and periods? C_LIO_LIWhat is the average lag between changes in key climate variables and malaria incidence? C_LI Monthly malaria incidence data and corresponding rainfall and temperature records spanning 2015-2024 were analysed using a continuous wavelet transform (CWT) framework. Scale-specific periodicities were identified using wavelet power spectra, while climate-malaria associations were quantified using cross-wavelet power and wavelet transform coherence (WTC). Phase difference analysis was employed to characterise time-varying lead-lag relationships between malaria incidence and climatic drivers at the annual timescale. Results show that in Lagos, malaria incidence is irregular and weakly linked to climate, reflecting the impact of interventions and socio-environmental factors that disrupt transmission. In contrast, Zamfara exhibits strong, regular annual cycles tightly coupled to rainfall and temperature, with malaria incidence lagging rainfall by about one month and temperature by three to four months. These findings highlight the need for region-specific strategies: sustaining intervention-driven disruption in low-burden urban areas, and intensifying climate-adaptive measures in high-burden rural settings. Integrating climate-sensitive surveillance and tailored intervention timing into Nigerias malaria control framework will strengthen resilience and accelerate progress toward elimination. Specifically, our findings demonstrate evidence-based framework to guide climate-adaptive intervention timing. In Zamfara state, extreme heat between March and May as shown in the temperature profile, may reduce use of LLINs, indicating that mass distribution before and during these periods, within same year, may be less effective. The start of rain comes with a cooling effect which may facilitates good weather condition that encourages LLIN utilization. Correspondingly, LLIN distribution campaigns conducted in June or July, prior to peak rainfall and peak malaria incidence typically observed between August and October, may enhance intervention effectiveness. Coupled with other climate-sensitive control interventions (for example, seasonal malaria chemo-prevention), such campaigns should be repeated at intervals of no more than three years, in alignment with the observed multi-annual cycles of malaria incidence, to effectively mask malaria risk in Zamafara state. This implementation strategy could be employed in other high transmission states of Nigeria to mitigate malaria risk.