Chaos, Solitons & Fractals

Cervical cancer remains a significant cause of mortality and economic burden, particularly in developing countries with low rates of human papillomavirus (HPV) vaccination and screening. To address this, we present a mathematical model for controlling cervical cancer by integrating strategic HPV vaccination, screening and treatment. The population is divided into seven compartment: susceptible, vaccinated, infected with HPV, screened, cervical cancer, under treatment, and recovered. The models well-posedness is first established by proving the boundedness and non-negativity of solutions, ensuring biological relevance. The basic reproduction number R0 is computed using the next-generation matrix. The local and global stability of the disease-free equilibrium is analysed using the Jacobian matrix and Lyapunov function respectively. Furthermore, bifurcation analysis is performed using the Castillo-Chavez and Song theorem and sensitivity analysis is conducted on key parameters to identify their influence on disease dynamics. Numerical simulations of the model supports the analytical results. The findings of the study indicate that if the reproduction number is less than one, the solution converges to the disease-free state, signifying the asymptotic stability of the HPV-Cervical cancer free steady state. Crucially, the model demonstrates that increasing vaccination, screening and treatment rates significantly reduces HPV and cervical cancer incidence. This study underscores the value of mathematical modeling in informing the public health policy and provides a framework for optimizing control measures against HPV and Cervical cancer.

To characterize tuberculosis transmission and assess the impact of important interventions, a data-driven SEITR TB model is created. The potential for disease persistence has been calculated using the basic reproduction number. To determine the factors most significantly affecting the spread of tuberculosis, stability and sensitivity analyses are conducted. Strengthened treatment measures and optimized distancing significantly lower infection levels, according to numerical simulations. The Least Squares Fitting technique is used to validate real epidemiological data with a model solution. And the results emphasize that the best combinations of social distancing and treatment not only reduce the number of infections but also provide a cost-effective strategy for public health planning. Additionally, two numerical techniques, namely Pearson correlation and Partial Rank Correlation Coefficients (PRCC), are utilized to assess the sensitivity of model parameters. It is noted that the outcomes of these two methods are in agreeable comparison with one another regarding sensitivity analysis.

This study presents a theoretical and mathematical framework for understanding the dynamical behavior of infectious disease spread using a compartmental modeling approach. The proposed model incorporates memory effects to capture temporal dependencies that are not adequately represented by classical formulations. Qualitative analysis is employed to investigate the stability properties of the system and the role of key mechanisms in shaping long term dynamics. Publicly available surveillance information is used only to illustrate the consistency of the model behavior with observed trends. The results highlight the value of memory based modeling structures for describing complex biological processes and provide a general mathematical perspective for studying epidemic dynamics.

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.

Monkeypox viral disease has been and continues to be a global public health concern. Currently, there are existing, though minimal measures to manage mpox and any future outbreaks. Relying on data-driven modeling for early detection of mpox and prediction of possible cases and deaths in the presence of an outbreak is thus imperative. The present study forecasted global mpox virus cases and deaths in Asia, Africa, Australia, Europe, North America, Oceania, and South America. Three forecasting models (deep neural network, gradient boosting, and polynomial regression) were trained on data from the seven geographical regions. The performance of the three models was assessed using coefficient of determination, mean squared error, root mean squared error, and mean absolute scaled error across each region. Prediction using the deep neural network revealed a potential of higher mpox deaths in Africa and higher mpox cases in South America. Prediction using gradient boosting showed a potential of mpox deaths in Africa and higher mpox cases in Asia and North America. Prediction using polynomial regression revealed a potential of higher mpox deaths in Africa and Asia while rapid rises in mpox cases from 2025 to 2028 were anticipated in all regions except Asia in case of a monkeypox outbreak. For the three models, the tree-based ML model (gradient boosting) outperformed the statistical model and deep learning model by R2 and MSE in predicting mpox case counts across all the seven geographical regions. This study showcases the worth in using data-driven modelling to predict emerging and re-emerging infectious diseases such as mpox.

Lassa Fever control remains a daunting task for authorities in poorly resourced settings where the costs of implementing the control strategies remain high as the disease has multiple hosts and environmental spread. An important metric based on community pathogen load may be useful in estimating the level of control over all in the community in order to budget for the costs of control effectively. We developed a model that accounts for the community contribution of Lassa viral load in humans, rodents as well as the environment accounting for Community Pathogen Load incorporating three control strategies. The model was calibrated and fitted to the Nigerian data and optimized to establish the most cost-effective strategy using cost-effective analysis. Our results suggested that targeting the human community pathogen load remains an important control focus but the control of rodent contribution was equally important. Overall, the combination of three control strategies was the best control measure that is cost-effective for curbing Lassa fever in the population.

This paper looked at the exploration of Lassa fever transmission dynamics through stochastic models which yielded valuable insights into the interplay of factors influencing the probability of extinction and persistence of the virus within a population. By embracing the inherent randomness and variability in the system, the model provided a more realistic representation of the complex ecological and epidemiological dynamics of Lassa fever. We developed the deterministic model using a system of ordinary differential equations and the stochastic model using the Continuous Time Markov Chain. The probability of extinction and persistence underscored the need for a proactive and flexible approach to public health management. Our study revealed that introducing Lassa virus at the onset of an epidemic through various routes affects the likelihood of pathogen extinction. The presence of multiple infection routes increased the probability of pathogen persistence, highlighting complex transmission dynamics. Variations in contact rates, particularly between susceptible rodents and the environment community pathogen load, play a crucial role in influencing pathogen dynamics. This interconnected nature of transmission pathways underscores the factors governing Lassa virus persistence or extinction in a population, providing valuable insights for targeted management and control strategies for Lassa fever.

Influenza forecasting in (sub-)tropical regions remains understudied due to year-round, irregular transmission patterns. Further, the variation in seasonality and transmission characteristic of influenza in post-COVID-19 pandemic could be attributed to various drivers to quantify for better understanding. To address this issue, this study introduced an ensemble forecasting approach that incorporates varied dataset lengths to forecast influenza activity in Hong Kong, integrating multi-stream surveillance data, including absolute humidity, temperature, ozone, and school closures/holidays. We applied temporal cross-validation to evaluate forecasting performance for short- and long term separately across different training-sets and model variants, ultimately constructing ensemble forecasts weighted by individual model performance. The optimal ensemble model could forecast the 2019/20 winter influenza season onwards and evaluate the impact of COVID-19 public health and social measures (PHSMs). We further extended the framework to forecast influenza in post-pandemic period since March 2023, accounting for the impact of cessation of PHSMs and COVID-19-induced cross-protection/competition in population susceptibility. Forecasts showed two peaks in 2019/20 season, which could account for 95.2% (95% prediction interval (PI): 89.1%, 98.3%) reduction in attack rate for COVID-19 PHSMs. The post-pandemic forecasts indicated changes in influenza transmission dynamics and seasonality, highlighting the need to consider factors such as population immunity and co-circulation with COVID-19 in future influenza forecasts. This study emphasizes the importance of incorporating diverse factors for better influenza forecasts in (sub-)tropical regions. The proposed framework offers a scalable tool for forecasting other respiratory virus transmissions, supporting healthcare agencies in managing future infection burdens and enhancing preparedness. Author summaryReliable and proactive forecasts of influenza activity and timing of epidemic outcomes enable public health officials to plan targeted responses. However, unlike temperate locations, the irregular seasonality of influenza in tropical/subtropical locations leads to highly variable forecasting patterns when models use varying lengths of historical data, reducing the robustness of forecasts. By leveraging multi-stream surveillance data in Hong Kong, we developed a mechanistic model-based ensemble forecasting framework that integrate potential combinations of data and models for short-, medium-, and long-term forecasts of influenza outcomes. Beyond methodological advancement, this framework has broader implications in assessing the impact of COVID-19-related interventions on influenza dynamics during pandemic and evaluating potential co-circulation risk of respiratory viruses including influenza and COVID-19 in post-pandemic era.

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.

Polygenic risk scores (PRSs) quantify an individuals genetic susceptibility to complex traits and diseases. Conventional PRSs, which are based on linear models, perform poorly for phenotypes with skewed distributions or with genetic effects that vary across the distribution. We propose a quantile regression-based PRS (QPRS) that can capture quantile-specific genetic effects. While existing PRSs provide only a single score, QPRS models genetic influences at multiple quantiles of the phenotype, thereby enhancing predictive performance by utilizing these multiple scores as covariates. We evaluate the performance of our method through both simulations and a real-data application. In simulations, QPRS significantly reduces the mean squared error (MSE) compared to the conventional PRS, both in the presence of variance quantitative trait loci and outliers. For real data analysis, we use data from Korea Genome and Epidemiology Study (KoGES) to evaluate predictive performance. We consider two prediction tasks: a continuous outcome (glucose level) and a binary outcome (diabetes status, derived from glucose level). For glucose-level prediction, the model incorporating QPRS achieves a R2 value 4.69 times higher than the model using conventional PRSs. For predicting diabetes status, the model with QPRS produces an area under the curve 1.06 times higher than the model with conventional PRSs.

Burn injuries are a significant concern in developing countries due to limited infrastructure, and treating them remains a major challenge. The manual assessment of burn severity is subjective and depends, to a large extent, on individual expertise. Artificial intelligence can automate this task with greater accuracy and improved predictions, which can assist healthcare professionals in making more informed decisions while triaging burn injuries. This study established a model pipeline for detecting burn injuries in images using multiple deep learning models, including U-Net, DenseNet, ResNet, VGG, EfficientNet, and transfer learning with the Segment Anything Model2 (SAM2). The problem statement was divided into two stages: 1) removing the background and 2) burn skin segmentation. ResNet50, used as an encoder with a U-Net decoder, performs better for the background removal task, achieving an accuracy of 0.9757 and an intersection over union (Jaccard index) of 0.9480. DenseNet169, used as an encoder with a U-Net decoder, performs well in burn skin segmentation, achieving an accuracy of 0.9662 and an intersection over Union of 0.8504. The dataset collected during the project is available for download to facilitate further research and advancements (Link to dataset: https://geninfo.iitr.ac.in/projects). TBSA was estimated from predicted burn masks using scale-based calibration

Age related macular degeneration is known to be one of the major causes of irreversible blindness among the older generation. We present a mathematical model of partial differential equations for the therapy of this disease, which is based on the intravitreal injection of a drug into the vitreous body. For the treatment to work, the drug has to travel past the inner-limiting membrane into the retina and reduce the free vascular endothelial growth factor (VEGF) concentration by binding to at least one of the two binding sites of the VEGF molecule. Therefore, our model consists of two compartments, the vitreous and the retina. In the vitreous we employ four coupled convection-diffusion-reaction equations with an additional coupling to the underlying aqueous humor flow and four coupled diffusion-reaction equations in the retina. The resulting PDE system is solved numerically in a realistic 3D eye geometry. Temporal discretization is based on one-step theta schemes and spatial discretization is done using the Finite Element method. The numerical results are used to demonstrate the therapy concept and to analyze the drug efficacy of aflibercept and ranibizumab. The results show, among other things, that only about 20 % of the drug reaches the retina through the inner-limiting membrane and that 50 % of the VEGF concentration has been rebuilt in the retina after 38.19 days for a single ranibizumab injection.

This study introduces a novel transformer-based ensemble framework for the multi-label detection of mental health disorders from social media posts. Unlike traditional multi-class approaches that often struggle with comorbidity, the proposed method employs a binary relevance strategy using fine-tuned DistilBERT models to identify co-occurring conditions, including depression, anxiety, and narcissistic personality disorder. To address class imbalance and optimize decision boundaries, the framework integrates a composite loss function (focal, dice, and log loss) and utilizes Youdens J statistic for threshold calibration. Validation on textual datasets demonstrates the efficacy of this approach, with an overall F1-score of 0.930 and AUC values exceeding 0.89. Comparative analysis suggests that decomposing complex diagnostic tasks into independent binary problems significantly reduces inter-class confusion relative to standard multi-class baselines. Furthermore, a qualitative error analysis highlights specific linguistic challenges, such as contextual polarity shifting, metaphorical ambiguity, and colloquial usage, that impact model specificity. The findings demonstrate the potential of the proposed framework as a robust screening tool for online mental health monitoring, while underscoring the necessity of human oversight to mitigate linguistic misinterpretations. Author summaryMental health disorders such as depression, anxiety, and narcissistic personality disorder represent a major global health challenge. This work proposes a method that employs transformer-based deep learning models to analyze social media posts for mental health assessment. A significant hurdle in automated diagnosis is that these conditions often occur together (comorbidity), whereas many existing Artificial Intelligence (AI) systems are designed to detect only a single disorder at a time. This study proposes a solution using a "multi-label" deep learning framework. Rather than relying on a single multi-class classifier, the approach utilizes an ensemble of specialized binary models, each trained to detect indicators of a specific disorder. This design reduces classification confusion between clinically similar conditions, such as depression and anxiety. The method was evaluated on publicly available datasets, had an F1-score of 0.930 which outperformed the existing approaches. The presented approach demonstrated high effectiveness, achieving better separation between clinically similar disorders compared to traditional methods. Crucially, the detailed investigation beyond the standard statistical metrics was performed which looked into specific models mistakes. It was found that, while the presented AI model is highly sensitive, it can be confused by the specifics of the language such as metaphors (e.g., "feeling like a pressure cooker"), negations (e.g., "I am not worried"), and the colloquial clinical terms. These results highlight that AI is a powerful tool which can be used for early screening and continuous monitoring on social media, while it still requires careful calibration and human oversight to distinguish between genuine symptoms and everyday emotional expression. The findings demonstrate that analyzing social media texts with advanced machine learning techniques can serve as a powerful complementary tool to clinical diagnostics. While not intended to completely replace professional evaluation, the proposed approach can help identify potential risks, promote earlier detection of mental health disorders, support preventive interventions, and ultimately improve access to care.

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.

Malaria affects almost 263 million people worldwide, most of whom live in sub-Saharan countries. In a strategy to reduce malaria-related mortality and limit transmission, diagnosis in endemic areas needs to be immediately available on the field, easy to perform and cheap. Therefore, it currently heavily relies on microscopic examination of blood smears. However, several studies comparing the sensitivity of this approach with qPCR, considered as the most sensitive method albeit not available on the field, found that up to half of the infected population failed to be detected by microscopy alone because no visible parasites could be found in blood smears. These so-called submicroscopic infections pose a diagnostic challenge, yet represent a huge reservoir for malaria transmission. In this study, we hypothesized that qPCR results could be predicted by deep learning from subtle cell signals present in thin blood smear images, even in the absence of visible parasites, making a sensitive diagnostic directly available on the field using a microscope and a smartphone. To test this hypothesis, we acquired a large smartphone-based blood smear images dataset from samples tested both for microscopy and qPCR. We then focused exclusively on these "negative" slides from the microscopic diagnostic point of view, among which half were qPCR positive. A range of standard deep learning models were evaluated to best predict the qPCR result from these microscopy images, using various backbones along with various aggregation functions at the slide level, from a simple vote to Multiple Instance Learning with attention. Our results show that the qPCR results can be predicted from parasite free blood smear images with 62.00% ({+/-}2.5 on 4-folds) accuracy and reaching 67.2 % ({+/-}9.6 on 4-folds) in sensitivity. We then used generative models to investigate the subtle morphological variations occurring in red blood cells that may contribute to predicting malaria infection in the absence of parasites. Leveraging thin blood smear and portable deep learning, we established the first proof of concept that the qPCR sensitivity can be approached through the detection of submicroscopic infections directly on the field without additional infrastructure and thus could significantly improve malaria surveillance and elimination efforts.

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.

IntroductionClimate change is a global adverse phenomenon affecting the health and wellbeing of all humans, and timely awareness can help mitigate these health effects. ObjectiveTo understand the knowledge and attitudes of Sri Lankan adults towards climate change and its effects on human health. MethodsA web based cross-sectional survey was conducted using a structured, pretested, web based, self-administered questionnaire, using a respondent driven sampling technique, among Sri Lankan adults. Data was collected over three months, from 1st September 2022. Responses were automatically stored in a cloud-based database and were imported into a spreadsheet and analysed using MS Excel. ResultsMajority of the 118 respondents were young, educated, employed adults from western province, and 56.78% were females. Overall knowledge on climate change was good among 82.20%, while over 90% had a good or favourable knowledge on health effects associated with climate change. Respondents demonstrated a good awareness of climate effects on skin cancer (92.37%), mental illnesses (82.2%) and asthma (82.2%), but were less aware of the effects on diabetes (28.8%), COPD (38.1%) and heart diseases (46.6%), and vector borne diseases such as Malaria (57.6%) and Dengue (61.8%). Over 90% had a good attitude towards the need for climate change mitigation and climate friendly activities being implemented. ConclusionsUrban, educated Sri Lankan adults had a good understanding and awareness on the health effects of climate change, and the importance of mitigating it in relation to its health effects, but further studies are needed to understand the awareness levels of the less educated rural communities.

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.

Collective interaction of individuals in various settings is crucial for exposure to infections, encompassing complex viral interplay and amplifying infectious risk through phenomena such as social reinforcement, clustering and superspreading events, during the COVID-19 pandemic. However, standard epidemic models often inadequately capture such heterogeneity, overlooking the higher-order social structural. Spatiotemporal variation in transmission, an essential feature of the pandemic, remains poorly quantified at various scales, particularly in integrating high-resolution data streams and complex network approaches. We introduced a higher-order simplicial model that unifies human mobility data, genetic diversity and antigenic drift to systematically investigate the role of social reinforcement, spatiotemporal variation and genetic mutations in SARS-CoV-2 transmission. We found a median of 5.3% to 14.4% of infections across provinces were attributed to social reinforcement, while cluster heterogeneity contributed to a median of 17%-71% increase in susceptibility. Multiple viral interactions elevated transmissibility by 68%-70% across the periods of dominant variants. The reconstructed transmission networks underscored distinct spatiotemporal variation, with dynamic critical locations, varying mobility patterns, and evolving geographic cluster structures, by assessing complex networks. The influence of human mobility was found to be positive on transmission, effective distance was negatively associated with infection risks, while greater genetic diversity and antigenic drift were linked to higher susceptibility and transmissibility. Our proposed data-driven higher-order framework could help us to understand epidemics better by accounting the role of collective interactions, population mobility, and genetic mutation in transmission, which could inform the targeted interventions to mitigate SARS-CoV-2 and other respiratory pathogens.

Timely forecasting of dengue hospitalizations is essential for public health preparedness but is frequently limited by delays in official reporting systems. Climatic conditions strongly influence dengue transmission, yet hospitalization data often become available weeks after patient admission, reducing their value for early response. Digital information generated during clinical practice may provide a more timely signal of emerging disease activity. This study evaluates whether integrating climate data with real-time records of physicians searches for dengue-related information improves short-term forecasts of dengue hospitalizations in Brazil under both ideal and realistic reporting conditions. Weekly hospitalization counts, weather indicators, and anonymized physician search data from a widely used clinical decision-support platform were combined to generate forecasts across multiple geographic regions. Model performance was compared under two scenarios: one assuming immediate availability of hospitalization data and another incorporating typical reporting delays. When hospitalization data were timely, climate-based models achieved the highest predictive accuracy. Under realistic reporting delays, however, models incorporating physicians search behavior consistently outperformed approaches relying solely on climate information or hospitalization history. In several regions, increases in physician search activity preceded or coincided with rises in hospital admissions, indicating early clinical engagement with dengue cases. These findings indicate that physician search behavior constitutes a valuable real-time indicator of dengue activity. Integrating digital clinical behavior with climate data enhances forecasting performance under real-world reporting constraints and may strengthen early-warning systems and public health decision-making for dengue and other climate-sensitive diseases. Author SummaryDengue is a major public health challenge in Brazil, where large outbreaks place sudden pressure on health services. Although climate conditions influence dengue transmission, public health responses often rely on hospitalization data that become available only weeks or months after patients are admitted, limiting the ability to act early. In this study, we explored whether real-time digital information generated by physicians could help overcome this delay. We combined weather data with anonymized records of physicians searches for dengue-related information within a widely used clinical decision-support platform in Brazil. We then tested whether these digital signals could improve short-term forecasts of dengue hospitalizations across different regions of the country, especially when official hospital data were delayed. We found that climate patterns were strong predictors when hospitalization data were timely. However, under realistic reporting delays, models that incorporated physicians search behavior produced more accurate forecasts. These findings show that digital clinical behavior can provide early insight into rising disease activity and support more timely public health responses.