Search Results (363)

Background: Electrocardiograms (EKGs) are essential tools in emergency medicine, often used to evaluate chest pain, dyspnea, and other symptoms suggestive of cardiac dysfunction. Yet, EKGs are not universally administered to all emergency department (ED) patients. Understanding and predicting which patients receive an EKG may offer insights into clinical decision making, resource allocation, and potential disparities in care. This study examines whether integrating structured clinical data with free-text patient narratives can improve prediction of EKG utilization in the ED. Methods: We conducted a retrospective observational study to predict electrocardiogram (EKG) utilization using data from 13,115 adult emergency department (ED) visits in the nationally representative 2021 National Hospital Ambulatory Medical Care Survey–Emergency Department (NHAMCS-ED), leveraging both structured features—demographics, vital signs, comorbidities, arrival mode, and triage acuity, with the most influential selected via Lasso regression—and unstructured patient narratives transformed into numerical embeddings using Clinical-BERT. Four supervised learning models—Logistic Regression (LR), Support Vector Machine (SVM), Random Forest (RF) and Extreme Gradient Boosting (XGB)—were trained on three inputs (structured data only, text embeddings only, and a late-fusion combined model); hyperparameters were optimized by grid search with 5-fold cross-validation; performance was evaluated via AUROC, accuracy, sensitivity, specificity and precision; and interpretability was assessed using SHAP values and Permutation Feature Importance. Results: EKGs were administered in 30.6% of adult ED visits. Patients who received EKGs were more likely to be older, White, Medicare-insured, and to present with abnormal vital signs or higher triage severity. Across all models, the combined data approach yielded superior predictive performance. The SVM and LR achieved the highest area under the ROC curve (AUC = 0.860 and 0.861) when using both structured and unstructured data, compared to 0.772 with structured data alone and 0.823 and 0.822 with unstructured data alone. Similar improvements were observed in accuracy, sensitivity, and specificity. Conclusions: Integrating structured clinical data with patient narratives significantly enhances the ability to predict EKG utilization in the emergency department. These findings support a personalized medicine framework by demonstrating how multimodal data integration can enable individualized, real-time decision support in the ED. Full article

Heart failure (HF) poses a substantial global burden due to its high morbidity, mortality, and healthcare costs. Accurate prognostication is crucial for optimizing treatment, resource allocation, and patient counseling. Prognostic tools range from simple clinical scores such as ADHERE and MAGGIC to more complex models incorporating biomarkers (e.g., NT-proBNP, sST2), imaging, and artificial intelligence techniques. In acute HF, models like EHMRG and STRATIFY aid early triage, while in chronic HF, tools like SHFM and BCN Bio-HF support long-term management decisions. Despite their utility, most models are limited by poor generalizability, reliance on static inputs, lack of integration into electronic health records, and underuse in clinical practice. Novel approaches involving machine learning, multi-omics profiling, and remote monitoring hold promise for dynamic and individualized risk assessment. However, these innovations face challenges regarding interpretability, validation, and ethical implementation. For prognostic models to transition from theoretical promise to practical impact, they must be continuously updated, externally validated, and seamlessly embedded into clinical workflows. This review emphasizes the potential of prognostic models to transform HF care but cautions against uncritical adoption without robust evidence and practical integration. In the evolving landscape of HF management, prognostic models represent a hopeful avenue, provided their limitations are acknowledged and addressed through interdisciplinary collaboration and patient-centered innovation. Full article

Early detection of clinical deterioration in emergency department (ED) patients is critical for timely interventions. This study evaluated the predictive performance of the National Early Warning Score (NEWS) parameters using machine learning. We conducted a single-center retrospective observational study including 27,238 adult ED patients admitted to Merano Hospital (Italy) between June 2022 and June 2023. NEWS vital signs were collected at triage, and mortality at 48 h, 7 days, and 30 days was obtained from ED database. Decision tree analysis (CHAID algorithm) was used to identify predictors of mortality; 10-fold cross-validation was applied to avoid overfitting. Mortality was 0.4% at 48 h, 1% at 7 days, and 2.45% at 30 days. For 48-h mortality, oxygen supplementation (FiO₂ >21%) and AVPU = “U” were the strongest predictors, with a maximum risk of 31.6%. For 7-day mortality, SpO₂ was the key predictor, with mortality up to 48.1%. At 30 days, patients with AVPU ≠ A, FiO₂ > 21%, and SpO₂ ≤ 94% had a mortality risk of 66.7%. Decision trees revealed different cut-offs compared to the standard NEWS. This study demonstrated that for ED patients, the NEWS may require some adjustments in both the cut-offs for vital parameters and the methods of collecting these parameters. Full article

Hemorrhoidal disease remains a prevalent yet often overlooked condition, affecting millions worldwide and imposing a substantial healthcare burden. Despite the availability of multiple treatment options, gaps persist in patient education, early symptom recognition, and optimal treatment selection. Recent advancements are evolving the pharmacist’s role in hemorrhoid management beyond traditional over-the-counter (OTC) and prescription approaches. The 2024 American Society of Colon and Rectal Surgeons (ASCRS) guidelines introduce updates on the use of phlebotonics, a class of venoactive drugs gaining recognition for their role in symptom management, yet largely underutilized in U.S. clinical practice. In parallel, novel clinical tools are reshaping how pharmacists engage in assessment and care. The integration of digital decision-support platforms and structured evaluation algorithms now empowers them to systematically evaluate symptoms, identify red flag signs, and optimize patient triage. These tools reduce diagnostic variability and improve decision-making accuracy. Given their accessibility and trusted role in frontline healthcare, pharmacists are well-positioned to bridge these critical gaps by adopting emerging treatment recommendations, leveraging algorithm-driven assessments, and reinforcing best practices in patient education and referral. This narrative review aims to equip pharmacists with updated insights into evidence-based hemorrhoid management strategies and provide them with structured assessment algorithms to standardize symptom evaluation and treatment pathways. By integrating these innovations, pharmacists can enhance treatment outcomes, promote patient safety, and contribute to improved quality of life (QoL) for individuals suffering from hemorrhoidal disease. Full article

Background and Objectives: this pilot study aimed to evaluate the diagnostic accuracy of ChatGPT-4o in analyzing oral mucosal lesions from clinical images. Materials and Methods: a total of 110 clinical images, including 100 pathological lesions and 10 healthy mucosal images, were retrieved from Google Images and analyzed by ChatGPT-4o using a standardized prompt. An expert panel of five clinicians established a reference diagnosis, categorizing lesions as benign or malignant. The AI-generated diagnoses were classified as correct or incorrect and further categorized as plausible or not plausible. The accuracy, sensitivity, specificity, and agreement with the expert panel were analyzed. The Artificial Intelligence Performance Instrument (AIPI) was used to assess the quality of AI-generated recommendations. Results: ChatGPT-4o correctly diagnosed 85% of cases. Among the 15 incorrect diagnoses, 10 were deemed plausible by the expert panel. The AI misclassified three malignant lesions as benign but did not categorize any benign lesions as malignant. Sensitivity and specificity were 91.7% and 100%, respectively. The AIPI score averaged 17.6 ± 1.73, indicating strong diagnostic reasoning. The McNemar test showed no significant differences between AI and expert diagnoses (p = 0.084). Conclusions: In this proof-of-concept pilot study, ChatGPT-4o demonstrated high diagnostic accuracy and strong descriptive capabilities in oral mucosal lesion analysis. A residual 8.3% false-negative rate for malignant lesions underscores the need for specialist oversight; however, the model shows promise as an AI-powered triage aid in settings with limited access to specialized care. Full article

Background: Overcrowded emergency departments (EDs) create significant challenges for patient management and hospital efficiency. In response, Amiens Picardy University Hospital (APUH) developed the “Prediction of the Patient Pathway in the Emergency Department” (3P-U) model to enhance patient flow management. Objectives: To develop and clinically validate an explainable artificial intelligence (XAI) model for hospital admission predictions, using structured triage data, and demonstrate its real-world applicability in the ED setting. Methods: Our retrospective, single-center study involved 351,019 patients consulting in APUH’s EDs between 2015 and 2018. Various models (including a cross-validation artificial neural network (ANN), a k-nearest neighbors (KNN) model, a logistic regression (LR) model, and a random forest (RF) model) were trained and assessed for performance with regard to the area under the receiver operating characteristic curve (AUROC). The best model was validated internally with a test set, and the F1 score was used to determine the best threshold for recall, precision, and accuracy. XAI techniques, such as Shapley additive explanations (SHAP) and partial dependence plots (PDP) were employed, and the clinical explanations were evaluated by emergency physicians. Results: The ANN gave the best performance during the training stage, with an AUROC of 83.1% (SD: 0.2%) for the test set; it surpassed the RF (AUROC: 71.6%, SD: 0.1%), KNN (AUROC: 67.2%, SD: 0.2%), and LR (AUROC: 71.5%, SD: 0.2%) models. In an internal validation, the ANN’s AUROC was 83.2%. The best F1 score (0.67) determined that 0.35 was the optimal threshold; the corresponding recall, precision, and accuracy were 75.7%, 59.7%, and 75.3%, respectively. The SHAP and PDP XAI techniques (as assessed by emergency physicians) highlighted patient age, heart rate, and presentation with multiple injuries as the features that most specifically influenced the admission from the ED to a hospital ward. These insights are being used in bed allocation and patient prioritization, directly improving ED operations. Conclusions: The 3P-U model demonstrates practical utility by reducing ED crowding and enhancing decision-making processes at APUH. Its transparency and physician validation foster trust, facilitating its adoption in clinical practice and offering a replicable framework for other hospitals to optimize patient flow. Full article

This study presents a next-generation clinical decision-support architecture for Clinical Decision Support Systems (CDSS) focused on emergency triage. By integrating Large Language Models (LLMs), Federated Learning (FL), and low-latency streaming analytics within a modular, privacy-preserving framework, the system addresses key deployment challenges in high-stakes clinical settings. Unlike traditional models, the architecture processes both structured (vitals, labs) and unstructured (clinical notes) data to enable context-aware reasoning with clinically acceptable latency at the point of care. It leverages big data infrastructure for large-scale EHR management and incorporates digital twin concepts for live patient monitoring. Federated training allows institutions to collaboratively improve models without sharing raw data, ensuring compliance with GDPR/HIPAA, and FAIR principles. Privacy is further protected through differential privacy, secure aggregation, and inference isolation. We evaluate the system through two studies: (1) a benchmark of 750+ USMLE-style questions validating the medical reasoning of fine-tuned LLMs; and (2) a real-world case study (n = 132, 75.8% first-pass agreement) using de-identified MIMIC-III data to assess triage accuracy and responsiveness. The system demonstrated clinically acceptable latency and promising alignment with expert judgment on reviewed cases. The infectious disease triage case demonstrates low-latency recognition of sepsis-like presentations in the ED. This work offers a scalable, audit-compliant, and clinician-validated blueprint for CDSS, enabling low-latency triage and extensibility across specialties. Full article

Background/Objectives: The demand for accessible and reliable digital health services has increased significantly in recent years, particularly in regions facing physician shortages. HomeDOCtor, a conversational AI platform developed in Slovenia, addresses this need with a nationally adapted architecture that combines retrieval-augmented generation (RAG) and a Redis-based vector database of curated medical guidelines. The objective of this study was to assess the performance and impact of HomeDOCtor in providing AI-powered healthcare assistance. Methods: HomeDOCtor is designed for human-centered communication and clinical relevance, supporting multilingual and multimedia citizen inputs while being available 24/7. It was tested using a set of 100 international clinical vignettes and 150 internal medicine exam questions from the University of Ljubljana to validate its clinical performance. Results: During its six-month nationwide deployment, HomeDOCtor received overwhelmingly positive user feedback with minimal criticism, and exceeded initial expectations, especially in light of widespread media narratives warning about the risks of AI. HomeDOCtor autonomously delivered localized, evidence-based guidance, including self-care instructions and referral suggestions, with average response times under three seconds. On international benchmarks, the system achieved ≥95% Top-1 diagnostic accuracy, comparable to leading medical AI platforms, and significantly outperformed stand-alone ChatGPT-4o in the national context (90.7% vs. 80.7%, p = 0.0135). Conclusions: Practically, HomeDOCtor eases the burden on healthcare professionals by providing citizens with 24/7 autonomous, personalized triage and self-care guidance for less complex medical issues, ensuring that these cases are self-managed efficiently. The system also identifies more serious cases that might otherwise be neglected, directing them to professionals for appropriate care. Theoretically, HomeDOCtor demonstrates that domain-specific, nationally adapted large language models can outperform general-purpose models. Methodologically, it offers a framework for integrating GDPR-compliant AI solutions in healthcare. These findings emphasize the value of localization in conversational AI and telemedicine solutions across diverse national contexts. Full article

This study applies stochastic discrete event modeling to demonstrate that reducing wait times for specialized outpatient clinics in the Costa Rican public healthcare system is possible. The classification process identified four medical specialties with the longest wait times. It includes the creation of a separate queuing theory model for each specialty. The birth and death model allowed for estimating the number of arrivals and consultations in the simulation. Validation was performed by comparing the model’s input and output data with real-world statistical reports. An analysis of medical specialists revealed that approximately 22% of patients referred to secondary care did not require specialized medical consultation. Through simulation and the use of stochastic input data, patient waiting times decreased. In an optimistic scenario, waiting times decreased steadily across all specialties over 24 months. Ophthalmology and orthopedics reduced their waiting times to less than 300 days. Otorhinolaryngology decreased from 370 to 250 days, and urology showed the most significant improvement, decreasing from 350 to 100 days in the first year and remaining stable. This evidence transforms the traditional paradigm of increasing capacity as the only solution to the waiting list problem and positions improving the referral process as an alternative. To achieve these results, the study highlights the importance of implementing improved triage protocols in primary care, integrating decision-support tools for general practitioners using machine learning, for example, to reduce unnecessary referrals. Training programs and feedback mechanisms could also align referral practices with specialty criteria. While these strategies were not implemented in this study, the simulation results provide a solid basis for their design and future evaluation. Full article

Telemedicine in emergency contexts presents unique challenges, particularly in multilingual and low-resource settings where accurate, clinical understanding and triage decision support are critical. This paper introduces TriagE-NLU, a novel multilingual natural language understanding system designed to perform both semantic parsing and clinical intervention classification from emergency dialogues. The system is built on a federated learning architecture to ensure data privacy and adaptability across regions and is trained using TriageX, a synthetic, clinically grounded dataset covering five languages (English, Spanish, Romanian, Arabic, and Mandarin). TriagE-NLU integrates fine-tuned multilingual transformers with a hybrid rules-and-policy decision engine, enabling it to parse structured medical information (symptoms, risk factors, temporal markers) and recommend appropriate interventions based on recognized patterns. Evaluation against strong multilingual baselines, including mT5, mBART, and XLM-RoBERTa, demonstrates superior performance by TriagE-NLU, achieving F1 scores of 0.91 for semantic parsing and 0.89 for intervention classification, along with 0.92 accuracy and a BLEU score of 0.87. These results validate the system’s robustness in multilingual emergency telehealth and its ability to generalize across diverse input scenarios. This paper establishes a new direction for privacy-preserving, AI-assisted triage systems. Full article

Chronic coronary syndrome (CCS) and atrial fibrillation (AF) are prevalent cardiovascular conditions that share numerous risk factors and pathophysiological mechanisms. While clinical scores commonly used in AF—such as CHA₂DS₂VA (which includes congestive heart failure, hypertension, age ≥ 75, diabetes, stroke/TIA, vascular disease, and age 65–74), HAS-BLED (which incorporates hypertension, abnormal renal/liver function, stroke, bleeding history, labile INR, elderly age, and drug/alcohol use), and C₂HEST (incorporating coronary artery disease, COPD, hypertension, elderly age ≥ 75, systolic heart failure, and thyroid disease)—are traditionally applied to rhythm or bleeding risk prediction, their value in estimating the angiographic severity of coronary artery disease (CAD) remains underexplored. We conducted a prospective, single-center study including 131 patients with suspected stable CAD referred for coronary angiography, stratified according to coronary angiographic findings into two groups: significant coronary stenosis (S-CCS) and non-significant coronary stenosis (N-CCS). At admission, AF-related scores (CHA₂DS₂, CHA₂DS₂VA, CHA₂DS₂VA-HSF, CHA₂DS₂VA-RAF, CHA₂DS₂VA-LAF, HAS-BLED, C₂HEST, and HATCH) were calculated. CAD severity was subsequently assessed using the SYNTAX and Gensini scores. Statistical comparisons and Pearson correlation analyses were performed to evaluate the association between clinical risk scores and angiographic findings. Patients in the S-CCS group had significantly higher scores in CHA₂DS₂VA (4.09 ± 1.656 vs. 3.20 ± 1.338, p = 0.002), HAS-BLED (1.98 ± 0.760 vs. 1.36 ± 0.835, p < 0.001), CHA₂DS₂VA-HSF (6.00 ± 1.854 vs. 5.26 ± 1.712, p = 0.021), and C₂HEST (3.49 ± 1.501 vs. 2.55 ± 1.279, p < 0.001). Multivariate logistic regression identified HAS-BLED and C₂HEST as independent predictors of significant coronary lesions. A threshold value of HAS-BLED ≥ 1.5 and C₂HEST ≥ 3.5 demonstrated moderate discriminative ability (AUC = 0.694 and 0.682, respectively), with acceptable sensitivity and specificity. These scores also demonstrated moderate to strong correlations with both Gensini and SYNTAX scores. AF-related clinical scores, especially HAS-BLED and C₂HEST, may serve as practical and accessible tools for early CAD risk stratification in patients with suspected CCS. Their application in clinical practice may serve as supplementary triage tools to help prioritize patients for further diagnostic evaluation, but they are not intended to replace standard imaging or testing. Full article

Background: Early identification of sepsis is critical for improving outcomes, particularly in low-resource emergency settings. In Thai community hospitals, where physicians may not always be available, triage is often nurse-led. Selecting accurate and practical sepsis screening tools is essential not only for timely clinical decision-making but also for timely diagnosis and promoting appropriate antibiotic use. Methods: This cross-sectional study analyzed 475 adult patients with suspected sepsis who presented to the emergency department of a Thai community hospital, using retrospective data from January 2021 to December 2022. Six screening tools were evaluated: Systemic Inflammatory Response Syndrome (SIRS), Quick Sequential Organ Failure Assessment (qSOFA), Modified Early Warning Score (MEWS), National Early Warning Score (NEWS), National Early Warning Score version 2 (NEWS2), and Search Out Severity (SOS). Diagnostic accuracy was assessed using International Classification of Diseases, Tenth Revision (ICD-10) codes as the reference standard. Performance metrics included sensitivity, specificity, predictive values, likelihood ratios, and the area under the receiver operating characteristic (AUROC) curve, all reported with 95% confidence intervals. Results: SIRS had the highest sensitivity (84%), while qSOFA demonstrated the highest specificity (91%). NEWS2, NEWS, and MEWS showed moderate and balanced diagnostic accuracy. SOS also demonstrated moderate accuracy. Conclusions: A two-step screening approach—using SIRS for initial triage followed by NEWS2 for confirmation—is recommended. This strategy enhances nurse-led screening and optimizes limited resources in emergency care. Early sepsis detection through accurate screening tools constitutes a feasible public health intervention to support appropriate antibiotic use and mitigate antimicrobial resistance, especially in resource-limited community hospital settings. Full article

Background/Objectives: Dyspnea is one of the main causes of visits to the Emergency Department (ED) and hospitalization, with its differential diagnosis representing a challenge for the clinician. Lung ultrasound (LUS) is a widely used tool in ED. The objective of this study was to evaluate the impact of LUS, performed by a non-expert operator, in determining diagnosis and prognosis of patients with dyspnea. Methods: A total of 60 patients presenting with dyspnea at the ED were prospectively enrolled and underwent LUS examination by a medical student, after brief training, within 3 h of triage. LUS findings were classified into four patterns: N.1, absence of notable ultrasound findings, attributable to COPD/ASMA exacerbation; N.2, bilateral interstitial syndrome, suggestive of acute heart failure; N.3, subpleural changes/parenchymal consolidations, suggestive of pneumoniae; and N.4, isolate polygonal triangular consolidation, attributable to infarction in the context of pulmonary thromboembolism. Results: The diagnostic hypothesis formulated after LUS was compared with the final diagnosis after further investigations in the ED, showing agreement in 90% of cases. The mean LUS score value was higher in patterns N.2 (18.4 ± 8.5) and N.3 (17 ± 6.6), compared to patterns N.1 and N.4 (9.8± 6.7 and 11.5 ± 2.1). Given the high prevalence of pattern N.2, the diagnostic accuracy of LUS in this context was further evaluated, showing a sensitivity of 82% and specificity of 100%. In terms of the prognostic value of LUS, hospitalized patients had a higher LUS score compared to those discharged (17.3 ± 8.1 vs. 8.5 ± 6.8, p value 0.004). A similar trend was obtained in the subgroup of patients requiring non-invasive ventilation (NIV), who present a higher LUS score (21.1 ± 6.6 vs. 13.1 ± 8.1, p value 0.002). When considering a combined outcome (death and NIV), patients with worse outcomes more often had a LUS score > 15 (p value < 0.001). Conclusions: In conclusion, this study confirms that LUS is a very useful tool in the ED, assisting the clinical evaluation for diagnosis, treatment decision, and determination of the appropriate care setting for patients with acute dyspnea. Its short learning curve allows even non-expert staff to use it effectively. Full article

Objectives: While pre-screening triage (PST) enhances healthcare efficiency in emergency and pediatric settings, its application in dental healthcare remains undervalued. This novel study implemented PST in dental services, identifying determinants of physician–patient dissatisfaction to optimize triage systems and promote dental health outcomes. Methods: A cross-sectional survey (July–September 2024) recruited 113 physicians and 206 patients via convenience sampling. Dissatisfaction levels were quantified using validated questionnaires and analyzed through t-tests, ANOVA, and regression models. Results: In total, 37.17% of physicians with prior PST experience demonstrated significantly higher dissatisfaction scores (37.67 ± 9.08 vs. 32.51 ± 10.08, p = 0.006). Multivariate analysis revealed that experienced physicians rated PST services 5.63 points higher than less experienced counterparts (95% CI: 0.75–10.51). Dental patients expressed dissatisfaction with nurse attitudes (β = 1.04, 95% CI: 0.07–2.01) and triage process inefficiencies. Conclusions: Key dissatisfaction drivers include a lack of physician PST exposure and nurse–patient interaction quality in dental settings. These findings advocate for the development of a specialized triage system to enhance clinical workflow efficiency and service effectiveness in dental healthcare. Full article

Objectives: Pneumothorax is a critical condition frequently encountered in emergency departments (EDs), with spontaneous pneumothorax (SP) and traumatic pneumothorax (TP) presenting distinct clinical challenges. This study aimed to evaluate the epidemiological characteristics, clinical outcomes, and treatment strategies for SP and TP across different age groups and provide insights for optimizing emergency management protocols. Methods: This retrospective cohort study analyzed 2797 cases of pneumothorax over five years (2018–2023) at a tertiary care center. Patients were stratified by age (18–39, 40–64, and >65 years) and pneumothorax type (SP vs. TP). Data on demographics, clinical presentation, treatment, hospital stay, recurrence, and complications were extracted from medical records. Comparative statistical analyses were also conducted. Results: The mean age of patients with SP was 32.5 ± 14.7 years, whereas patients with TP were older (37.8 ± 16.2 years, p < 0.001). Male predominance was observed in both groups: 2085 (87.0%) in the SP group and 368 (92.0%) in the TP group (p = 0.01). The right lung was more frequently affected in the SP (64.2%) and TP (56.0%) groups (p < 0.001). Age-related differences were evident in both groups of patients. In the SP group, younger patients (18–39 years) represented the majority of cases, whereas older patients (≥65 years) were more likely to present with SSP and required more invasive management (p < 0.01). In the TP group, younger patients often had pneumothorax due to high-energy trauma, whereas older individuals developed pneumothorax due to falls or iatrogenic causes (p < 0.01). SP predominantly affected younger patients, with a history of smoking and male predominance associated with younger age (p < 0.01). TP is more frequent in older patients, often because of falls or iatrogenic injuries. Management strategies varied by age group; younger patients were often managed conservatively, whereas older patients underwent more invasive procedures (p < 0.01). Surgical intervention was more common in younger patients in the TP group, whereas conservative management was more frequent in elderly patients (p < 0.01). The clinical outcomes differed significantly, with older patients having longer hospital stays and higher rates of persistent air leaks (p < 0.01). Recurrence was more common in younger patients with SP, whereas TP recurrence rates were lower across all age groups (p < 0.01). No significant differences were observed in re-expansion pulmonary edema, empyema, or mortality rates between the age groups, suggesting that age alone was not an independent predictor of these complications when adjusted for pneumothorax severity and management strategy (p = 0.22). Conclusions: Age, pneumothorax subtype, and underlying pulmonary comorbidities were identified as key predictors of clinical outcomes. Advanced age, secondary spontaneous pneumothorax, and COPD were independently associated with recurrence, prolonged hospitalization, and in-hospital mortality, respectively. These findings highlight the need for risk-adapted management strategies to improve triaging and treatment decisions for spontaneous and traumatic pneumothorax. Full article