Browse Articles

Dynamic properties of Kermack-McKendrick-like models

We rigorously analyze the dynamic properties of Kermack-McKendrick-like compartmental models for infectious diseases, extending the classical SIR framework to include exposed individuals, mild and severe infections, hospitalization, and intensive care unit (ICU) compartments. Using a 13-compartment model, we establish mathematical results on well-posedness, the basic reproduction number R 0 , a first integral leading to a unique final epidemic size, and the global stability of the disease-free equilibrium under permanent immunity. When temporary immunity is included, we prove the existence of an endemic equilibrium for R 0  > 1. An age-stratified multi-group version of the model is also studied, demonstrating similar convergence properties and highlighting the impact of age structure on epidemic dynamics. Our results provide a rigorous mathematical framework for understanding how immunity duration, clinical progression, and age structure shape epidemic outcomes.

Machine learning–based long-term degradation and LCOE Analysis of floating PV with custom pontoon design

Floating Photovoltaic (FPV) systems are a feasible alternative for solar energy utilization in areas where land is scarce. However, issues related to performance degradation, ecological effects, and economic viability have restricted the widespread acceptance of this technology. This study examines the eco-compatibility of a 5 MW FPV solar plant, which is suitable for a wetland ecosystem in Bangladesh, and examines the technical feasibility, environmental sustainability, and economic viability of the plant. In this study, a model was developed using detailed system-level simulations, and the model validated the structural feasibility of a specially designed light-permeable annular pontoon by performing hydrostatic buoyancy and stability tests. Ecological compatibility is assessed using a light-transmission-based photosynthetic viability model, whereas the long-term degradation of the performance ratio and energy yield over a 25-year lifetime is predicted using a climate-aware machine learning framework that incorporates irradiance, temperature, humidity, and system aging effects. Furthermore, an economic model sensitive to inflation was used to examine the levelized cost of electricity (LCOE) in the case study. The results show an average performance ratio of 82.4%, lifecycle energy outputs of approximately 222 GWh, and an LCOE of 0.0315 USD/kWh. In addition, approximately 70% of the aquatic photosynthesis potential is retained, and approximately 104,149.5 tCO₂ is eliminated.

Analysis of the efficiency and sustainability of the underground loop in Hangzhou Future Science and Technology City

In order to improve the low utilization efficiency and limited environmental benefits of the underground loop in Hangzhou Future Science and Technology City, this study systematically analyzes the operation status of the loop road by collecting traffic flow data through field research and combining questionnaire interviews to understand the driver’s experience. It is found that the current underground loop system faces several challenges, including low public awareness, insufficient precision in traffic guidance, inconsistent management of land parcel access interfaces, limited operational flexibility, and inadequate realized carbon reduction. As a result, only about 10% of traffic uses the loop for arrival and departure, and the utilization of underground transportation resources remains far below the expected target. Based on the comparative analysis of similar projects such as Xidong New City in Wuxi and Zhenru Subcenter in Shanghai, this study proposes strategies such as constructing a comprehensive four-tier traffic guidance system, strengthening intelligent information management and control, unifying the collaborative management standards of parcel access interfaces, and deepening low-carbon and sustainable operation. These strategies can effectively alleviate surface-level traffic, provide practical references for the planning and operation of urban underground transportation facilities, and help the core area realize the synergistic and sustainable development of transportation and space.

Development and validation of the Manipal Interstitial Lung Disease Education Booklet (MILD EduB) for individuals with interstitial lung disease

In India, Interstitial Lung Disease (ILD) is a significant healthcare burden. Poor outcomes are caused by delayed diagnosis and insufficient patient education (PE). In this context, the Manipal Interstitial Lung Disease Educational Booklet (MILD EduB), a culturally relevant resource for individuals with ILD in India, was developed and validated. The booklet was developed in four stages: (1) a scoping review to identify important educational themes; (2) development of a draft booklet shared with experts for review and feedback; (3) validation of the booklet by healthcare professionals using the content validity index (CVI); and (4) face validation by individuals with ILD. The initial scale-level CVI (S-CVI/Ave) was 0.79, which improved to 1.0 following modifications, indicating good expert consensus via the healthcare professionals. Face validation by individuals with ILD (n = 10) using a 4-point Likert scale demonstrated 100% agreement for readability, relevance, and design. With a Flesch Reading Ease Scale score of 80.9, the finalised booklet is considered easy to read and accessible to individuals with ILD with a wide range of reading levels. MILD EduB fills an important gap in ILD-specific educational resources, particularly in low-resource settings. Its evidence-based content and patient-centered design aim to improve adherence and self-management.

A local sequence alignment approach to recognizing fixed poetic forms across languages

Fixed poetic forms such as the sonnet, ottava rima, or terza rima are an important feature of European literary traditions, yet large-scale empirical research on their cross-lingual distribution and evolution has been limited so far. This paper introduces a fully language-independent, unsupervised method for identifying recurrent rhyme-based forms using local sequence alignment. Drawing on 187,719 poems from six European traditions (Czech, English, French, German, Italian, Russian) in the PoeTree collection, we encode rhyme schemes in a compact eight-symbol alphabet and apply the Smith–Waterman algorithm via the Metronome package to compute pairwise distances. Dimensionality reduction (UMAP) and density-based clustering (HDBSCAN) yield 61 clusters, many of which align with known fixed forms. Evaluation against existing Czech and Russian annotations shows strong recall, while supervised classification experiments—both within and across languages—demonstrate that form categories are robustly learnable in the induced vector space. We illustrate the potential of such data for literary research in three showcases: cross-tradition influence in 19th-century Czech poetry, topical affinities of selected forms using multilingual topic modeling, and geographic associations revealed through geonym analysis.

Agrivoltaic panel shading influences soil nutrient dynamics and microbial community structure in a wheat-cultivated field

Agrivoltaic systems (AV), which combine photovoltaic panels with crop cultivation, represent a dual land use strategy to address food and energy security challenges under growing population pressures. However, the effects of AV systems on soil biogeochemical processes and microbial communities remain insufficiently characterized. This study investigated the effects of AV trackers with varying Ground Coverage Ratio (GCR) on soil physicochemical properties, and microbial community composition and diversity at two time points (T1 and T2). Field experiments were conducted in a wheat-cultivated field using two AV configurations: standard (STV1, GCR = 13%) and extended (STV2, GCR = 41%) trackers, compared to a full-sun control (CI). Soil chemical analysis performed using a combustion analyzer showed that organic carbon, total nitrogen, Olsen phosphorus, and exchangeable calcium, potassium, and magnesium were all increased under STV2 compared to both STV1 and the control; this overall enrichment was further confirmed by ICP-OES-based elemental profiling. Amplicon sequencing of the 16S rRNA and ITS1 regions revealed reduced bacterial and fungal diversity under STV2. Significant taxonomic shifts in bacterial and fungal communities were also observed at both phylum and genus levels relative to the control. These findings suggest that extended shading from STV2 altered soil chemical properties while restructuring microbial communities, reducing diversity and favoring copiotrophic taxa, especially at T1. Overall, the results underscore the central role of agrivoltaic design, specifically shading intensity and GCR, in shaping belowground processes with relevant implications for soil health and the agroecological performance of dual land use energy–food systems.

Emergency packets handing in queue aware congestion avoidance schemes in IoHT

The Internet of Healthcare Things (IoHT) emerged as an enabling technology that is strongly required for real-time patient monitoring and emergency healthcare services. However, the massive number of messages sent from heterogeneous sensors during the emergency situations generally leads to congestion in the network, causing it to overflow the buffer and lose some vital packets. Existing queue-aware congestion avoidance schemes based on priority queue or FIFO scheduling methods fail to differentiate among several levels of severity in emergency packets, resulting in the delayed or dropped transmission of packets. To addresses the issue of handling the emergency packets in the high and low-priority queues, we propose a scheme called Dual Queue Aware Congestion Avoidance, which operates with a dual queue structure combined with an Emergency-aware Packet Placement Algorithm (EPPA). The scheme classifies packets into high-priority and low-priority queues according to sensor type and assigns an emergency flag to indicate the severity. Unlike conventional FIFO approaches, emergency packets are placed immediately after the currently processed packet to ensure timely treatment without disrupting the ongoing transmission. A large number of simulations are carried out using NS-2.35 on Ubuntu. The results show that D-QACA significantly reduces buffer loss probability and high-priority packet loss—achieving up to 80% lower buffer loss and 90% fewer emergency packet drops compared to state-of-the-art schemes such as QACA and DCCA. The scheme enhances performance through an emergency-aware repositioning mechanism that allows critical packets to be served after the packet currently in processing.

Retraction: Mechanisms of adipocyte regulation: Insights from HADHB gene modulation

Research on the application of LLaVA model based on QLoRA fine-tuning in medical teaching

The augmented reality large language model medical teaching system (ARLMT) integrates augmented reality (AR)with a medical multimodal large language model (LLaVA) based on specifically designed for biomedical applications(LLaVA-Med), employing Quantized Low-Rank Adaptation (QLoRA) to advance medical education. Deployed on resource-constrained AR devices, such as INMO Air2 AR glasses, ARLMT overlays real-time visual annotations and textual feedback on medical scenarios to create an immersive and interactive learning environment. Key advancements include a 66% reduction in memory footprint (from 15.2 GB to 5.1 GB) through QLoRA, enabling efficient operation without compromising performance, and an average response time of 1.009 seconds across various medical imaging categories, surpassing the GPT-4 baseline in both speed and accuracy. The system achieves 98.3% diagnostic accuracy, demonstrating its reliability in real-time applications. By combining visual and textual elements, ARLMT enhances comprehension of complex medical concepts, providing a scalable, real-time solution that bridges technological innovation and pedagogical needs in medical training.

Artificial intelligence meets pediatric orthopedics: A comparative analysis of ChatGPT-4o, Gemini 2.0, and Claude 3.5 in detecting supracondylar humeral fractures

Background Supracondylar humeral fractures constitute 10–16% of pediatric skeletal injuries, requiring timely diagnosis to prevent neurovascular complications. Developmental variations in pediatric bone structures pose diagnostic challenges for clinicians. This study evaluated three next-generation large language models (LLMs) (ChatGPT-4o, Gemini 2.0, Claude 3.5) for detecting pediatric supracondylar humeral fractures and their classification according to the Gartland system. Methods This retrospective observational study included 300 pediatric patients (150 with supracondylar humeral fractures confirmed by expert consensus, 150 without fractures) aged 2–10 years presenting to the Emergency Department of the Bilkent City Hospital (October 2022-January 2025). Two-view elbow radiographs were presented to each LLM three times on different days. Diagnostic accuracy was evaluated using overall accuracy (all three responses correct), strict accuracy (≥2 correct responses), and ideal accuracy (≥1 correct response). Response consistency was assessed using Fleiss’ Kappa coefficient. Fractures were classified according to modified Gartland criteria. Results Gemini 2.0 demonstrated highest sensitivity (68.4%) followed by Claude 3.5 (58.7%) and ChatGPT-4o (19.3%) for fracture detection (p < 0.001). Ideal accuracy rates were 83.3%, 78.7%, and 27.3% respectively. Although ideal accuracy rates exceeded 91% in non-fracture cases, specificity remained low (33.1–36.0%), indicating a high rate of false-positive classifications. Response consistency was very good for ChatGPT-4o (κ = 0.69) and Gemini 2.0 (κ = 0.61), good for Claude 3.5 (κ = 0.44). For Gartland classification, Gemini 2.0 achieved highest accuracy: Type I (83.3%), Type II (62.4%), Type III (68.7%). Conclusion Current LLMs demonstrate limited capability as independent diagnostic tools for pediatric supracondylar humeral fractures. Gemini 2.0’s 68.4% sensitivity indicates these technologies require specialized pediatric training before clinical implementation. However, their potential as assistive tools for triage and assessment warrants further development of pediatric-specific models.

ADRD: Detecting diffusion-generated images via adversarial perturbation induced reconstruction discrepancy

The rapid advancement of diffusion models has raised concerns about their misuse in generating deceptive visual content, motivating growing interest in AI-generated image detection. Many existing detection methods rely on image semantic features, but modern diffusion models are optimized to closely match the semantic structure of real images, reducing the effectiveness of semantic-based detection. An alternative line of work exploits differences revealed through diffusion reconstruction; however, most existing approaches treat reconstruction error as a static and passive metric, which can be sensitive to generators or post-processing, thereby limiting robustness. In this work, we propose Adversarial Diffusion Reconstruction Distance (ADRD), a detection framework that models diffusion reconstruction as a dynamic response process rather than a fixed descriptor. ADRD actively probes the reconstruction behavior by introducing perturbation in latent space and measuring how reconstruction deviations respond under identical perturbations. We empirically observe that real images typically exhibit larger and more variable reconstruction responses, while diffusion-generated images tend to show more stable reconstruction behavior, reflecting differences in their alignment with the diffusion model’s implicit data manifold. By characterizing reconstruction sensitivity instead of absolute reconstruction error, ADRD provides a complementary perspective to existing reconstruction-based detectors. Experimental evaluations on multiple benchmarks suggest that reconstruction response under controlled perturbations constitutes a meaningful signal for diffusion-generated image detection. The code is available at https://github.com/ezell-chou/adrd

Interference of phototherapy with blue LED light on the behaviour of mice infected with Toxoplasma gondii

Toxoplasma gondii is an intracellular protozoan capable of promoting physiological and behavioural changes in hosts. When female mammals acquire T. gondii for the first time during pregnancy, Congenital Toxoplasmosis (CT) can occur, posing a significant risk to the foetus. Due to challenges in diagnosing and treating CT, a new blue LED light therapy (BLLT) was proposed to eliminate parasites during placenta and tissue invasion in mice; however, its effects on the behaviour of the animals are unknown. Thus, behavioural analysis was carried out in pregnant infected Swiss mice under BLLT (applied continuously for 12 hours, from 7 am to 7 pm, maintaining an intensity of 460 nm and 7 μW/cm²). Infected mice, regardless of light exposure, exhibited increased inactivity and reduced maintenance behaviours. BLLT influenced certain behaviours, such as an increase in abnormal behaviours in infected mice and higher food and water intake in non-infected mice, suggesting a potential stress effect. Grooming decreased under BLLT in infected mice, while affiliative interactions were reduced in non-infected mice under conventional light. Activity levels were largely unaffected in infected mice, but blue light exposure increased activity in non-infected mice. While our study highlights the potential of BLLT to reduce parasite load, further research is needed to investigate the long-term behavioural and physiological effects, as well as to clarify the mechanisms involved. In conclusion, BLLT may mitigate parasite load and influence behavioural outcomes, while prolonged exposure can act as a stressor, moderately affecting welfare. Future research should explore optimized light regimens, integrate physiological and behavioural measures, and evaluate long-term effects to balance therapeutic efficacy and animal welfare.

Localized sample-based quantum diagonalization for strongly correlated chemistry

We develop a hybrid quantum-classical workflow combining sample-based quantum diagonalization (SQD) and the localized active space self-consistent field method (LASSCF) to solve for the ground states of transition-metal complexes, a longstanding challenge for both classical and quantum algorithms. The resulting approach, named LASSQD, integrates quantum sampling with fragment-based multireference theory to reduce the computational cost of solving strongly correlated active spaces. We test LASSQD on multiple iron-based complexes and demonstrate that it agrees with LASSCF within 1 kcal/mol, albeit at a much reduced computational cost. The cost reduction originates from the use of a sparse approximation of the exact and combinatorially large ground-state wavefunction, which also enables LASSQD to treat fragment sizes that are computationally inaccessible to LASSCF, as demonstrated by our computation of the spin gap of iron-porphyrin. These results establish that LASSQD is a scalable strategy for generating reliable multireference wave functions, providing a robust starting point for post-SCF correlation methods that recover dynamic correlation beyond the active space, and a promising pathway toward quantum-enhanced electronic structure calculations.

Ocular injury awareness, knowledge, and safety practices among dental professionals, students, and supporting staff: A cross-sectional analysis

Background Ocular health is important in dentistry because the profession relies on visual acuity and has a high risk of splatter, aerosols, and projectiles. There is a great deal of promotion with respect to personal protective equipment (PPE), but little is known about how the awareness is turned into protective behaviour in dental practice in Indian contexts. Objective To assess the extent of knowledge, awareness, and clinical practices related to ocular safety among dental professionals, students, and auxiliary staff in India, and to identify significant discrepancies between their theoretical knowledge and practical implementation. Methodology An online cross-sectional survey was conducted among 512 dental professionals across various institutions in India by using a 27-item questionnaire that was diligently piloted and validated. The questionnaire serves to assess self-reported awareness, the incidence of ocular injuries, and patterns of eye protection for both clinicians and their patients. Results Although 86.1% of respondents reported being familiar with eye safety regulations, 62.5% reported eye injuries during clinical procedures, mainly as a result of spatter, aerosol, and removing restorations, which doesn’t suggest lack of protection but lack of protective eyewear. Responsive to the conditions mentioned, only 42.6% were compliant overall and consistent in wearing protective eyewear, while 93.6% report acknowledging the need for protective eyewear. Additionally, 13.3% had observed ocular injuries in patients, and 38.35% admitted to not providing eye protection during procedures. Conclusion There is a significant differential between the high awareness and poor uptake of ocular safety practices in Indian dentistry. This requires enhancing policy in institutions, introducing ocular protection in dental education and establishing standard procedures. Reducing preventable injuries and ensuring the safety of patients and practitioners is justified.

Molecular diversity of deep-sea fishes (Actinopterygii: Teleostei) in the western South Atlantic: A high diversity and new findings revealed by DNA barcoding

The deep sea hosts approximately 15% of all fish species, but the difficulty of sampling limits our knowledge, causing large gaps in geographic distribution, and many deep-sea fish groups still require taxonomic revisions. Herein, the deep-sea fishes collected in the southern Brazil, western South Atlantic, are studied through DNA barcoding resulting in a better understanding of the biodiversity and genetic diversity of the region. The samples were obtained during oceanographic cruises aboard the Brazilian R/V Alpha Crucis, focusing on the continental slope off southern Brazil at depths of 200–1,500 meters. A total of 170 sequences of actinopterygian fishes from 102 different species of 49 families and 19 orders were generated. The DNA barcoding identified 84 sequences at the species level, with the remaining being identified at least to the genus level. A high level of genetic divergence between species was observed, ranging from 3.9% to 36.7%. We provided the first COI sequences for 13 species that, to date, were not represented in databases. Among the species already represented in online databases with COI sequences, we expanded the geographic coverage of existing data. As a result, we increased representation for 16 species in the Atlantic, 26 species in the South Atlantic, and 28 species in the western South Atlantic. Our efforts revealed a diverse deep-sea fish fauna in the region, highlighting new occurrences and demonstrating a high genetic divergence between some taxa.

Multiplexing in networks and diffusion

We study multiplexing—the extent to which people have multiple, distinct types of relationships with the same others. We document empirical multiplexing patterns in Indian village data. We show that relationships such as socializing, advising, helping, and lending are correlated but distinct, while commonly used proxies for networks based on ethnicity and geography are nearly uncorrelated with actual relationships. We also show that these layers and their overlap affect information diffusion in a field experiment. On the theoretical side, we show that increased overlap in people’s different types of relationships impedes the spread of simple contagions—such as diseases or basic information that spread through a single interaction. In contrast, such higher levels of multiplexing enhance the spread of a complex contagion (e.g., a behavior or belief that requires multiple interactions for transmission) when infection rates are low, but impede complex contagion if infection rates are high. Finally, we identify empirical differences in multiplexing by gender and connectedness.

Embedding routine hearing health checks within existing Meals on Wheels services – A protocol for the SOUND-BITES Program pilot study

Introduction There is a high prevalence of undiagnosed and untreated hearing loss among older adults. Due to finite resources, task shifting to trained non-specialists is a strategy to improve equity of access to hearing health care. This pilot study aims to implement and evaluate a novel model of hearing care known as the SOUND-BITES program. This will leverage evidence-based mobile health technologies (Arclight otoscopy and Sound Scouts hearing screening app) and partnerships with Meals on Wheels New South Wales and Master of Clinical Audiology students. Methods SOUND-BITES provides a hearing health assessment involving otoscopy using Arclight, hearing screening using the Sound Scouts tablet-based app, and hearing health education to consenting Meals on Wheels clients and their household members. Volunteers from Meals on Wheels and audiology students will be trained to deliver the screening, education and referral components of this program. This pilot study utilises a mixed-methods approach to evaluation. The primary outcomes are program acceptability and feasibility; and secondary outcomes are actions taken by clients to address hearing loss, changes in hearing handicap and cost-benefits six months after program completion. Quantitative data will be collected from a pre- and post-program survey, and qualitative data from an optional post-program interview with clients and volunteers. Findings from the project will be disseminated through peer-reviewed journals and conferences as well as to relevant communities including the Meals on Wheels network.

YOLO-Net: A lightweight edge-enhanced detection model for small-object recognition in tennis match scenarios

The rapid advancement of deep learning has enabled intelligent analysis in professional sports, yet tennis remains particularly challenging due to small and fast-moving objects, frequent occlusions, and complex backgrounds. To address these difficulties, we propose YOLO-Net, a lightweight detection framework tailored for tennis event analysis. Built upon YOLO11n, the framework integrates three task-oriented improvements: a C3k-MSEIS module for multi-scale edge enhancement and dual-domain feature selection to refine fine-grained boundaries; an ECA channel attention mechanism inserted after C2PSA to strengthen inter-channel dependency modeling and improve feature discriminability; and a Focaler-IoU loss function to emphasize hard and small samples while reducing localization errors. In addition, we construct and annotate a dedicated tennis dataset containing 6,648 images across three categories—player, racquet, and ball—covering diverse scenes, camera angles, and lighting conditions. Experimental results show that YOLO-Net achieves 84.5% precision and 78.2% mAP@0.5 with only 2.58M parameters, outperforming the YOLO11n baseline by 2.5% in precision and 0.9% in mAP while maintaining real-time inference. These findings demonstrate that YOLO-Net is an efficient, accurate, and deployable solution for applications such as referee assistance, tactical analysis, and intelligent broadcasting in tennis competitions.

Provable cluster-preserving visualizations with curvature-based stochastic neighbor embeddings

Stochastic Neighbor Embedding (SNE) algorithms like UMAP and tSNE often produce visualizations that do not preserve the geometry of noisy and high dimensional data. In particular, they can spuriously separate connected components of the underlying data submanifold and can fail to find clusters in well-clusterable data. To address these limitations, we propose EmbedOR, a SNE algorithm that incorporates discrete graph curvature. Our algorithm stochastically embeds the data using a curvature-enhanced distance metric that emphasizes underlying cluster structure. Critically, we prove that the EmbedOR distance metric extends consistency results for tSNE to a much broader class of datasets. We also describe extensive experiments on synthetic and real data that demonstrate the visualization and geometry-preservation capabilities of EmbedOR. We find that, unlike other SNE algorithms and UMAP, EmbedOR is much less likely to fragment continuous, high-density regions of the data. Finally, we demonstrate that the EmbedOR distance metric can be used as a tool to annotate existing visualizations to identify fragmentation and provide deeper insight into the underlying geometry of the data.

Psychiatric nurses versus psychiatrists and pharmacists 'knowledge on polypharmacy practices in psychiatry: An interprofessional mixed-methods exploration

Background Polypharmacy is often crucial for managing complex and treatment-resistant psychiatric disorders, yet it carries risks such as adverse drug interactions, medication non-compliance, and suboptimal health outcomes. Interprofessional perspectives on polypharmacy significantly influence clinical decision-making and prescribing practices. Aim This research evaluates healthcare providers’ knowledge and attitudes regarding psychiatric polypharmacy, comparing the views of psychiatric nurses, psychiatrists, and pharmacists. It also explores how these factors impact prescribing behaviors and interprofessional collaboration. Methods A convergent mixed-methods approach was employed at the Erada Complex for Mental Health and Addiction in Jeddah, Saudi Arabia. The study involved 221 healthcare providers, including psychiatrists (n = 32), psychiatric nurses (n = 158), and pharmacists (n = 31). Quantitative data were collected using validated scales to assess knowledge and attitudes, while qualitative insights were gathered through open-ended responses and group discussions. Results Knowledge levels varied among the professionals, with psychiatrists possessing the most comprehensive understanding (84.2 ± 11.0), followed by pharmacists (81.5 ± 10.0) and psychiatric nurses (79.5 ± 9.8). Attitudes toward polypharmacy also differed, with psychiatrists showing the most favorable views (3.79 ± 0.49), whereas nurses and pharmacists were more cautious due to concerns about adverse effects and medication burden. A significant positive correlation (r = 0.653, p < 0.05) was observed between knowledge and attitude scores. Sociodemographic factors, such as professional experience and confidence in medication management, influenced both knowledge and attitudes regarding medication management. Qualitative findings highlighted interprofessional tensions, with psychiatric nurses advocating for more conservative approaches, psychiatrists emphasizing clinical necessity, and pharmacists focusing on optimizing medication safety. Conclusion Healthcare providers demonstrated varying levels of awareness and attitudes toward psychiatric polypharmacy, shaped by their professional roles and responsibilities. While psychiatrists were more accepting of polypharmacy, psychiatric nurses expressed concerns about patient burden, and pharmacists prioritized safety considerations. Enhancing interprofessional collaboration and ongoing education on polypharmacy practices are essential for improving patient outcomes.