The early prediction of student success through machine learning holds transformative potential for improving educational outcomes. However, the development of robust predictive models is fundamentally constrained by the scarcity of large-scale, high-quality educational datasets and privacy regulations. This paper introduces a machine learning framework that addresses these challenges through privacy-preserving synthetic educational data, with the key insight that synthetic data provides maximum value as a pre-training foundation rather than a complete replacement for real data. The framework is built upon SynEdu-HEDL, a novel synthetic dataset comprising 20,000 students across 180 courses with six interconnected tables. The dataset was generated using a hybrid rule-based and probabilistic approach that provides strong privacy guarantees (no real student records are used). Using this synthetic foundation, we evaluate predictive architectures including LSTM networks with attention, transformers, graph neural networks, and a novel hybrid LSTM-GNN architecture. Experimental results on the real-world OULAD dataset demonstrate that while direct transfer from synthetic to real data yields limited performance (AUC-ROC = 0.714), fine-tuning with only 5% of real data achieves an AUC-ROC of 0.781-a 12.7% relative improvement over training from scratch on the same limited real data (0.693). With 20% real data, fine-tuning reaches AUC-ROC of 0.831, approaching the full-data upper bound of 0.842. This synthetic pre-training benefit represents the central practical contribution of this work. Temporal analysis reveals that meaningful early warnings can be generated by week four, with the LSTM-GNN model correctly identifying 47.3% of students who would ultimately drop out. Fairness evaluation identifies acceptable disparities for gender and program type but meaningful differences for first-generation students (disparate impact ratio 0.84 on OULAD), which is successfully mitigated with only 1.8% reduction in predictive performance. Domain gap analysis reveals that performance degradation stems primarily from differences in forum participation (KS=0.34) and login frequency (KS=0.21) between synthetic and real data. The SynEdu-HEDL dataset and all code are publicly released. The framework provides validated evidence that synthetic educational data offers the greatest practical value when used for pre-training, enabling institutions with limited historical data to develop effective early warning systems without compromising student privacy.
Healthcare fog computing enables low-latency processing of sensitive medical data by distributing computation closer to data sources such as medical IoT devices and edge systems. Despite these advantages, the decentralized and heterogeneous nature of fog environments introduces substantial challenges related to security, privacy, and performance, particularly under sophisticated cyber threats, data tampering risks, and emerging quantum computing attacks. To address these challenges, this paper proposes Q-ZeroFog, a post-quantum security framework designed to enhance privacy, resilience, and performance in healthcare fog computing environments. Q-ZeroFog integrates a quantum-resistant blockchain layer, an AI-driven software-defined networking (SDN) control plane, and an adaptive zero-trust security model to ensure secure and efficient operation of healthcare fog networks and e-health applications. The framework employs post-quantum cryptographic schemes to protect sensitive healthcare data against quantum-enabled attacks, reinforcement learning-based SDN control for dynamic traffic optimization, and digital twin technology for proactive fault detection and autonomous recovery of fog nodes. In addition, homomorphic encryption is applied to lightweight healthcare data analytics (e.g., aggregation and threshold-based computations), enabling privacy-preserving processing without exposing raw patient data. Extensive simulations conducted in a controlled, permissioned healthcare fog environment demonstrate that Q-ZeroFog consistently outperforms existing security models across multiple performance metrics. The framework achieves intrusion detection rates of 99.15%, 98.45%, and 98.00% under low, medium, and high network loads, respectively, surpassing benchmark models such as ZT-1 and FogGuard. Data integrity reaches up to 99.20%, while task completion rates exceed 97.85% across all load conditions. Furthermore, Q-ZeroFog delivers reduced latency (130-210 ms), lower energy consumption (95.5 Wh for fog nodes and 21.75 Wh for IoT devices), improved scalability, and minimal privacy leakage (as low as 1.25% under high load). These results validate Q-ZeroFog as a scalable, privacy-preserving, and high-performance post-quantum security framework capable of meeting the stringent requirements of latency-sensitive healthcare fog computing applications.
Path planning for cloud based autonomous systems such as smart transportation, Internet of Things (IoT) installations and robot fleets need to be secure, energy efficient, time efficient and fulfil privacy constraints. Current reinforcement learning (RL) techniques mainly consider optimisation of single objective, centralised or loosely secured model updates which are susceptible to data poisoning, privacy breach and adversarial model updates. We present Blockchain enabled Energy and Time efficient Multi Objective Reinforcement Learning (BlockE2T MORL) a new decentralised approach for secure, cloud assisted path planning. BlockE2T MORL has three main components: (i) a dynamic multi objective reward function that reduces energy, travel time and security threat; (ii) a lightweight blockchain inspired trust mechanism that assigns continuous trust values to agents, and is incorporated in the reward function to punish dishonest or malicious agents; and (iii) a hybrid actor critic learning strategy that facilitates exploration and exploitation in dynamic environments. Unlike conventional blockchain systems, our approach incurs low computational overhead ([Formula: see text] instead of [Formula: see text] for validation) and operates without heavy consensus protocols. We evaluate BlockE2T-MORL on a simulated grid-based cloud environment with up to 100 agents and varying adversarial ratios. After 200 training episodes (5 independent runs), the proposed framework achieves: (i) energy consumption = 124.3 ± 8.7 J (23.4% reduction vs. standard RL, p < 0.01), (ii) latency = 45.2 ± 3.8 ms (17.4% improvement, p < 0.05), and (iii) trust score = 0.87 ± 0.04 (67% improvement, p < 0.001). The framework converges faster (210 ± 25 episodes in the final optimized configuration, compared with ≥ 520 episodes for baseline methods). BlockE2T-MORL offers a scalable, privacy-preserving, and computationally lightweight solution for next-generation intelligent path planning in cloud-based autonomous systems.
The study aimed to assess obstetric care providers' practices and associated factors towards birth companionship in public hospitals of South Gondar zone, North West Ethiopia, 2025. An institution-based cross-sectional study was conducted from January 01 to February 28, 2025 among obstetric care providers working in 10 public hospitals of South Gondar zone, North West, Ethiopia. A census of 218 participants was conducted. Data were collected using a structured, self-administered questionnaire. The data were entered into EpiData version 4.6 and analyzed using Statistical Package for Social Science version 27. Descriptive statistics were computed, and both bivariate and multivariate logistic regression analysis were performed, and statistical significance was declared at P< 0.05. Ethical clearance was obtained and informed consent was secured from all the participants. A total of 211 respondents participated in the study, making the response rate 96.7%, of these the prevalence of birth companionship practice among obstetric care providers was 16.59% (95% CI, 11.6-21.6). Obstetric care providers who had received training on Compassionate and Respectful care, Respectful Maternity Care, Basic Emergency Obstetric and Newborn Care were 1.55 times (AOR: 1.55; 95%CI: 1.09- 2.21) more likely to practice companion presence during childbirth. Additionally, health professionals working in facilities equipped with privacy measures between delivery couches were 1.48 more likely to permit companion presence during childbirth (AOR1.95; 95%CI: 1.08-2.02). Birth companionship remains poorly practiced among obstetric care providers in South Gondar zone public hospitals, pointing deep rooted gap between the existing policies and the implementation. Training on respectful maternity care and presence of privacy measures emerged as key factors shaping provider practice. Strengthening policy enforcement, embedding companionship with provider training programs and ensuring adequate privacy infrastructures in labor wards are essential steps forward to encourage of practice.
This cross-sectional study investigates healthcare practitioners' perceptions of the implementation of artificial intelligence (AI) to enhance patient safety culture in Riyadh, Saudi Arabia, conducted from March to June 2025. The focus is on understanding how AI is perceived in the context of improving patient safety culture and aligning with Saudi Vision 2030 goals. The study employed a cross-sectional design and administered self-administered online surveys via convenience sampling among healthcare practitioners across multiple healthcare settings in Riyadh. The study targeted a population including doctors, nurses, and allied health professionals. A structured questionnaire was developed to assess perceptions of AI, including key variables such as perceived AI benefits, concerns about data privacy, and the necessity of training. Descriptive statistics were computed to characterize the sample, including age, gender, and profession. Findings revealed that most participants perceived AI as supportive of patient safety through improved diagnostic accuracy, reduced medical errors, and streamlined workflows, which participants believed may contribute to a stronger patient safety culture. Descriptive analyses suggested variation in perceptions across professional groups. The study found that healthcare practitioners in Riyadh generally perceived AI as a potentially valuable tool for supporting patient safety culture, particularly through improving diagnostic accuracy and reducing errors; however, concerns about data privacy and insufficient training remain significant barriers that must be addressed to ensure effective and safe AI integration. This study highlighted healthcare practitioners' positive perception of AI's role in enhancing patient safety culture in Riyadh. While AI is seen as beneficial in improving accuracy and reducing errors, challenges such as data privacy concerns and a lack of training were identified as barriers to its implementation. Not applicable.
Access to real-world healthcare data is increasingly hindered by privacy concerns and stringent legal frameworks, including the General Data Protection Regulation (GDPR) and the Health Insurance Portability and Accountability Act (HIPAA). Consequently, synthetic healthcare data generators, notably Synthea™, have emerged as essential tools to generate high-quality datasets with minimal privacy and legal concerns. However, many simulation engines including, Synthea™, predominantly rely on command-line interfaces (CLI), imposing significant technical barriers for clinical researchers. Furthermore, raw output formats such as HL7 Fast Healthcare Interoperability Resources (FHIR) JSON create a substantial "interpretability gap" for end-users. To address these challenges, we present SyntheaWeb, a web-based platform that simplifies the generation and interactive inspection of synthetic patient cohorts. It provides a user interface, visual cohort dashboard, structured longitudinal patient records, semantic terminology linking (e.g., SNOMED CT, LOINC), and the capability for selective subset export.
The design of a privacy-preserved intrusion detection system for supply chain networks is challenging because of strict data privacy requirements, heterogeneous data distributions, and unreliable participating nodes. This study proposes BlockFedZTA, a framework that integrates federated learning, XGBoost, trust-aware aggregation, and a lightweight commitment-based integrity verification mechanism. In the proposed approach, each participant trains a local model and shares only a salted SHA-256 commitment without exposing model parameters. The aggregation mechanism assigns weights according to validation performance, reducing the influence of low-quality or potentially malicious updates. Experiments were conducted using a unified dataset containing 100,000 instances and 130 features representing five classes (Normal, DoS, Probe, R2L, and U2R), distributed among three organizations under non-IID conditions. The framework was evaluated under no-drift, moderate-drift, and severe-drift scenarios. Five-fold cross-validation produced average accuracies of 0.966, 0.964, and 0.963, respectively. Statistical analysis confirmed that the trust-aware aggregation strategy significantly outperformed FedAvg under drift conditions (p < 0.01). Additional comparison with FedAvg, Krum, Multi-Krum, Median Aggregation, Trimmed Mean, FLTrust, and FoolsGold revealed higher performance with an accuracy of 0.96470 in both mild and severe situations of the data drift problem. Moreover, our model was highly resistant to label poisonings, ensuring an accuracy of more than 0.962 even with a high level of 60%. Scaling analysis with up to 50 clients again confirmed high performance and superiority over FedAvg with moderate communication overhead. For example, communication costs went up from 1640.74 KB to 20451.89 KB per round; meanwhile, the number of audit log bytes needed rose only from 3.40 KB to 174.02 KB. Repeated runs of the algorithm ensured a stable average accuracy of 0.9647 with a standard deviation of 0.0002. Thus, BlockFedZTA ensures a robust federated IDS approach in a supply chain environment.
The protection of sensitive patient information is central to modern healthcare delivery and is legally established through the Health Insurance Portability and Accountability Act (HIPAA). For plastic surgery practices, HIPAA compliance presents unique and increasingly complex challenges regarding clinical photography, public-facing marketing, and digital communication. We conducted a focused review to synthesize essential regulatory principles, common pitfalls, and practical strategies for maintaining compliance when creating or operating a plastic surgery practice. We outline historical developments of HIPAA, foundational Privacy and Security Rule requirements, and considerations specific to photography, metadata, electronic communication, and patient consent. Further, we describe common violations encountered in routine workflows and the associated civil and criminal penalties enforced by the Office for Civil Rights. By providing practical, specialty-specific guidance, this article aims to help plastic surgeons strengthen patient privacy protections without compromising clinical efficiency, thus reinforcing the trust at the core of the patient-provider relationship. Level of Evidence: 5 (Risk) For image description, please refer to the figure legend and surrounding text.
Smart wearable devices are increasingly used in physical education to support data-driven monitoring and feedback. However, limited research has examined why K-12 physical education teachers continue using these devices. This study investigates teachers' continuance intention by integrating the Task-Technology Fit model and the Unified Theory of Acceptance and Use of Technology, with self-efficacy and perceived data privacy added as extended variables. Using convenience sampling, 755 valid questionnaires were collected from K-12 physical education teachers, including 500 male teachers (66.22%) and 255 female teachers (33.78%). The model was tested using covariance-based structural equation modeling (CB-SEM), and fuzzy-set qualitative comparative analysis (fsQCA) was used to identify configurations leading to high and low continuance intention. The CB-SEM results showed that task characteristics significantly predicted task-technology fit (β = 0.916, p < 0.001), whereas technological characteristics did not. Task-technology fit was the strongest predictor of continuance intention (β = 0.504, p < 0.001), followed by performance expectancy, social influence, self-efficacy, and effort expectancy. Perceived data privacy had a significant negative effect. The fsQCA identified four configurations leading to high continuance intention and five leading to low continuance intention, indicating that continued use results from multiple combinations of task-related, technological, psychological, and social conditions. This study advances understanding of teachers' technology adoption by integrating symmetric and asymmetric analytical approaches within a unified theoretical framework. The findings highlight the central role of task-technology fit and show that continuance intention is not determined by a single factor, but by multiple interacting conditions. These results provide practical implications for policymakers, school administrators, and technology developers seeking to support the sustainable integration of smart wearable devices into physical education and school-based physical activity promotion.
Integrating artificial intelligence (AI) has the potential to transform healthcare. AI can enhance medication management, patient outcomes, and streamline pharmacy operations. However, addressing AI limitations and pharmacists' concerns is essential to fully grasp its real future potential in healthcare and its impact on the workforce. A cross-sectional study was conducted using a previously developed pilot-tested online questionnaire with demonstrated content validity and reliability among licensed pharmacists in the UAE. The questionnaire assessed participants' demographics, perceived benefits (9 items), and concerns/barriers (18 items) regarding AI implementation in pharmacy practice. Responses were recorded using a 5-point Likert scale. Perception scores were analyzed using descriptive statistics (mean ± standard deviation) to reflect the distribution of responses. Internal consistency was high (Cronbach's α: benefits, 0.88; concerns/barriers, 0.90). Associations between participants' demographics and perception scores were analyzed using appropriate statistical tests, with significance set at p < 0.05. A total of 400 participants were invited to participate in the study, of whom 340 returned completed questionnaires (response rate: 85%). Overall, participants reported positive perceptions of AI, particularly in relation to multitasking and rapid data analysis (4.1 ± 0.9), improved service quality (3.9 ± 1.0), and enhanced patient follow-up (3.8 ± 1.1). However, perceptions of AI's clinical impact were more moderate, including improved patient outcomes (3.5 ± 1.1), reduced medication errors (3.3 ± 1.2), and reduced healthcare costs (3.1 ± 1.2). Major concerns included cybersecurity risks (4.2 ± 0.8), data privacy issues (4.1 ± 0.9), and potential job displacement (3.9 ± 1.1). In this study, AI was perceived as a promising advancement in pharmacy practice, particularly in enhancing operational efficiency, multitasking, and patient follow-up. However, significant concerns were identified, including data privacy, cybersecurity, job displacement, and the potential loss of the human element in patient care. Addressing these concerns requires ensuring that AI complements, rather than replaces, pharmacists' roles, supported by updated education, targeted training, and clear regulatory frameworks.
Telemedicine adoption in gastroenterology has accelerated rapidly in recent years. Virtual interventions offer disease activity control comparable to standard in-person care. However, factors determining patient satisfaction and personal perceptions of virtual clinics remain poorly explored. This study identifies the clinical, demographic, and socioeconomic predictors of clinic format preference among patients with inflammatory bowel disease. We conducted a cross-sectional observational study at King Fahad Medical City in Riyadh, Saudi Arabia. The study included adult patients diagnosed with Crohn's disease or ulcerative colitis. Authors collected demographic parameters, clinical covariates, and patient experience ratings using a structured questionnaire. A multivariable binary logistic regression model was employed to isolate independent predictors of virtual clinic preference. The cohort comprised 1,005 patients. Overall, 55.6% preferred in-person appointments and 44.4% preferred virtual clinics. Prior telehealth exposure served as the strongest predictor for future virtual care (odds ratio [OR] = 4.10; 95% confidence interval [CI]: 2.90, 5.82). Patients utilizing subcutaneous injections (OR = 2.02) or oral medications (OR = 1.81) demonstrated significantly higher odds of selecting virtual care. Conversely, an ulcerative colitis diagnosis predicted a preference for physical visits (OR = 0.63). Prioritizing ease of access (OR = 1.47) and the use of technology (OR = 2.14) underpinned virtual choices. Strict privacy concerns (OR = 0.57) and inadequate physician communication (OR = 0.05) deterred patients from the remote clinic. Prior telehealth experience and the use of home-based therapies drive virtual clinic adoption. Disease phenotype, reliance on intravenous treatments, and privacy concerns may necessitate accessible in-person care. Health care systems must abandon uniform digital strategies and implement tailored, hybrid care models to optimize resource allocation.
To explore the data risk perception structure and connotation in the entire process of generative AI-enabled nursing research and to identify healthcare management and training needs as applied to digital health developments. Nursing research highly relies on contextualized and unstructured data. General generative AI still faces shortcomings in professional adaptation, data governance, and responsibility definition, which may lead to risks such as privacy leaks, amplified bias, academic misconduct and accountability vacuums. The study focusses on the perceptions of Future nursing professionals. Purposeful maximum variance sampling was used to recruit 20 participants from 3 universities, and semi-structured one-on-one interviews were conducted. The report followed the COREQ Protocol checklist. Five data risk awareness themes were identified: data adaptation risk, data security risk, data quality risk, data ethics risk, and response risk, presenting risk concerns throughout the entire process of "use-generation-sharing-responsibility". The data risk perception of nursing master's students regarding generative AI-enabled nursing research presents a clear five-dimensional structure, unfolding along the chain of "input-processing-output-diffusion-attribution." This structure supports the development of a framework for defining boundaries of AI use, data governance, ethical compliance, and capacity building in nursing research settings and digital health developments.
This paper describes a scoping review that explored the use and usability of wearable devices in assistive living with a focus on barriers to the real-world use of this technology in the home for the elderly. Published research was reviewed from the databases: PubMed, CINAHL, IEEE Xplore, and Web of Science. Using Arksey's et al.'s scoping review methodology relevant studies were identified, resulting in 37 reviewed for thematic analysis. The thematic analysis resulted in specific themes to barriers and facilitators in usability. Themes include privacy and security, technical challenges, providing a sense of safety and continued independence, and knowing connection to support is available if required. Wearable technology can positively contribute to elder care but a number of key issues and barriers remain.
As the number of cancer survivors continues to grow, optimizing long-term survivorship care models has become increasingly important. Telehealth has the potential to improve access to health care for survivors; however, studies evaluating telehealth in this population remain limited. Additionally, concerns persist regarding equity in technology access and digital literacy. This study aimed to examine demographic factors and patient attitudes influencing telehealth use among cancer survivors compared to the general population. Adult participants were identified from the nationally representative database Health Information National Trends Survey 6 (HINTS 6). Multivariable logistic regression was used to calculate the predictors of telehealth use among cancer survivors. χ2 tests compared the prevalence of reported reasons of not using telehealth in the last 12 months between cancer survivors and the general population. A total of 5793 (weighted n=239,557,883) individuals were included in this study, 7.7% (weighted n=18,545,434) who are cancer survivors. 5092 individuals from the general population and 701 cancer survivors were included. Older age was associated with lower telehealth use (adjusted odds ratio [aOR] 0.11; 95% CI 0.02-0.59 for patients aged ≥65, compared to those under 40 y old). Higher education (aOR 2.55; 95% CI 1.24-5.27) and heart disease history (aOR 2.52; 95% CI 1.20-5.28) were associated with increased telehealth use. Employed (aOR 0.46; 95% CI 0.22-0.97) and retired (aOR=0.37; 95% CI: 0.18-0.77) cancer survivors were less likely to use telehealth than unemployed individuals. Of the nonusers, over 60% reported that telehealth options were not offered, and 80% preferred in-person visits. Technical issues and privacy concerns were not major factors in utilizing telehealth. Despite greater telehealth use among cancer survivors, a negative association between older age and telemedicine utilization persists. Efforts should focus on improving access for older cancer survivors and addressing employment-related factors, patient attitudes, and telehealth availability. Future studies should explore personalized approaches to enhance cancer survivors' health care experiences. Our findings emphasize the need to address specific factors including age and employment related disparities, patient preferences and telehealth availability to optimize equitable access to telehealth and enhance the delivery of cancer survivorship care.
Artificial Intelligence (AI) has improved orthodontic tasks, but clinical adoption remains fragmented. This paper presents a multi-layered framework that combines privacy-preserving data infrastructure, multimodal intelligence, and human-AI collaboration into one coherent system. Its main contribution is a structured design blueprint that integrates isolated AI tools into a clinically deployable ecosystem while addressing governance, integration, and trust.
Automated glaucoma subtype classification from clinical notes remains clinically unactionable without subspecialty-aligned explanations supporting clinician-facing deployment. We extended our Ci-SSGAN with a GPT-5.2-to-Qwen3-8B teacher-distilled reasoning module, fine-tuning Qwen3-8B on 2,660 de-identified ophthalmology notes using expert-reviewed rationales. On 294 notes, the fine-tuned model achieved ROUGE-L 0.792 ± 0.013 and BERTScore F1 0.955 ± 0.004, surpassing eight zero-shot comparators including GPT-4o and GPT-4.1, establishing privacy-preserving distillation as a path to interpretable AI.
Vibe coding tools enable software generation through natural language prompts powered by generative artificial intelligence, substantially lowering the technical barrier to application development. Their potential to empower health professionals without programming backgrounds is promising, yet remains largely unexplored in the literature. To analyze whether vibe coding tools effectively bridge the gap between non-technical health professionals and digital health application development, and to identify the enablers, barriers, and risks associated with their use. This study used a design science approach to create and preliminarily evaluate digital health artifacts. In a 3-hour in-person workshop, medical students used a vibe coding platform to build mHealth prototypes for dementia care challenges. The process included problem identification, ideation, prototyping, presentation, and reflection. Data came from participant observation, artifact review, and a post-workshop satisfaction survey. All three groups successfully developed functional prototypes within the allotted time. Key enablers included conversational accessibility, immediate visible results, and direct clinical applicability. Critical barriers included unfamiliarity with health data privacy regulations, absence of security measures in all prototypes, and a tendency to define excessively broad problem scopes. Notably, no group implemented user authentication or data encryption. Vibe coding effectively brings health professionals closer to digital application development, but this democratization is not sufficient on its own. It must be accompanied by training in data security, health informatics and AI fundamentals.
Medical imaging plays a crucial role in modern diagnostic practices, but traditional techniques often face limitations in accuracy, efficiency, and scalability. The emergence of deep learning (DL) has led to significant improvements that are transforming this field. This review discusses how DL algorithms are enhancing diagnostic imaging by improving accuracy, enabling automated analysis, and supporting personalized treatment plans. It focuses on key deep learning (DL) frameworks, including convolutional neural networks (CNNs), recurrent neural networks (RNNs), and generative adversarial networks (GANs). The review examines their applications in important medical imaging tasks such as image classification, segmentation, reconstruction, and disease prediction. It also considers how DL techniques are integrated with tools like radiomics, data augmentation strategies, and predictive analytics models. DL methods have shown superior performance in detecting and classifying diseases like pneumonia, tuberculosis, and Alzheimer's. They also improve the quality and speed of imaging modalities such as MRI, CT, and ultrasound. Despite these advances, challenges remain in data availability, model interpretability, clinical validation, and ethical issues related to bias and privacy. Addressing these challenges is essential for the successful clinical use of DL in medical imaging. This review ends with suggestions for future directions and best practices for ethically and practically integrating DL technologies into routine healthcare.
Remote patient monitoring (RPM) has become an important approach for managing patients with heart disease. This paper reviews current evidence on RPM and its impact on outcomes such as mortality and hospitalization. While RPM improves early detection and intervention, challenges remain, including data overload and integration into clinical workflows. To address these issues, this paper proposes an AI-enabled framework that integrates data collection, intelligent analysis, and clinical decision support. The framework aims to support earlier detection of risk and more effective clinical responses. Key considerations related to ethics, data privacy, and implementation within national digital health initiatives are also discussed. Overall, integrating AI into RPM offers a practical path toward more proactive and patient-centered care.
To evaluate the impact of the NLSE program on perioperative nurse capabilities, nurse and patient satisfaction, nursing practices, and patient safety, and to explore experiences related to leadership, workflow, safety culture, and standardization in 3 Ethiopian hospitals. A prospective quasi-experimental pre/postintervention mixed-methods study design was conducted from December 2023 to August 2024 at 3 major Ethiopian hospitals. The NLSE intervention comprised tailored training, mentorship, and nurse-led quality improvement initiatives. Quantitative data were collected from 9 fellows, 3 facility mentors, 112 perioperative nurses (65 at baseline, 47 postintervention), and 257 surgical patients using validated Likert-scale questionnaires. Statistical analyses, including paired t tests, χ2 tests, and ordinal logistic regression, were performed using SPSS Version 27. Qualitative data from 18 key informant in-depth interviews were analyzed thematically using ATLAS.ti. Nurse fellows demonstrated significant improvements in self-reported capabilities (knowledge, skills, and confidence) on quality improvement (QI), evidence-based practice, patient safety, and clinical leadership (mean difference: 0.2-0.8; P<0.001). Patient satisfaction scores also increased, with courtesy, comfort, privacy, and a quiet environment as significant predictors. Nurse satisfaction improved in areas such as motivation, recognition, and support. Qualitative findings supported these results, highlighting enhanced teamwork, communication, and patient safety culture. The NLSE program significantly enhanced perioperative nurse capabilities, nurse and patient satisfaction, and patient safety practices. These findings support the adoption of multimodal, nurse-led interventions to strengthen perioperative care and outcomes in low-resource settings. We highly recommend scaling up the NLSE program framework in such environments.