Laparoscopic posterior sectionectomy (LPS) is a challenging liver resection procedure often requiring advanced techniques and equipment to ensure precision. This video demonstration highlights a cost-effective, resource-efficient technique for LPS using a single-energy device approach, showcasing safe and precise liver resection while emphasizing meticulous surgical techniques. An 81-year-old male patient presented with a solitary lesion in segment VII of the liver, close to the hepatocaval confluence. Preoperative imaging confirmed a well-circumscribed lesion without vascular invasion. LPS was performed using Ligasure, which served a dual purpose: 'Kelly-clysis' for parenchymal dissection while simultaneously sealing and dividing vascular structures. Key procedural steps included precise anatomical exposure, isolation of major vascular structures, and systematic parenchymal transection under continuous inflow control. Critical aspects of the technique were highlighted, including real-time intraoperative ultrasound guidance for vascular mapping and ensuring minimal blood loss. The procedure was completed successfully without intraoperative complications. Operative time was 220minutes, and estimated blood loss was 60mL. Cumulative Pringle clamp time was 72minutes. The postoperative course was uneventful, with the patient discharged on postoperative day 5. Histopathological analysis confirmed R0 resection. The video highlights critical operative strategies to mitigate challenges typically encountered during LPS without cavitron ultrasonic surgical aspirator, demonstrating the feasibility of this technique in resource-limited settings. This video highlights the feasibility of laparoscopic posterior sectionectomy using a cost-effective, resource-efficient, single-energy device approach, offering a practical alternative for surgeons. Proper planning, expertise, and use of conventional tools can ensure safe and effective outcomes, even in resource-constrained environments.

Simplification of Hematopoietic Stem Cell Transplantation and Continued FACT Accreditation: Strategies for Low- and Middle-Income Countries.

Temperature-sensitive medicines require controlled storage and transport conditions because unsuitable temperature and humidity exposure can compromise product quality and drug-delivery reliability. This study presents a low-cost Internet of Things (IoT)-assisted and lightweight machine learning framework for cold-chain risk assessment in temperature-sensitive drug delivery. A realistic cold-chain monitoring dataset was prepared to represent normal refrigerated storage, door-opening disturbance, transport exposure, freezing-risk events, humidity variation, and high-temperature excursions. Temperature, humidity, deviation-based variables, and exposure-duration features were used to classify storage conditions into Safe, Warning, and Unsafe states. Since unsafe excursions were naturally less frequent, bounded sensor-level augmentation was applied to construct an ML-ready dataset while preserving the raw monitoring profile for descriptive analysis. Four lightweight classifiers, namely Logistic Regression, Decision Tree, Random Forest, and support vector machine, were evaluated. Random Forest achieved the highest baseline performance with 0.9659 accuracy and 0.9636 macro F1-score. In feature-set ablation, the derived feature configuration without the direct risk score achieved 0.9636 accuracy and 0.9616 macro F1-score using seven features. The findings indicate that low-cost IoT monitoring combined with lightweight learning can support early warning and decision assistance in resource-constrained temperature-sensitive drug-delivery environments.

Obstructive jaundice has contributed a sizable burden of global mortality, morbidity, economic cost, and hospitalization worldwide cause by benign and malignant conditions. To assess the etiological pattern, predictors for postoperative complications, and short-term outcomes among patients undergoing surgical intervention for EHBO at Tikur Anbesa Specialized Hospital, Addis Ababa, Ethiopia, 2024. A hospital-based retrospective cohort study was conducted to evaluate the etiological patterns, outcomes, and predictors of complications following surgical intervention for extrahepatic biliary obstruction (EHBO). The study included 122 EHBO patients who were consecutively enrolled during the study period. Normality of continuous variables was assessed using the Shapiro-Wilk test; normally distributed data were expressed as means with standard deviations, whereas skewed variables were reported as medians with interquartile ranges. Univariate and multivariate binary logistic regression analyses were performed to identify predictors of 30-day postoperative complications. Results are reported as adjusted odds ratios (AOR) with 95% confidence intervals (CI), and a p-value < 0.05 was considered statistically significant. Among 122 patients undergoing surgical intervention for EHBO, the mean age was 51.8 years (SD 13.2), with a slight male predominance (53.3%). Most patients had good performance status (ECOG 0-1 in 91.8%). Malignant biliary obstruction (MBO) was present in 67 (54.9%), predominantly due to pancreatic cancer (23.0%) and periampullary tumors (18.9%). In MBO patients, mean serum bilirubin decreased from 18.8mg/dL preoperatively to 1.7mg/dL at 4 weeks, representing a 90.1% reduction. Curative resection was achieved in 22 (32.8%) patients (Whipple procedure in 19, bile duct excision in 3); the remainder underwent palliative surgery. The 30-day postoperative complication rate was 25.4% (31/122), with mortality of 1.6% (2/122). Surgical site infection was the most frequent complication (19.7%), followed by anastomotic leakage (6.6%). Multivariable analysis identified poorer ECOG status (AOR 6.1, 95% CI 2.0-18.4; p = 0.001) and jaundice duration > 8 weeks (AOR 2.1, 95% CI 1.7-6.3; p = 0.003) as independent predictors of postoperative complications. Our findings demonstrate that favorable surgical outcomes for EHBO are achievable even in resource-constrained environments. However, further mitigating postoperative morbidity requires targeted perioperative optimization of high-risk patients, specifically those presenting with prolonged jaundice (> 8 weeks), preoperative cholangitis, poor ECOG performance status, hypoalbuminemia, and malignant biliary obstruction (MBO). Consequently, implementing systematic risk stratification and addressing these key clinical predictors are imperative to minimize complications and optimize patient outcomes. Not applicable.

Access to reliable healthcare information remains a significant challenge in many low-resource settings, particularly in developing countries where shortages of healthcare professionals and inadequate healthcare infrastructure hinder timely access to quality healthcare services. This study presents the design, implementation, and evaluation of ePACK, an intelligent Artificial Intelligence (AI)-based health assistant developed to support symptom checking and preliminary clinical decision support within the Nigerian healthcare context. The system enables users to describe symptoms in natural language, which are processed using Natural Language Processing (NLP) techniques and analysed through machine learning algorithms to identify potential health conditions. Clinical recommendations are further validated using a rule-based decision-support engine derived from the Practical Approach to Care Kit (PACK) Nigeria clinical guidelines. The proposed framework integrates a responsive web-based user interface, NLP-driven symptom extraction, machine learning-based disease prediction, and a guideline-informed clinical recommendation engine. A Random Forest classifier was adopted as the primary prediction model due to its superior performance among the evaluated algorithms. The system was assessed using 200 patient cases obtained from healthcare facilities across Adamawa, Nasarawa, and Ondo States, Nigeria. Evaluation results demonstrated an overall diagnostic agreement rate of 78% when compared with physician-confirmed diagnoses, while user satisfaction assessments indicated positive perceptions regarding system usability, accessibility, and response efficiency. The findings demonstrate the potential of AI-enabled digital health assistants to enhance healthcare accessibility, support patient triage, and provide preliminary health guidance in resource-constrained environments. Furthermore, the study highlights the value of integrating machine learning techniques with locally adapted clinical guidelines to improve the relevance, safety, and effectiveness of digital health interventions within the Nigerian healthcare system.

The increasing penetration of photovoltaic (PV) systems requires accurate and stable short-term forecasting to ensure reliable grid operation under operational constraints. This paper investigates short-horizon multi-step PV power forecasting using one full year of high-resolution (5 min) real-world data from a 111-kW grid-connected rooftop installation. The forecasting problem is formulated as a direct multi-output supervised learning task with a 30 min prediction horizon. A comprehensive comparative evaluation is conducted across baseline (persistence), tree-based (XGBoost), and deep learning architectures (LSTM, GRU, and Temporal Convolutional Networks—TCN). Results show that deep learning models significantly outperform conventional baselines, with LSTM achieving the lowest normalized RMSE (≈10.3%), while TCN provides a competitive trade-off between predictive accuracy, temporal stability, and computational efficiency. The direct multi-step formulation was adopted to reduce potential error propagation effects commonly observed in recursive forecasting approaches. Beyond forecasting accuracy, the study evaluates computational complexity and inference latency to assess practical deployability in resource-constrained environments. The proposed framework demonstrates that high-resolution real-world PV forecasting can achieve both strong predictive performance and operational feasibility. These findings contribute to the development of robust short-term forecasting strategies for distributed renewable energy systems operating under regulatory export constraints.

Aspiration pneumonia is a prevalent and severe complication in stroke survivors, substantially affecting morbidity and mortality rates. Early identification of risk factors is crucial for implementing timely prevention and management strategies. We aimed to determine the prevalence and clinical associations of aspiration pneumonia in stroke patients admitted to Bolan Medical Complex Hospital, Quetta. This descriptive cross-sectional study looked at 201 stroke patients who were admitted over 6 months. Clinical, demographic, and radiological data were gathered utilizing a systematic proforma. Aspiration pneumonia occurred in 36.3% of patients. Hemiplegia was identified as the most significant independent association (odds ratio = 1.19, 95% confidence interval: 1.01-1.41, p=0.04). Nasogastric tube feeding was present in 77.1% of aspiration pneumonia cases. Dysphagia and dysarthria showed associations in univariate analysis but lost significance in multivariate modeling. Hospital-acquired cases exhibited elevated rates of fever, positive chest X-rays, and significant neurological impairments. Aspiration pneumonia is common among stroke patients, with motor impairments and mechanical feeding being major risk factors. Early screening and comprehensive management are crucial for decreasing its prevalence, particularly in resource-constrained environments.

MindGrab: A Spectrally-Motivated Architecture for Accessible Deep Learning in Neuroimaging.

ABSTRACT The large-scale application of IoT technology in the medical industry also brings the advantage of continuous monitoring of patients and timely detection of heart diseases; however, it also results in challenges like ensuring data privacy, resources, and management of medical records. The paper presents a secure hybrid machine learning framework incorporated with IoT for diagnosing heart diseases at their infant stage, mostly meant for healthcare settings with low resources. The proposed protocol integrates lightweight encryption to secure medical data during early-stage diagnosis and data transmission in IoT-based healthcare systems combined with machine learning models, while nature-inspired optimization improves feature selection, classifier tuning, accuracy, and convergence. A variety of cardiac datasets are extensively used for the system’s proposal, and it is evaluated under inconsistent and noisy IoT conditions to determine its generalization capacity, robustness, and interpretability, which are assessed through the deployment of explainable artificial intelligence techniques. The framework is also analyzed regarding scalability, computational efficiency, and security – performance trade-offs to verify its suitability for real-time deployment in large-scale IoT healthcare systems. The findings of the study indicate that the suggested method has a well-balanced integration of security, accuracy, interpretability, and efficiency, thereby giving a practical and reliable solution for intelligent IoT-based healthcare diagnostics.

Addressing the computational efficiency and cross-device generalization challenges faced by acoustic scene classification in resource-constrained environments, this study proposes a structure-aware dual-stream feature disentanglement framework based on harmonic-percussive source separation and asymmetric convolutions. The framework achieves structured decomposition of acoustic signals through HPSS techniques, performs independent modeling targeting temporal-frequency dimension differentiated characteristics through cascaded asymmetric convolution kernels, and realizes adaptive feature fusion through a dual attention mechanism. Systematic validation on the DCASE 2020 and TAU Urban Acoustic Scenes 2022 datasets demonstrates that the proposed method significantly reduces parameter count and inference time while maintaining competitive classification accuracy, exhibiting superior robustness compared to Transformer-based methods in cross-device scenarios and consistent generalization capability under reduced input duration conditions. This study provides a solution balancing performance and efficiency for acoustic scene classification deployment on edge devices and mobile platforms, with the proposed structure-aware feature disentanglement concept offering a new perspective for enhancing the interpretability of deep learning models through leveraging physical prior knowledge in the audio signal processing domain.

Pleural effusion is a common occurrence in many different types of health care facilities and has a broad spectrum of potential causes, from benign systemic disorders to potentially fatal infections and tumors. In developing countries and resource-poor environments where sophisticated imaging techniques, specialty laboratory tests, and trained personnel are not always available, timely and accurate diagnosis presents major obstacles. As a result, clinicians utilize both clinical judgment and basic testing to guide the diagnosis of pleural effusion using pragmatic, cost-effective methods. To review and synthesize available evidence on the diagnostic approach to pleural effusion, with a focus on strategies applicable in resource-constrained healthcare environments. The electronic database literature was reviewed through a narrative process. Literature was reviewed from electronic databases (e.g., PubMed, Scopus, and Google Scholar) through the use of keywords: pleural effusion, diagnosis in resource-limited settings, thoracentesis, etc., All articles that met the inclusion criteria (i.e., published articles in English that identify diagnostic methods and challenges faced when diagnosing patients with a pleural effusion in a low-resource setting) and that were published as guidelines, original research articles, and review articles were included in the review. Articles that were not related to diagnostic methods or simply dealt with elaborate methods of imaging were removed from the database before analysis. Clinical assessment, chest X-ray, and diagnostic thoracentesis are key to diagnosing pleural effusions in settings with limited resources. While Light's criteria remain commonly used to help differentiate between transudative and exudative effusions, their use will be limited by laboratory constraints. Where feasible, point-of-care ultrasound represents an incredibly helpful adjunct to these procedures. Tuberculosis and parapneumonic effusions are the most common causes of pleural effusion in low-resource locations, thus requiring a high degree of clinical suspicion. Simplified diagnostic algorithms that combine clinical presentation with findings from simple investigations can help direct management choices. To maximize the benefit of a systematic diagnostic evaluation of pleural effusion in resource-poor environments, an organized approach combining patient history taking (clinical assessment) with a limited range of readily available laboratory tests will yield the greatest results. Investing time in developing cost-effective resources for key diagnostic procedures, establishing standardized clinical pathways for diagnosing pleural effusion, and providing the appropriate level of education/training for physicians to use these resources effectively will enhance both the diagnostic accuracy of the clinician and the overall survival rates of patients diagnosed with pleural effusion in the low-resource setting.

The global food security is severely threatened by various bacterial and fungal diseases that significantly degrade the quality, yield and productivity of wheat crop. This increases the need for an accurate and efficient system to improve wheat yield and mitigate these losses by enabling early intervention. The dataset used in this research comprises of 10,000 images from brown rust, yellow rust, powdery mildew, loose smut diseases and healthy wheat plants. The existing neural networks, ensembling and transformer-based models used for classifying wheat diseases are limited by high computational resource requirements that leads to inefficient feature extraction. These challenges are addressed by proposing a customized, lightweight and optimized Swin-Streamlined High Accuracy and Reduced Parameters (Swin-SHARP) transformer, a lightweight and optimized transformer model that enhances feature extraction while significantly reducing computational overhead. In particular, Swin-SHARP results in 82.5% reduction (48.9M to 8.5M parameters), making it an attractive solution for resource-constrained environments. The extracted features are further optimized by integrating the Swin-SHARP transformer with boosted Model-Agnostic Meta-Learning (MAML) and a weighted ensembling strategy to enhance generalization and classification accuracy. Our proposed model achieves a remarkable 98.1% accuracy, significantly outperforming existing CNN-based solutions, ensemble approaches, transformer, and deep learning models. We also cross-validated our proposed model on an unseen wheat plant diseases dataset, achieving 95.57% accuracy. Our proposed model is also compared against prominent models such as Inception-v3, ResNet-18, and VGG-16, which outperforms them by 1.6%, 1.7%, and 2.6%, respectively. The comparison with existing state-of-the-art models, including Sequential CNN, SGDR-S, Inception-v3, Cereal Conv, Darknet-53 CNN, EfficientNet B3, GLNet, CNN & SVM, Customized CNN, CaiT-YOLOv9 and MSFNet revealed that our method outperforms them by 0.6%, 5.8%, 5.3%, 0.75%, 2.8%, 2.68%, 1.42%, 1.3%, 3.31%, 3.29%, and 2.4% respectively. These results demonstrate the effectiveness and practicality of the Swin-SHARP transformer for wheat disease classification, particularly for real-time agricultural applications on mobile and embedded systems aimed at early disease detection and crop management.

Brain stroke occurs due to blockage or rupture in the cerebral blood supply and represents a critical medical emergency requiring rapid and accurate diagnosis. However, manual interpretation of CT scans is time-consuming and may delay clinical decision-making. To address this challenge, this study proposes the Deep Neural Brain Stroke Detection (DNBSD) system, a lightweight deep learning-based framework for automated stroke detection from CT images. The proposed model employs a task-specific convolutional neural network (CNN) architecture consisting of Conv2D, MaxPooling, Batch Normalization, Flatten and Dense layers, containing only 1.67 million trainable parameters and a computational complexity of 0.2973 GFLOPs, making it suitable for resource-constrained clinical environments. Additionally, preprocessing techniques including image resizing and normalization were applied to optimize performance. The model is trained and evaluated on two publicly available datasets: Brain Stroke CT Image Dataset (BSCI) and Brain Stroke Prediction CT Scan Image Dataset (BSPCSI), each divided into training, validation, and testing subsets. Experimental results demonstrate that the DNBSD system achieves high performance, with an accuracy of [Formula: see text] and an AUC of [Formula: see text] on the BSCI dataset, and an accuracy of [Formula: see text] with an AUC of [Formula: see text] on the BSPCSI dataset, while showing improved performance compared with several baseline approaches and state-of-the-art deep learning models. To enhance interpretability and support clinical decision-making, explainable artificial intelligence techniques, including LIME and Grad-CAM, are integrated to highlight critical regions influencing predictions. Additionally, a web-based diagnostic tool is developed to enable real-time stroke prediction. The findings suggest that the proposed approach can serve as an effective and interpretable tool for automated stroke detection, with potential to enhance clinical diagnostic workflows.

Intracranial hemorrhage is life-threatening and requires prompt and accurate diagnosis. Non-contrast head computed tomography is the standard first-line examination, but detecting small hemorrhages and classifying multiple subtypes require substantial expertise. Workforce shortages and increasing diagnostic workloads, especially in emergency settings, further challenge timely decision-making. Artificial intelligence (AI)-assisted interpretation has shown promise for improving accuracy and efficiency. This retrospective study evaluated the effect of AI assistance on the diagnostic performance of radiologic technologists (RTs). We analyzed the data for 100 non-contrast head computed tomography examinations (50 positive and 50 negative for hemorrhage) obtained from the Japan Medical Image Database. The interpretations of the five RTs (5-12 years of experience) with and without AI assistance were compared with those of two radiologists. The detection targets were intraparenchymal, intraventricular, subarachnoid, subdural, epidural, and any hemorrhages. We calculated the Area Under the Receiver Operating Characteristic Curve (AUC), accuracy, sensitivity, and specificity. The differences in the AUC for the AI-assisted and unassisted readings were tested using the DeLong method with Bonferroni correction. Significant AUC improvements were observed for five of the 30 reader-task comparisons (17%) after Bonferroni correction. These improvements were all related to intraventricular (p = 0.0001 to 0.0071) and subdural (p = 0.0022 to 0.0071) hemorrhages. AI assistance significantly improved RT detection of challenging subtypes such as intraventricular and subdural hemorrhages. However, it did not improve the diagnostic accuracy for detecting any hemorrhage overall (p = 0.0689 to 0.9669). AI can strengthen the role of RTs within task-sharing models and help stabilize preliminary assessments, especially in emergency care and resource-constrained environments.

Diabetic foot ulcers (DFU) represent a major global burden, requiring more effective prevention strategies. Digital solutions have been proposed to improve DFU early-detection and reduce DFU burden, but user-centered solutions for low-and-middle income countries (LMIC) are lacking. The aim of this paper is to describe the process and lessons learned from the co-creation of (1) an intervention based on a mobile health (mHealth) app for early detection of pre-ulcerative lesions and prevention of DFU in people living with diabetes, and (2) the strategy to pilot test the intervention in rural and urban settings in Peru. We conducted a co-creation process between December 2024 and April 2025. Four workshops were held in Lima (urban) and four in Piura (semi-rural), involving 54 people living with diabetes, their caregivers, and healthcare professionals (HCPs). In workshop #1, patient journey mapping was used to identify barriers and facilitators to DFU care. Workshops #2 and #3 focused on the co-development of a mHealth app, while workshop #4 gathered participant feedback on study procedures for the upcoming pilot study. A total of 23 people with diabetes, 21 HCPs, and 10 caregivers participated in the workshops. Patient journeys revealed barriers to DFU management, including health system fragmentation, limited clinician training, and insufficient education for people with diabetes and caregivers. In workshops #2 and #3 a beta version of the mHealth app was tested by the participants who proposed recommendations on how to improve it and other features to be added. Study procedures for the coming pilot study of the developed intervention were tested with role play. Participants recommended community-based recruitment of participants for the pilot study, manuals and videos to train users, family engagement, WhatsApp reminders, a helpline, interim feedback calls, and graphical results in multiple formats. Participant engagement was sustained throughout the study, with participants contributing from barrier identification through to intervention and procedure design. This work describes the usage of a structured co-creation processes to develop a digital health tool for DFU prevention and detection. The co-creation process was conducted with the aim to develop an intervention that is contextually relevant, acceptable, and more likely to be implemented successfully in the Peruvian context. The processes described here offer a practical model for developing DFU-prevention interventions in similar resource-constrained environments. The upcoming pilot study will be essential to evaluate usability, adherence, and early signals of impact, ultimately guiding future scale-up.

The Joint Expeditionary Trauma Training (JETT) Course addresses critical gaps in military surgical readiness by providing the first joint, standardized training for Role 2 and small surgical teams. This scenario-based course prepares teams to deliver lifesaving care in austere, resource-constrained environments, ensuring operational effectiveness and reducing risk on future battlefields.

Structured physical activity programs are increasingly promoted to maintain functional capacity in older adults; however, evidence regarding their implementation in real-world community settings, particularly in socioeconomically vulnerable populations, remains limited. Multimodal interventions integrating physical and cognitive components may provide additional benefits, yet feasibility data from resource-constrained environments are scarce. This single-arm feasibility study evaluated the feasibility, safety, and adherence of a combined physical and cognitive training program delivered in a community aging center serving a socioeconomically vulnerable population with exploratory assessment of functional and cognitive outcomes. In this single-arm feasibility study, 50 community-dwelling adults aged > 55 years attending a community aging center serving a socioeconomically vulnerable population were enrolled following medical, physical, and neuropsychological screening; 46 completed the intervention and post-intervention assessment. Participants attended one-hour sessions conducted on alternate days for four months, alternating between structured physical exercise and supervised cognitive activities. Primary feasibility outcomes included recruitment, retention, adherence, and safety. Secondary exploratory outcomes included functional capacity assessed by the Six-Minute Walk Test (6MWT) and cognitive performance assessed using the abbreviated NEUROPSI battery. Within-participant changes were analyzed descriptively and inferentially. Of 50 enrolled participants, 46 (92%) completed the intervention and post-intervention assessment. Mean attendance was 18.7 ± 9.3h (range 3-39). No adverse events attributable to the intervention were reported. Mean 6MWT distance increased from baseline to post-intervention by 140m (Cohen's d = 1.58, p < 0.001), while oxygen saturation and blood pressure responses remained stable. Improvements in walking distance were moderately associated with session attendance. Modest increases were also observed in global cognitive performance at post-intervention. Given the single-arm design and interval between baseline and intervention, outcome findings should be interpreted as exploratory. A structured combined physical and cognitive training program was feasible and safe in a community aging center serving a socioeconomically vulnerable population, with high retention and no intervention-related adverse events. Improvements in functional capacity were observed, with exploratory evidence of cognitive change. These findings support the practicality of implementing supervised multimodal programs in real-world community environments and provide a foundation for future controlled studies.

Efficacy and safety of underwater endoscopic mucosal resection (UEMR) for non-pedunculated colorectal polyps: a prospective study IN Vietnam.

Epileptic seizure prediction using an Electroencephalogram (EEG) can improve patient safety by enabling early intervention, yet most existing approaches focus on short prediction horizons with limited personalization or computational efficiency. This study presents a unified deep learning framework evaluated across ten pre-ictal prediction windows up to 300 min before seizure onset, using recordings from 161 patients and 1023 seizure events. At the 5 min horizon, the generalized model achieved 96.30% accuracy and 91.62% sensitivity. Two complementary personalization strategies are introduced: incremental transfer learning, which progressively fine-tunes the generalized model using patient-specific data, and Dynamic Time Warping (DTW)-based similarity personalization, which constructs a morphology-aware training cohort from a single reference seizure. Personalized models consistently outperform generalized baselines, particularly at longer horizons, with the DTW-based approach achieving 89.68% accuracy using only 70 similar patients. Reliable prediction is demonstrated up to 60 min prior to onset, while model optimization reduces computational complexity with minimal performance loss, supporting deployment in resource-constrained clinical environments.

South Asia has the highest rates of suicide fatalities among women globally. Existing suicide prevention interventions are largely based on Western-informed risk reduction programs and theories. This pilot trial aims to test the acceptability and feasibility of a co-designed suicide prevention program for peri-rural Pakistani women called the Khushal Pur-Umeed Zindagi (KPZ, خوشحال پُرامید زندگی پروگرام), compared to enhanced usual care (EUC). The study seeks to innovate suicide prevention by grounding interventions in decolonized frameworks, leveraging peer mothers, and developing robust health system integration strategies. We will conduct a two-arm, mixed-methods, hybrid type 2, stratified pilot cluster randomized controlled trial in peri-rural areas of the Islamabad Capital Territory (ICT). KPZ is a decolonized, culturally salient brief intervention combining narrative-based safety planning and contact follow-up, delivered by peer mothers over 6months. EUC entails the WHO Mental Health Gap Action Programme (mhGAP) suicide prevention module, along with referral support delivered through Primary Health Care (PHC) by Medical Officers (n = 2) and Lady Health Worker teams (n = 11). KPZ is integrated into the PHC system and is implemented through close collaboration between peer mothers (delivery agents) and the government-employed LHWs. A cohort of peer mothers (n = 11), identified from the same communities as participants, will be trained and supervised for the study duration. We will enroll 50 women aged 18-45 with a child under 3years who report suicidal ideation and will conduct follow-ups at 3- and 6-months post-recruitment. Qualitative interviews with trial participants and periodic reflections from peer mothers will be conducted iteratively over 6months. A mixed-methods approach will be used to assess both clinical and implementation outcomes. Primary clinical outcomes include suicidal ideation severity and suicidal behaviors. Implementation outcomes include feasibility, acceptability, fidelity, appropriateness, and peer mothers' competence. Secondary outcomes assess additional clinical domains (depression, anxiety) and culturally relevant constructs (moral injury, cultural suicide cognitions). This pilot trial will provide evidence on the feasibility and acceptability of integrating a decolonized, peer-delivered suicide prevention program for women into the existing Pakistani health system. The findings have the potential for large-scale public health impact by improving the availability, accessibility, and quality of culturally appropriate suicide prevention interventions in resource-constrained and complex environments. Results will inform a larger definitive trial of a community-initiated suicide prevention program in Pakistan. NCT06208293.