Clinical prediction is challenged by heterogeneity and noise in healthcare data. We evaluated SHAP-Enhanced Predictive Modeling (SEPM), a data-centric framework that ranks training observations using aggregated absolute Shapley Additive Explanations (SHAP) and optimizes subset retention through cross-validation. Across four clinical datasets and five classifiers, SEPM was compared with full-data training (TML) and randomized subset selection (SEPM-RAND) using nested cross-validation. A separate common-protocol XGBoost benchmark compared SEPM with Truncated Monte Carlo (TMC) Data Shapley estimation. Twenty simulation scenarios varied training sample size and outcome prevalence, with evaluation on full, high-SHAP, and low-SHAP test groups. SEPM achieved higher mean F1 scores than TML in 18 of 20 dataset–classifier combinations and SEPM-RAND in 16, retaining 42.1
Abstract Background Antibiotic overuse is a major driver of antimicrobial resistance (AMR), a global public health concern, particularly in sub-Saharan Africa where a high burden of infectious diseases and predominantly empirical diagnosis contribute to inappropriate use. The World Health Organization (WHO) AWaRe (Access, Watch, Reserve) classification supports rational prescribing by enabling the monitoring of antibiotic use patterns. However, data on AWaRe-based antibiotic use in Mozambique, especially in outpatient settings, remain limited. This study aimed to evaluate antibiotic prescribing patterns among outpatients using the WHO AWaRe classification in health facilities in Tete and Sofala, Mozambique. Methods Secondary data from outpatient prescriptions at two health facilities in Mozambique, Health Center No. 3 in Tete Province and Caia District Hospital in Sofala Province, were analyzed. Patients of all ages attended between November 2025 and January 2026 with at least one antibiotic prescribed were included ( n = 387). Antibiotic prescribing patterns were evaluated using the WHO AWaRe classification and WHO/INRUD prescribing indicators. Results Among the 387 prescriptions included, a total of 435 antibiotics were prescribed. Most patients received one antibiotic (89.4%), while 6.7% received two and 3.9% received three or more. Cotrimoxazole was the most frequently prescribed antibiotic (42.8%), followed by amoxicillin (20.2%) and metronidazole (11.3%). According to the AWaRe classification, 85.7% of antibiotics belonged to the Access group and 12.4% to the Watch group, with no prescriptions from the Reserve group. All antibiotics were prescribed by their generic names and were included in the essential medicines list. Total antibiotic consumption was 121.91 DDD/1,000 inhabitants/day, with amoxicillin (50.22) and ciprofloxacin (23.59) showing the highest utilization rates. Conclusion Antibiotic prescribing among outpatients in this study met the WHO target for the Access group, indicating overall alignment with recommended practices. However, the use of Watch antibiotics, particularly macrolides, highlights the need to strengthen antimicrobial stewardship efforts to ensure more rational use. The implementation of WHO recommendations, including the AWaRe classification, together with improved adherence to clinical guidelines and targeted training of healthcare providers, is essential to optimize prescribing practices in outpatient settings.
In Germany, the dual system of statutory and private health insurance can contribute to healthcare inequalities. It remains unclear if and how far nursing care in hospitals is affected. The objective of this scoping review was to map the existing evidence on social inequalities in nursing care in German hospitals, with an emphasis on the role of health insurance type where reported. A scoping review was conducted following the JBI methodology and reported according to PRISMA-ScR. The databases PubMed, Scopus, CINAHL, and Thieme were systematically searched for records from January 2000 to September 2025; backward and forward snowball sampling and an exploratory Google Scholar search were additionally conducted. Zotero was employed for reference management. Eligibility was assessed following the Population, Concept, Context framework. Study selection was performed by a single reviewer with consensus sought for ambiguous cases. Data were extracted into an Excel spreadsheet and synthesized narratively, categorizing findings thematically. Of 188 initial records, eight studies met the inclusion criteria, comprising four theoretical-conceptual works, one qualitative study, and three quantitative studies. Findings were organized around five themes, including Bourdieu’s capital and social milieus, mechanisms of nursing inequality, systemic versus individual level, diversity-related barriers, and the influence of insurance type. Only one study examined insurance-based differences, finding no significant difference in perceived waiting times for nurses, highlighting the limited empirical evidence on insurance type in nursing in German hospitals. Future research should empirically examine the role of insurance type in nursing care disparities, adding intersectional dimensions and quality measures, including patient-reported experiences. Open Science Framework (https//doi.org/10.17605/OSF.IO/RYV6U, Registered on October 5, 2025).
This conceptual paper undertakes a critical examination of the potential of value co-creation to transform healthcare delivery in Africa, a continent grappling with myriad healthcare challenges. By synthesizing existing literature and leveraging the Service-Dominant Logic framework, the study identified and categorized key barriers to value co-creation in African healthcare into two distinct categories: structural barriers (including technology, physical environment, service design system, and fragile healthcare system) and human-centered barriers (encompassing patient empowerment, limited health literacy, and cultural barriers). Through a qualitative desktop research approach, this study develops a novel conceptual framework that illustrates the value co-creation process in African healthcare, underscoring the imperative of addressing these multifaceted barriers to facilitate effective co-creation. The analysis contributes to the understanding of value co-creation in the African healthcare context, highlighting its potential to enhance patient outcomes, improve satisfaction, and foster more resilient healthcare systems. The findings have profound implications for theory, practice, and policy, informing the development of more patient-centered, contextually relevant, and sustainable care models that can improve healthcare delivery in Africa. By reimagining healthcare through the lens of co-creation, this study offers valuable insights into innovative solutions for Africa’s complex healthcare challenges, with potential applications for policymakers, healthcare practitioners, and researchers.
Abstract The 1984 Bhopal Gas Tragedy remains one of the deadliest industrial disasters in human history, killing thousands instantly and exposing over half a million people to the toxic methyl isocyanate gas. Despite its profound lessons for emergency response, inter-agency coordination, and public health preparedness, a few educational interventions have leveraged this catastrophe as a teaching case for health professionals. We designed, implemented, and evaluated a comprehensive simulation-based educational intervention that recreates key aspects of the Bhopal disaster to prepare the students for disaster risk preparedness in complex emergency scenarios requiring multi-sectoral coordination. We engaged 59 students from two institutions in a five-station simulation exercise. Twenty-two Master of Public Health students served as facilitators following intensive training, while 37 undergraduate Allied Health Sciences students participated as learners. The simulation included personification exercises, role-play scenarios, pre-hospital emergency care, hospital surge management, and multi-sectoral policy response stations. We measured learning outcomes using 15-item pre-post assessments and gathered qualitative feedback through open-ended questions. There was a significant improvement (p < 0.001) in the competencies measured. The overall mean score for different learning aspects increased from 44.8 at baseline to 61.6 (a 37.4% gain) after the intervention. Qualitative analysis revealed four major themes: nearly all participants (96.6%) valued the experiential learning approach, while most recognized gains in systems thinking (93.2%), empathy (89.8%), and critical problem-solving (81.4%) skills. Both facilitators and participants rated the experience highly satisfying (mean scores 4.4–4.8 on 5-point scales). The study demonstrates that well-planned, historically grounded simulation exercises can powerfully enhance disaster preparedness education among health science students.
