Abstract: Artificial Intelligence (AI) systems are increasingly embedded in development contexts across the Global South, yet limited evidence explains how individuals within marginalized communities behaviorally adapt to these technologies beyond structural access and governance conditions. Building on prior framework-based analysis, this study examines the micro-level processes through which users internalize and operationalize AI-enabled systems in everyday livelihood and learning activities. A mixed-method sequential explanatory design was employed using the same population across urban, peri-urban, and rural settings, integrating structured surveys with ethnographic observations, digital usage tracing, and behavioral mapping. The findings identify three dominant adaptation pathways: instrumental adoption driven by efficiency gains, socially negotiated use shaped by contextual constraints, and reflexive adaptation linked to learning and trust formation. Quantitative analysis indicates that user agency significantly mediates the relationship between access and effective utilization, while qualitative insights reveal that learning styles and socio-cultural conditions influence the depth and sustainability of engagement. The study concludes that inclusive AI outcomes depend not only on infrastructure and governance but also on dynamic human–technology interactions, where cognitive engagement and iterative feedback mechanisms play a central role. These findings extend existing models by introducing a behavioral adaptation dimension critical for designing context-sensitive and sustainable AI interventions.

Abstract: Background: Mandibular condylar fractures continue to be a subject of debate, traditionally framed as a choice between open and closed treatment. However, this binary approach fails to adequately account for fracture-level anatomy, Temporomandibular joint (TMJ) involvement, and functional outcomes. Purpose: To present an imaging-guided, fracture-level-based algorithm that integrates radiologic evaluation, surgical approach selection, fixation biomechanics, and functional rehabilitation. Review Strategy: This invited review combines current evidence with a 13-year institutional experience involving 495 joints. High-resolution CT was used to assess fracture morphology, displacement, and ramal height, while MRI was selectively employed in intracapsular fractures to evaluate disc-condyle relationships when intra-articular involvement was suspected. Management decisions, including surgical approach and fixation strategy, were guided by an institutional algorithm tailored to fracture characteristics. Results: Implementation of this approach yielded consistent and predictable outcomes. Mouth opening improved from approximately 18.77 mm preoperatively to 40 mm at 6 months. Lateral excursions became symmetrical (~9.6 mm), occlusion was restored in all patients, and bite force returned to near-physiological levels. Pain scores showed near complete resolution within 1 month. Postoperative morbidity remained low, with predominantly transient facial nerve weakness. Conclusion: This imaging-guided, algorithmic framework provides reproducible functional outcomes and signifies a shift toward structured, anatomically driven management of condylar fractures.

Abstract: The proliferation of large language model (LLM) agents has enabled increasingly complex 2 multi-step automation; however, composing multiple agents into coherent systems intro3 duces significant orchestration challenges that remain poorly documented. This survey 4 examines LLM-based multi-agent orchestration from 2023 through early 2026 (literature 5 cutoff: March 2026). We propose a three-topology, one-adaptivity taxonomy—centralized, 6 decentralized, and hierarchical coordination topologies, each optionally augmented with 7 a dynamic/adaptive control axis—grounded in classical multi-agent systems theory and 8 recent empirical evidence. We compare four leading frameworks (LangGraph, CrewAI, 9 AutoGen/Microsoft Agent Framework, and OpenAI Agents SDK) along axes directly rele10 vant to practitioners: state-management granularity, token cost structure, failure-recovery 11 options, and design philosophy. The emerging protocol stack is examined in terms of why 12 MCP (agent-to-tool) and A2A (agent-to-agent) occupy complementary layers, how the 13 ACP–A2A merger signals protocol convergence, and where ANP’s decentralized-discovery 14 design fits. Production design considerations—state management, task planning, error 15 handling, scalability, and security—are evaluated with reference to published benchmarks. 16 We close by identifying five open challenges and proposing a six-dimension evaluation 17 framework for multi-agent coordination quality. This paper provides practitioners with 18 a decision framework spanning taxonomy, framework selection, protocol adoption, and 19 production deployment.

Abstract: AGI is often framed as a problem of aligning model objectives with human values or constraining agent behavior. That framing becomes incomplete once AI systems move into the infrastructures through which people and institutions perceive, evaluate, remember, and decide. Cognitive integrity is introduced as the first infrastructure of intelligence, in humans and AGI-mediated systems alike: the evolving capacity of a bounded system to maintain calibrated attention, trust, contestability, and decision under pressure. The central risk is not boundary change as such, but maladaptive boundary reorganization: transitions that leave persons or institutions unable to reform a viable, reality-linked, self-directing boundary after coupling with AI. This reframing surfaces a conceptual vocabulary for AGI governance centered on integrity boundaries and health, failed reintegration, cognitive rails, and successor-safe continuity.

Abstract: This work analyses 164 galactic rotation curves from the SPARC database and develops a field-based interpretation of the dark matter effect within the framework of the Universal Quantum Foam Hypothesis (UQSH). The empirical excess term C(r) = v^2_obs(r) − v^2_bar(r) reveals, after normalisation, a consistent structure of preferred dynamical regimes. Global fits identify two dominant states: a peak regime with scale parameter q ≈ 0.5–1.0, encompassing mainly low-surface-brightness galaxies and dwarf galaxies, and a diffuse regime with q ≈ 3.0, dominated by more massive spiral galaxies. Individual fits yield a distribution of roughly 62% peak systems, 26% diffuse systems, and 12% in the transition zone. An analysis of the dynamic factor D = g_obs/g_bar as a function of the maximum rotation curve radius reveals a statistically significant negative correlation (r = −0.31, p = 0.0001). Beyond approximately 50–80 kpc, D converges systematically toward 1. This empirical instability boundary marks the spatial range within which coherent field organisation produces measurable amplification. In the UQSH framework, light is interpreted as a spherically propagating tension front that follows the accumulated field geometry. In this picture, the convergence κ does not measure the instantaneous mass density, but the projected field curvature. An updated UQSH model of the Bullet Cluster reproduces the observed offsets between gas centres and κ-peaks of 219 kpc and 228 kpc without requiring an additional non-baryonic matter component. Within this framework, the dark matter effect is not attributed to missing particles, but arises as an intrinsic property of the field medium. Baryonic structures act as stable field configurations that spatially pre-stress the medium. Through continuous radiation, they excite the field and generate overlapping deformations. While individual contributions partially relax, their continuous renewal leads to a dynamically maintained, time-averaged field configuration that appears as a large-scale effective deformation. The nonlinear superposition of these contributions — bound baryonic mass, continuous excitation, and the resulting field organisation — produces a large-scale field tension that appears observationally as the dark matter effect. On galactic scales, the empirical instability boundary at approximately 50–80 kpc sets a natural spatial limit for coherent field organisation. In galaxy clusters, the contributions of many saturated structures superpose into a collective field tension that systematically enhances the lensing signal beyond the baryonic expectation. The universal fits show high internal consistency within each regime, with mean squared errors of MSE ≈ 0.016 in the peak regime and MSE ≈ 0.06–0.13 in the diffuse regime. This universality contrasts with expectations from continuous halo models and supports the interpretation of preferred dynamical states arising from field organisation.

Abstract: The big data revolution transformed how we think of data analytics in many ways. Critical amongst them are the somewhat interconnected ideas of volunteered geographic information, crowdsourcing, and the big data property of variety. The robust literature concerning conceptual neighborhood graphs in two of these cases considers objects whose datatypes are held stable between the relations under consideration. This, however, is a limiting factor in these three application spaces due to the unknown form that data will take. This paper considers two avenues for the conceptual neighborhood graph to take as directions for future research: discretization conceptual neighborhood graphs (changing between corresponding vector and raster spaces) and cartographic generalization conceptual neighborhood graphs (changing the form of the objects in question). This paper provides insights as to what considerations should be considered when embarking upon this idea and demonstrates these concepts applied to prior conceptual neighborhood graphs.

Abstract: Li-ion batteries (LIBs) are seeing increasingly widespread adoption across consumer electronics, electric vehicles, and grid-scale energy storage systems, yet their susceptibility to thermal runaway remains a concern. This study evaluates ethoxy(pentafluoro)cyclotriphosphazene (PFPN) as an electrolyte additive to improve electrolyte flammability and thermal stability without compromising electrochemical performance. Electrolyte flammability was quantified using Self-Extinguishing Time (SET) measurements, which revealed that PFPN significantly suppresses combustion. At 4 wt% PFPN, 67% of electrolyte samples failed to ignite despite extended ignition exposure, and the average SET decreased by 43% (from 51 s g⁻¹ to 29 s g⁻¹). Differential Scanning Calorimetry (DSC) further demonstrated improved thermal stability, with the onset of solvent decomposition delayed by ~30 °C at 4 wt% PFPN. Ionic conductivity modestly decreases (11%, from 10.26 to 9.12 mS cm⁻¹ at 4 wt% PFPN). Electrochemical testing showed negligible impact on battery performance. Graphite||Li and NMC811||Li half-cells containing PFPN exhibited comparable capacity retention to baseline cells. NMC811||Graphite pouch cells were used to further evaluate extended cycling and rate capability, PFPN containing cells demonstrated similar capacities even after prolonged cycling and high-rate operation. Overall, PFPN provides effective flame retardance at concentrations as low as 4 wt% while maintaining electrochemical compatibility, making it a promising additive for enhancing thermal stability of LIB electrolytes.

Abstract: Black-hole accretion systems exhibit a characteristic coexistence of activities: broad-band X-ray variability, hot coronae, wide-angle winds, and both steady and discrete jets. This coexistence suggests a persistently time-dependent magnetic background in which noisy fluctuations and explosive release are both essential. In this paper, we connect them all to intermittent magnetic reconnection and propose a Synchronized Spin Model (SSM) in which multiple local dynamos in a rotating accretion flow are represented as interacting macro-spins. Their synchronization, partial synchronization, excursion, and reversal define a compact set of collective variables that organize both timing statistics and large-scale morphology. In this picture, multiscale magnetic reconnection sustains coronal heating, flares, intermittent outflows, and discrete jet activity, while the same synchronization dynamics produce amplitude modulation and demodulation, providing a route to 1/f-like variability, rms–flux/Taylor-like scaling, and approximately log-normal statistics of the demodulated envelope. We further argue that, although the continuous flux distribution in black-hole systems is more naturally discussed in multiplicative or log-normal terms, broader event- catalog statistics remain useful for describing suitably defined burst hierarchies, particularly by analogy with solar and stellar flare systems. The hard/soft cycle of X-ray binaries is then interpreted as motion through magnetic state space.

Abstract: Olive leaf spot, caused by Pseudocercospora cladosporioides, is one of the main foliar diseases affecting olive crops. This study determined the sensitivity of eighteen Uruguayan isolates of P. cladosporioides to eighteen fungicides belonging to eight chemical groups. Mycelial growth inhibition assays were performed on PDA with increasing fungicide concentrations, and the EC₅₀ for each isolate-fungicide combination was calculated. Copper-based fungicides showed moderate levels of inhibition. Among contact fungicides, mancozeb exhibited the best inhibition compared with ziram and captan, while dodine showed a similar level of inhibition to ziram. Most fungicides from the benzimidazole, strobilurin and triazole groups demonstrated the highest efficacy with carbendazim, trifloxystrobin and mefentrifluconazole showing the greatest inhibitory activity. These findings guide the selection of fungicides for cercospora leaf spot management within sustainable pest control programs and establish a reference framework for future resistance monitoring.

Abstract: The paper proposes an implementation of Kolmogorov-Arnold networks (KANs) for the purpose of dynamic proportional-integral-derivative (PID) control tuning in first- and second-order linear systems under noisy and time-varying reference conditions. Toy da-tasets, based on instantaneous system error, output and reference trajectory, are used for training the networks and comparing KANs results over a performance of: i) a PID with fixed coefficients, taken from MATLAB’s Simulink PID Autotune; ii) an MLP based neural network (NN), trained on the same datasets; iii) a traditional adaptive PID scheme with gain scheduling; iv) an LMS-based online tuning approach. Results show that KANs out-perform MLPs and LMS even with less optimized datasets under noisy and quick-changing conditions and perform on par with methods, such as gain scheduling, while allowing for more flexibility and easier setup.

Abstract: The gut, brain, fat body, wing, epidermis, corpora allata, salivary gland, pheromone gland, prothoracic gland, and many other tissues are not included in the olfactory/chemosensory function. However, they are the tissues where CSPs are most prevalent, with the exception of the antennae and legs. In part-1, we proposed renaming the "chemosensory protein (CSP)" family to "4-Cysteine Soluble Proteins (4CSPs)" in order to avoid designating a protein present in hemolymph or eggs as chemosensory. In part-2, we broaden the report's focus from ubiquitous tissue distribution to potential intracellular functions in order to bolster our idea. We go over our studies on insecticide resistance, the Mp10 story in aphids, and other systems relevant to lipid transport and immunity. Most of the data gradually tilt toward the non-chemosensory features of 4CSPs, and this adds even more evidence to support those aspects. We provide a second review (part-2) and analysis that shows a stronger association between 4CSPs and mucins, translation initiation factors, and proteins belonging to the actin complex family. This is yet another compelling evidence that the benefits of renaming "chemosensory proteins" far outweigh the drawbacks, when taking tissue distribution and intracellular localization into account.

Abstract: Multi-Stage Transmission Network Expansion Planning (MS-TNEP) is critical for adapting power grids to long-term renewable integration. However, simultaneously incorporating N-1 security, active power losses, and spatial generation uncertainties imposes prohibitive computational complexity. This paper proposes a probabilistic MS-TNEP model evaluated over a 20-year horizon. To overcome intractability, a hybrid decomposition framework is employed, delegating discrete combinatorial investment decisions to an upper-level metaheuristic while resolving operational feasibility, power losses via fictitious nodal demand, and N-1 contingencies through lower-level linear programming. Furthermore, a novel Pack-Based Grey Wolf Optimizer (PBGWO) is introduced to enhance convergence in this constrained domain. The approach is validated on the modified Garver and the 46-bus Southern-Brazilian systems under multiple wind and conventional generation scenarios. Comparative analysis against the Genetic Algorithm, standard GWO, and Whale Optimization Algorithm reveals that PBGWO consistently identifies the optimal expansion schedules.

Abstract:

This study evaluates the techno-economic feasibility of LNG regasification alternatives, including offshore platform conversion, floating storage and regasification unit (FSRU) retrofit, and onshore LNG terminals, under conceptual design conditions at a capacity of 100 MMSCFD. The analysis integrates cost estimation, project schedule, and technical maturity within a multi-criteria decision-making framework based on the Analytic Hierarchy Process (AHP), combining quantitative techno-economic results with expert judgment to support structured comparison of alternatives. Cost estimation is conducted using two approaches, namely cost–capacity scaling and analogous estimation, to examine their influence on feasibility outcomes. The results indicate that the conventional scaling method, using an exponent of 0.6, produces inconsistent results across configurations, overestimating costs for offshore-based systems while underestimating costs for onshore LNG terminals. Back-calculation of effective scaling exponents yields values of approximately 0.43 for offshore platform conversion, 0.37 for FSRU retrofit, and 0.78 for onshore LNG terminals, demonstrating that cost–capacity relationships are configuration-dependent and cannot be represented using a single uniform exponent. The AHP evaluation, conducted under two scenarios based on the applied cost estimation methods, shows that offshore platform conversion consistently achieves the highest feasibility ranking, followed by FSRU retrofit and onshore LNG terminals. While the ranking remains unchanged, the choice of cost estimation method influences the magnitude of score differences, affecting the strength of preference among alternatives. These findings highlight the limitations of conventional scaling approaches and demonstrate that offshore platform conversion can serve as a cost-competitive and time-efficient alternative for LNG infrastructure development, particularly in regions with existing offshore assets.

Abstract: Rockfall from slope unstable rock masses, a typical geological hazard induced by brittle failure, is characterized by abrupt occurrence, negligible macroscopic deformation prior to failure, and extremely short lead time for early warning, posing a severe threat to the safety of mountainous transportation systems, water conservancy and hydro-power projects, and urban settlements. Conventional static analysis methods have sig-nificant limitations in real-time acquisition of damage evolution of structural planes and dynamic assessment of stability changes, which can hardly meet the practical re-quirements of early warning for unstable rock masses. The dynamic evaluation method for the stability state of unstable rock masses, based on the principles of structural dy-namics, establishes a correlation model between dynamic parameters (natural fre-quency, damping ratio, mode shape, etc.) and the damage degree of structural planes, providing a new paradigm for dynamic identification and quantitative evaluation of the stability of unstable rock masses. This paper systematically reviews the dynamic behavior mechanism and theoretical evaluation framework of slope unstable rock masses, and elaborates on the damage evolution of structural planes, the disturbance effect of environmental dynamic loads, and the key dynamic parameter system. The single-degree-of-freedom dynamic models and their theoretical derivation for three typical types of unstable rock masses (sliding-type, toppling-type, and falling-type) are thoroughly analyzed, and the cutting-edge advances such as multi-block chain collapse model and data-physics dual-driven surrogate model are reviewed. Meanwhile, the contact and non-contact monitoring methods based on Micro-Electro-Mechanical System (MEMS) and Laser Doppler Vibrometer (LDV) techniques, as well as the de-velopment status of cloud-edge collaborative intelligent early warning architecture, are systematically summarized. On this basis, the core challenges are pointed out, includ-ing the long-term evolution under multi-field coupling, high-fidelity inversion calcu-lation for large-scale rock masses, and the scientific correlation between early warning thresholds and failure probability. The full-life-cycle dynamic simulation based on digital twin is also prospected. The research results provide a systematic reference for the improvement of the theoretical system of dynamic evaluation of slope unstable rock masses and the engineering practice of disaster prevention and mitigation.

Abstract: Wearable electrochemical biosensors often producevoltammetric signals that are corrupted by noise and long-term drift.Effective on-device denoising is critical to improve signal quality anddetect anomalies due to sensor drift or interference. This paperexplores lightweight TinyML models for denoising and drift detectionin wearable sensor voltammograms under the strict memoryconstraints of microcontrollers. We apply compact 1D convolutionaland dense autoencoder networks, as well as a PCA-basedreconstruction, to remove noise and identify drifting signals. Using apublic NIST dataset of cyclic voltammograms with added syntheticnoise and artifacts, we evaluate each model’s denoising performance(signal reconstruction MSE) and drift/anomaly detection capability (ROC-AUC) versus its memory footprint (quantized int8 model size). Results show that a small Conv1D autoencoder (8KB weights) canreduce noise by 75% and achieve 0.89 AUC for drift detection,approaching the performance of a larger dense autoencoder (35KB)and outperforming PCA. We observe a trade-off between model sizeand generalization: the larger autoencoder nearly perfectly flaggedanomalies (AUC 1.0) but smaller models remain competitive whileusing 4–6× less memory. These findings demonstrate that drift-resilient signal enhancement can be achieved on-device with minimalresource usage, enabling more robust wearable electrochemicalsensing.

Abstract:

Rice (Oryza sativa L.) is a staple food, and its eating quality is largely determined by amylose content (AC), which is controlled by the Wx gene. Current low-AC alleles often lead to excessive reductions, limiting their utility. The promoter region, particularly the pyrimidine-rich region (PRR) from -45 to -7, is hypothesized to fine-tune Wx expression, yet its functional motifs remain unexplored. In this study, we applied prime editing (PE) to precisely edit the Wx promoter in the japonica cultivar Huaidao 5. We generated two novel alleles, Wx^1T/32d (deleting PRR1 [-45 to -19] and PRR2 [-11 to -7]) and Wx^5d (deleting only PRR2), which are absent in natural germplasm. These edits resulted in significant, fine-tuned reductions in AC (3.95% and 2.77%, respectively) without affecting grain transparency. Our results identify PRR1 and PRR2 as important functional motifs whose deletion synergistically downregulates AC. Furthermore, the edited lines exhibited improved taste without compromising agronomic performance. This study not only provides novel insights into the regulatory motifs of the Wx promoter but also demonstrates the precision and potential of PE for improving rice grain quality.

Abstract: Objectives: This study aimed to evaluate very low HBV DNA viral load below the limit of quantification and to identify associated risk factors in different patient groups, including individuals with chronic hepatitis B, occult HBV infection (OBI) and others. Methods: We retrospectively analyzed 390 patients with very low-level viremia (VLLV). HBV DNA levels were measured in plasma samples using real-time RT-qPCR. Serological markers were evaluated in serum samples using chemiluminescence microparticle immunoassay (CMIA). Demographic variables, HBV serological markers (anti-HBs, anti-HBe, anti-HBc), and DNA results were evaluated. Results: The study included 193 patients in the sustained virological response (SVR) group and 197 patients in the other group; 60 of these patients had occult hepatitis B infection (HBV DNA positive, HBsAg negative) and 137 had no occult hepatitis B infection. Very low viral load was more common in men (53.3%) and in individuals aged ≥50 years (63.3%). In univariate analysis, OBI was associated with anti-HBe (odds ratio (OR)=2.874, 95%CI:1.256-6.172, p=0.012), anti-HBc seropositivity (OR=2.368; 95%CI:1.216-4.527, p=0.008), and being ≥50 years age (OR=1.845, 95% CI:0.963-3.473, p=0.077). In multivariate analysis, being ≥50 years of age, anti-HBe positivity, and anti-HBc positivity were independently associated with OBI. Among SVR patients, VLLV was associated with older age and anti-HBc positivity, but no significant association was observed with gender, anti-HBe, or anti-HBs status. Conclusions: In patients with VLLV, anti-HBc and anti-HBe seropositivity, along with older age, were independently associated with lower HBV DNA levels. Although anti-HBe positivity reflects reduced viral replication, it does not indicate complete viral suppression and may be detected at very low viremia levels, particularly in occult HBV infection. These findings highlight the complex interplay between viral replication dynamics and host immune responses across the VLLV spectrum, and emphasize the need for individualized monitoring strategies. Further validation through prospective studies is warranted.

Abstract: In Europe, travel-associated Legionnaires’ disease (TALD) cases require timely environmental investigations to support risk assessment, rapid control measures, and prompt reporting of investigation findings to the European Legionnaires’ Disease Surveillance Network (ELDSNet). This study evaluated TALD-related environmental investigations conducted during 2025 and early 2026 in Crete, Greece, following notifications through ELDSNet. Overall, 30 notifications corresponded to 24 unique confirmed TALD cases with illness onset in 2025 and 24 implicated hotels, with some cases involving stays in multiple hotels and Regional Units and clusters identified in some implicated hotels. The investigation framework combined microbiological, physicochemical, and operational data, focusing on delays from symptom onset, notification, sampling, and laboratory reporting. Overall, 516 environmental samples were collected, of which 503 yielded valid analytical results. Among the 503 analyzed samples, Legionella spp. was detected at ≥50 CFU/L in 127 samples (25.25%). This included 123 samples positive for L. pneumophila (24.45%), of which 31 were serogroup 1 (6.16%). Concentrations exceeding the >1000 CFU/L threshold were recorded in 53 samples (10.54%). Operational indicators varied, with median values of 31.0 days for reporting delay (RD), 14.5 days from notification to first sampling (TTF), 47.5 days from symptom onset to first sampling (TDS), and 67.0 days from symptom onset to first laboratory result (OELR). These findings support documenting response delays, strengthening inspector capacity and cross-regional coordination, and integrating microbiological results with operational indicators for timelier source attribution, faster reporting, and TALD public health action.

Abstract: This study aimed to determine whether intranasal hinokitiol modulates short-term sal-ivary secretory IgA (SIgA) secretion dynamics and IgA antibody-forming cell (AFC) ac-tivity in the submandibular glands of aged mice, a model of age-associated mucosal immune decline. Aged BALB/c mice received intranasal hinokitiol (50 μg) once weekly for 4 weeks. Saliva was collected on days 0, 7, 14, and 21 at baseline, 0.5 h, 1.5 h, 3 h, and 6 h after each administration. SIgA levels were measured using an enzyme-linked im-munosorbent assay. On Day 21, IgA AFCs were enumerated using an enzyme-linked immunosorbent spot assay, and their viability and proliferative activity were assessed using the MTT assay. Salivary SIgA rose transiently after each dose, peaking at 1.5 h and returning to baseline by 6 h. By Day 21, baseline SIgA secretion was significantly higher than at Day 0, indicating a cumulative effect. IgA AFCs were unchanged in number, but viability and proliferation increased at 0.5 and 1.5 h, coinciding with SIgA peaks. Flow cytometry revealed significant expansion of B220⁺CD38⁺ memory B-cells; B220⁺CD138⁺ plasma cells were unaffected. Intranasal hinokitiol transiently enhances salivary SIgA secretion in aged mice, likely through short-term modulation of salivary gland immune activity. This non-invasive approach may aid mucosal defense in aging populations.

Abstract: Intimate partner violence is a serious public health issue with detrimental consequences on the victim's health. This study explores the perspectives of mental health professionals on their role in supporting women following their experience of intimate partner violence in the rural areas of the Eastern Cape, South Africa. The findings highlight the association between the experience of violence, poverty, lack of access to health services, and rural women's limited ability to escape abuse and recover. These structural factors contribute significantly to poor health outcomes, as this study found barriers preventing rural women from receiving counselling following the traumatic experience of abuse. Most notably, our findings uncover the impact of the shortage of shelters in rural areas, which often forces mothers to separate from their young children as they try to rebuild their lives after escaping the abuse. This unique insight distinguishes this study from previous work on this topic. The findings reveal a major gap in the public health response to intimate partner violence in rural areas of the Eastern Cape, particularly regarding the lack of accessible shelters for rural women. The study concludes that addressing the major shortcomings in the public health response to abuse of women in rural areas is critical to address poor health outcomes for women. We recommend increasing the quantity of shelters across the Eastern Cape and making them accessible to rural women and their children.