Dilla University

Data-driven diagnosis and treatment in the medical field are becoming more and more common, yet they are hindered by the shortcomings of the datasets available: they are limited, unbalanced, and require privacy protection. Nonetheless, such extreme scarcity, disproportionality, and privacy issues of medical datasets are major obstacles towards realizing robust and generalizable Artificial Intelligence (AI) models. Privacy-preserving generative frameworks are deemed critical to provide significantly advanced performance that will be able to boost medical data quality and variety without the need to affect the structural and clinical integrity. This research experiment aims to provide a new Generative AI-solution Medical Data Enhancement (GAIMD-E) framework with a dual-stream structure encompassing a Contour-Aware Generative Adversarial Network (CA-GAN) combined with a Variational Transformer Auto-Encoder (VTAE). The CA-GAN contains a contour computation component that explicitly learns the anatomical boundaries and the structural outlines of medical images, and has the capacity of generating high-quality synthesized examples with morphological consistency preserved. At the same time, the VTAE simultaneously represents local feature uncertainty as well as global temporal dependencies on multi-modal medical records (for example, radiology, pathology, and so on). Besides, a new Gradient-Guided Feature Consistency Loss (GGFCL) is proposed to achieve structural consistency in synthetic and real samples that can improve domain credibility. Also, experimental tests on three benchmark datasets Brain Tumor Segmentation (BraTS), Medical Information Mart for Intensive Care III (MIMIC-III), and National Institute of Health (NIH) Chest X-ray reveal up to 31.6% gain in model generalization, 27.4% decrease in false negatives and 21.2% increase in F1-score when training downstream classifiers on augmented data. These findings confirm the effectiveness of the framework to remedy data gaps when maintaining clinical applicability, and it is a viable and privacy-safe option towards advancements in AI-based medicine.

BackgroundUnsafe abortion is a significant global health concern, contributing to high rates of maternal mortality, particularly in developing countries. Women in pastoralist communities face unique challenges to access comprehensive abortion care (CAC) due to sociocultural, economic, and geographical barriers. This study explores the lived experiences of women in their journey to utilize CAC services in pastoralist communities of the Oromia region, Ethiopia.MethodsA phenomenological study was conducted in pastoralist communities of the Oromia region, including the Borana and Guji zones. A total of nine women were interviewed in-depth for their lived experiences in their journey to access the CAC service. All interviews were audio-recorded, transcribed verbatim, translated to English, and coded using Open Code version 4.03 software. Both inductive and deductive thematic analysis were employed to analyze the data. The lived experiences of women in their journey to access CAC services were analyzed under three categories: ‘before’, ‘during’, and ‘after’ utilizing the CAC service.ResultsBefore receiving CAC service, women faced emotional instability, fear, decision-making difficulties, and barriers such as transportation challenges, stigma, and lack of family support. During the service, women reported mixed experiences with healthcare providers, with some receiving supportive care while others reported disrespect and a lack of privacy. Following receiving CAC service, experiences included both relief and satisfaction for some, while others expressed regret and guilt due to cultural and religious beliefs surrounding abortion.ConclusionThe study reveals the complex experiences of women in their journey of accessing CAC services in pastoralist communities, worsened by socio-cultural norms, economic hardship, and limited healthcare infrastructure. Addressing the barriers identified, particularly those related to stigma, provider attitudes, and healthcare accessibility, is essential for improving CAC service delivery in these underserved regions.

The escalating pressure on land and water resources necessitates precise tools for identifying suitable irrigation areas, especially in rain-fed agricultural systems vulnerable to climate variability. This study presents a comprehensive, basin-wide land suitability assessment for surface irrigation in Ethiopia’s Genale Dawa River Basin (GDRB), the nation’s third-largest basin with significant untapped potential. We employed a GIS-based Analytic Hierarchy Process (AHP) to integrate 11 biophysical and infrastructural criteria: soil depth, texture, pH, organic matter, cation exchange capacity, water-holding capacity, rockiness, slope, land use/land cover, and proximity to rivers and roads. Factor weights were derived via expert-driven pairwise comparisons (Consistency Ratio = 0.079), and a weighted overlay analysis generated a final suitability map classified into four classes: highly suitable (S1), moderately suitable (S2), marginally suitable (S3), and not suitable (N). The analysis reveals that only 8.32% of the GDRB is highly suitable (S1) for surface irrigation, while a vast 73.29% is moderately suitable (S2). The highly suitable zones are primarily located in the central plains, characterized by deep Vertisols, gentle slopes, and better access to water, whereas moderately suitable lands require targeted interventions to overcome constraints related to soil texture, rockiness, or infrastructure access. The findings advocate for a two-phase development strategy: immediate investment in highly suitable areas for rapid gains, coupled with a long-term program of land improvement and infrastructure development to unlock the potential of moderately suitable lands. This study provides a critical, evidence-based geospatial tool for policymakers to optimize land and water resources, enhance climate resilience, and achieve sustainable agricultural intensification in Ethiopia and similar data-scarce basins.

A hyperglycemic crisis represents a critical metabolic emergency with potentially fatal consequences and significant health complications. These crises account for approximately 1% of hospital admissions for individuals with diabetes. While hospital admission is essential for providing the intensive care and inpatient services necessary to sustain life, prolonged hospital stays are a major risk factor for increased morbidity and in-hospital mortality among patients admitted for hyperglycemic crises. Therefore, this study aimed to determine the prevalence of prolonged hospital stays and associated factors among hyperglycemic crisis patients admitted to public hospitals in Ethiopia between July and November 2022. This research utilized a hospital-based multicenter prospective observational study design over a period of five months. Patient data were gathered through interviews and chart reviews. The duration of hospital stay was assessed by following the patients from admission to discharge. The collected data were systematically cleaned, coded, and entered into Epi Data Manager version 4.6, subsequently exported to the Statistical Package for the Social Sciences (SPSS) version 25.0. Multiplestep-wisebackward logistic regression was employed to identify factors influencing the length of hospital stay. The strength of associations between dependent and independent variables was examined using a 95% confidence interval (CI), adjusted odds ratio (AOR) and a p-value of less than 0.05. This study included a total of 213 patients with hyperglycemic crisis. A mean age of participants was 40.37 ± 16.87 SD. About 58.2% (95% CI: 51.3-64.9%) of admitted patients were hospitalized for longer than a week, and the mean of hospital stay was 7.21 days. Age of older than 45 years [AOR:6.0,95%CI:2.27-15.93], female gender [AOR:2.76,95%CI:1.4-5.42], type 2 diabetes [AOR: 3.78,95%CI:1.36-10.48], and hypernatremia [AOR:4.31,95%CI: 1.1-16.98] were factors associated with prolonged hospital stays. The findings from this study underscore significant factors contributing to prolonged hospitalization in patients experiencing hyperglycemic crises. These group of patients need more emphasize and tailored management strategies to potentially mitigate the duration of hospitalizations and improve overall healthcare outcomes.

This study presents a higher-order uniformly convergent numerical method for a singularly perturbed nonlinear reaction–diffusion equation. The quasilinearization technique is used to transform a nonlinear term into a linear boundary value problem. The equivalent singularly perturbed reaction–diffusion differential equation is discretized using a central finite-difference method on the Shishkin and Bakhvalov meshes. The stability and parameter-uniform convergence are rigorously analyzed and established. Theoretically, the proposed method is of second order. The application of Richardson extrapolation demonstrates that the Bakhvalov mesh achieves a pure CM−4 rate, whereas the Shishkin mesh attains an ε-uniform convergence rate of CM−4ln4M. Three numerical examples have been solved to corroborate the theoretical findings.

In low and middle-income countries, access to clean and microbiologically safe drinking water remains a major challenge. Contaminated drinking water promotes the transmission of waterborne pathogens and the dissemination of antimicrobial-resistant bacteria, which is a global concern. Between May and July 2024, a cross-sectional study was conducted to determine the bacteriological quality of tap water, antimicrobial resistance patterns of E. coli, and associated risk factors in Gulelle and Yeka sub-cities of Addis Ababa, Ethiopia. A total of 341 household tap water samples were collected and processed for bacteriological quality assessment following standard microbiological procedures. Coliforms were isolated using membrane filtration, and aerobic colony counts (ACC) were determined via the pour plate method. Antimicrobial susceptibility testing was performed using the Kirby-Bauer method. Statistical analysis was performed using SPSS version 27, with logistic regression identifying significant associations (p < 0.05). Bacteriological contaminations of tap water were found to be 18.8% (n=64/341) for total coliforms, 14.7% (n=50/341) for thermotolerant coliforms, 8.8% (n=30/341) for E. coli, and 81.5% (n=278/341) for ACC. E. coli isolates showed the highest resistance to amoxicillin (100%), ampicillin (80%), and nitrofurantoin (70%) while exhibiting 100% susceptibility to ciprofloxacin and gentamicin. Additionally, the E. coli isolates demonstrated moderate sensitivity to Ceftriaxone (60.0%) and Cefepime (73.0%), respectively. Significant associations with contamination were identified, including damaged pipes (AOR: 9.9, 95% CI: 4.09–24.15), the proximity of latrines to water sources (AOR: 7.19, 95% CI: 2.93–17.64), and inadequate waste management practices (AOR: 2.5, 95% CI: 1.02–6.3). A high level of bacterial contamination, including E. coli, and antimicrobial resistance was observed in drinking tap water, highlighting the need for continuous sanitary inspections to prevent waterborne infections and the spread of antimicrobial resistance within the community.

Conventional soybean oil contains high levels of linoleic acid, which reduces oxidative stability and necessitates hydrogenation, leading to trans-fat formation. In this study, 40 Indian soybean genotypes were screened for fatty acid composition, and Gujarat Junagadh Soybean-3 (GJS-3) was selected for CRISPR/Cas9-mediated editing of the fatty acid desaturase-2 (FAD2) gene. Agrobacterium-mediated transformation produced 22 regenerated plants, of which 57.1% were PCR-positive for Cas9/sgRNA. Targeted single-nucleotide substitutions were confirmed by Sanger sequencing in three edited lines (T3, T7, and T15), corresponding to an editing efficiency of 13.63%. These lines exhibited a marked increase in oleic acid content (42-45%) compared with the wild type (22%) and a concomitant reduction in linoleic acid (30-32% vs. 54%), resulting in nearly a two-fold improvement in the oleic/linoleic acid ratio. PCR analysis confirmed the absence of Cas9 and U3 sequences, indicating transgene-free edited plants. This study provides the first evidence of CRISPR/Cas9-mediated FAD2 editing in an Indian soybean cultivar and demonstrates its effectiveness in improving oil quality, oxidative stability, and processing efficiency.

Relapses from substance usage continue to be a major global public health concern. Many individuals find it difficult to sustain long-term recovery, often cycling between phases of recovery and relapse, even with improvements in treatments. This study employs a center-based phenomenological design to examine the experiences of adults who relapse from substance dependence. For in-depth interviews, nine patients were chosen. An interpretative phenomenological analysis (IPA) was used to analyze the data. The results revealed that a complex interplay of personal and environmental factors influences relapse. Self-motivation, relocating, maintaining supportive social networks, improving mental resilience, making plans for future stability, and actively engaging in spiritual healing were all effective coping strategies for reducing relapse. To prevent recurrence and encourage long-term recovery, the findings emphasize the need for comprehensive intervention strategies that address social influences, environmental triggers, and emotional regulation.

Sustainable groundwater management in regions with limited data remains a significant challenge. Identifying potential groundwater zones by combining geological and hydrological information is crucial for the effective use and protection of this resource. In the Southwest Shewa Zone of Ethiopia, groundwater is the main source for household, agricultural, and industrial needs. However, limited hydrogeological data often leads to drilling ineffective wells, emphasizing the need for precise, data-driven mapping methods. This study uses a stacked ensemble machine learning approach to identify groundwater potential zones. The framework combines Adaptive Boosting, Random Forest, Histogram-Based Gradient Boosting, and Extreme Gradient Boosting using a meta-learner to improve prediction accuracy and minimize bias. The area was classified into five groundwater potential levels: very low, low, moderate, high, and very high. Results indicate that 55.8% of the area falls within high to very high groundwater potential zones, while 29.8% corresponds to very low to low potential. Model evaluation using recall, precision, F1-score, and Receiver Operating Characteristic (ROC) demonstrates strong predictive capability and reliable class discrimination across all groundwater potential zones.These findings demonstrate the strong predictive performance of the stacked ensemble learning model and provide a scientific basis for identifying groundwater prospecting zones and guiding well siting in the study area.