Mbarara University of Science and Technology Institutional Repository (MUST-IR)

MUST-IR preserves research output from the MUST Community

Communities in MUST-IR

Select a community to browse its collections.

Now showing 1 - 10 of 10

Central Administration 129
This community will contain collections from the general Administrative staff
Faculty of Applied Sciences and Technology 63
Faculty of Business and Management Sciences 117
Faculty of Computing and Informatics 148
Faculty of Interdisciplinary Studies 216
Faculty of Medicine 2388
Faculty of Science 567
Institute of Tropical Forest Conservation 127
Maternal Newborn and Child Health Institute 18
MUST Research Institute 3

Recent Submissions

Pathways to suicide attempts in rural Uganda: a qualitative study of survivors, family members, and healthcare providers, and community health workers
(International Journal of Mental Health, 2026-07-19) Lee,Yang Jae; Asasira,Mildred; Kazungu,Rauben; Aud,Ryan M.; Moffett,Elizabeth; Kang,Loukas; Ager,Katherine; Mugamba, Tribius; Rosenheck,Robert; Tsai,Alexander C.
Introduction: Suicide is a critical public health issue worldwide. Risk factors are exacerbated by cultural stigma, socioeconomic hardship, and limited mental healthcare access. However, few studies have examined the lived experiences of suicide attempt survivors in Africa to probe the underlying drivers. Objectives: To explore the lived experiences and perceived drivers of suicide attempts in rural Uganda among individuals with lived experience, their families, healthcare providers, and community health workers. Methods: We conducted in-depth qualitative interviews in Buyende District, Uganda from June to August 2023 with a purposive sample of 54 participants aged 23 to 70: people who had survived a suicide attempt (n = 18) and their family mem¬bers (n = 17), healthcare providers (n = 10), and community health workers (n = 9). Interviews were recorded, transcribed, translated, and analyzed using a thematic framework method. Results: Four main themes emerged as critical factors leading to suicide attempts: (1) familial and spousal conflict, involving relationship breakdown, resource denial, and social isolation; (2) financial stress, including chronic poverty, acute debts, and per¬ceived socioeconomic failure; (3) mental illness and substance use, notably depression, psychotic symptoms, and alcohol use complicating existing psychological distress; and (4) religious and spiritual influences on the attribution of suicidal thoughts. Conclusion: In this qualitative study from rural Uganda, suicide attempts across were perceived to result from intertwined inter-personal, economic, mental health, and spiritual factors. Participants rarely framed suicide as the result of a single cause, but rather as a response to cumulative and unresolved distress. Moments of missed support suggest opportunities for targeted intervention.
Enhancing Tuberculosis Care in Southwestern Uganda: Facilitators and Barriers to Utilizing Mobile Health Technologies
(Global Implementation Research and Applications, 2022-08-30) Tumuhimbise,Wilson; Atwine,Daniel; Kaggwa,Fred; Musiimenta,Angella
Although mobile health technologies are low-cost approaches that can enhance tuberculosis care, there is a dearth of research about the facilitators and barriers regarding their utilization in enhancing the public–private mix. The aim of this study is to explore the facilitators and barriers to utilizing mobile health technologies in fostering public–private mix for tuberculosis care in Southwestern Uganda, this was a qualitative study that involved in-depth interviews with 13 key informants (private healthcare workers) purposively selected between June and July 2020 due to their active involvement in tuberculosis care from four private hospitals in Mbarara city. An inductive content analysis approach based on the constructs of the consolidated framework for implementation research was used in this study. The main facilitators were related to design quality and packaging, cosmopolitanism, relative advantage, networks and communication, and engaging constructs and these include mapping and linking patients from private hospitals to the referral units; supporting patient care and medication adherence; reducing stigma among patients; enhancing communication between facilities; improving the knowledge and beliefs of healthcare workers and patients; as well as notifying and reporting tuberculosis cases to the Ugandan Ministry of Health. Poor internet connections and lack of knowledge to use mobile health interventions were the barriers that emerged from networks and communications; and knowledge and benefits about the intervention constructs. Mobile health could enhance tuberculosis care by linking the referred patients from private hospitals to public hospitals. Future research should focus on the effectiveness of mobile health technologies in enhancing the public–private mix in tuberculosis care in low-resource settings.
From centralised policy development to uneven implementation: a qualitative study of adolescent reproductive health policy processes at the national level and in Southwestern Uganda
(BMC Health Services Research, 2026-08-29) Asiimwe,Susan; Nyakato,Viola Nilah; Michielsen,Kristien
Background: Adolescent pregnancy remains a significant public health and health systems concern in Uganda and across sub-Saharan Africa. Although Uganda has developed several adolescent reproductive health policies and strategies, implementation remains uneven. Less is known about how policy development and implementation processes, financing arrangements, institutional structures, and frontline discretion shape translation of adolescent reproductive health policy into practice. This study examined how adolescent reproductive health policies in Uganda are developed and implemented, and how policy actors and frontline implementers cope with related constraints. Methods: A qualitative study was conducted between 14th July and 30th October 2025 at the national level and in Kabale and Mbarara study areas in southwestern Uganda. Twenty-eight key informants were purposively selected, including eight policymakers at national and subnational levels and twenty frontline implementers involved in adolescent reproductive health service delivery. Semi-structured interviews were recorded, transcribed, and analysed using thematic analysis combining inductive and deductive approaches, guided by Braun and Clarke’s six-phase process. Themes were developed inductively and subsequently mapped onto Walt and Gilson’s Health Policy Triangle domains, while Street-Level Bureaucracy theory informed interpretation of frontline discretion and coping practices Results: ARH policy development followed a structured, ministry-led process shaped by centralised decision-making and donor-dependent financing. Although institutional frameworks and development partner support facilitated policy formulation, limited subnational influence, delayed policy revision, and dependence on external funding constrained responsiveness and sustainability. At the implementation level, weak dissemination, resource shortages, limited supervision, and ambiguity around adolescents’ consent expanded frontline discretion, contributing to variable service delivery. Policymakers and implementers adopted coping strategies, including service integration and resource improvisation, phased implementation, and community engagement, which sustained services but also risked normalising systemic gaps. Conclusions: Uneven implementation of adolescent reproductive health policy in Uganda does not simply reflect operational failure at the facility level. It is shaped by the interaction of financing arrangements, centralised governance, policy ambiguity and frontline discretion within a resource-constrained health system. Strengthening implementation requires stable domestic financing, clearer operational guidance, stronger subnational engagement, and institutional mechanisms that connect policy development with frontline realities.
A Theory of Planned Behaviour Analysis of Barriers and Facilitators of Immediate Postpartum Contraception Uptake in Southwestern Uganda
(Open Access Journal of Contraception, 2026-08-27) Ainomugisha,Brenda; Owaraganise,Asiphas; Tumuhimibise,Wilson; Naiga,Patience; Asiimwe,Josephine; Kamugisha,Arnold; Bwengye,Martin; Alinda,Nedeen; Kayondo,Musa; Ngonzi,Joseph
Background: Closely spaced pregnancies increase the risks of maternal and perinatal morbidity. Initiation of contraception in the immediate postpartum period offers a critical opportunity to address this unmet need during a window when women are in contact with health services. We explored barriers and facilitators of immediate postpartum contraceptive uptake at a tertiary hospital in southwestern Uganda, using the Theory of Planned Behavior. Methods: We conducted a qualitative descriptive study between August and December 2025 at Mbarara Regional Referral Hospital. Using purposive sampling, we conducted 7 focus group discussions among postpartum women (24), spouses (19), midwives (20), and doctors (12) involved in maternity care using semi-structured interview guides grounded in the TPB. Audio-recorded interviews were transcribed verbatim, translated where necessary, and analysed using thematic analysis. Themes were deductively mapped onto the TPB domains of attitudes, subjective norms, and perceived behavioural control until no new themes emerged across successive focused group discussions. Results: Participants generally viewed postpartum family planning positively and recognized its role in child spacing, maternal recovery, and family wellbeing. However, uptake was hindered by fears of side effects, misconceptions about infertility and contraceptive failure, concerns regarding immediate postpartum initiation, and limited awareness of methods suitable for use. While partner support facilitated uptake, opposition and community myths discouraged use. Perceived behavioural control was constrained by inadequate counselling, limited provider confidence, poor integration of family planning services within maternity care, and inconsistent availability of contraceptive methods. Conclusion: Immediate postpartum contraceptive uptake is influenced by interconnected individual, social, and health-system factors. It requires multi-level interventions: integrated contraceptive counselling across the antenatal–postnatal continuum, structured male partner engagement, correction of community misconceptions, training of health care providers, and onsite availability of methods within labour wards and theatres
Modeling the impact of biomedical prevention on HIV/AIDS transmission dynamics
(Discover Public Health, 2026-08-26) Byamukama,Michael; Kizito, Mohammed; Atwebembeire,Jeninah; Mugisha,Arthur; Mwesigwa,Ronald
The HIV/AIDS epidemic has significantly reshaped global public health priorities, prompting unprecedented investments in research, treatment programs, and prevention strategies. Despite these efforts, the disease continues to place substantial pressure on healthcare systems, particularly in resource limited settings. In this study, a deterministic dynamical system model is developed to assess the impact of key biomedical interventions on the transmission dynamics of HIV/AIDS. The well-posedness of the proposed model is established using the theory of positivity and boundedness of solutions. Analytical results show that the HIV/AIDS-free equilibrium is both locally and globally asymptotically stable whenever the effective reproduction number is less than unity (R∗H < 1). Furthermore, the analysis reveals that in the presence of antiretroviral therapy (ART) defaulting, the model may exhibit a backward bifurcation. Such a result provides a plausible explanation for the sustained transmission of HIV at the national level despite achieving the effective reproduction number less than unity. In contrast, under conditions of full adherence (100%) to ART, the model exhibits a forward bifurcation, whereby the HIV/AIDS-free equilibrium is stable whenever R∗H < 1 and loses stability as R∗H exceeds unity. Sensitivity analysis of the effective reproduction number highlights the importance of increased ART uptake across all stages of infection, together with expanded coverage of pre-exposure prophylaxis (PrEP) and post-exposure prophylaxis (PEP), as critical strategies for reducing HIV transmission. These findings are further supported by numerical simulations. A key outcome of the study indicates that ART defaulting has the greatest impact in increasing the infectious population. While promoting ART uptake remains essential, strengthening community education on sustained adherence to ART is even more critical for achieving the World Health Organization (WHO) 90-90- 90 targets.
Optimizing Chest X-ray Referral Using Clinical Metadata: Pulmonary Disease Patterns and Diagnostic Access in Southwestern Uganda
(Cureus, 2026-08-23) Obungoloch,Johnes; Tumusiime,Julius; Buri,Gershom; Mukama,Martin; Nkwanga,Jacob; Mbusa,Chrispus; Kaggwa,Fred; Murungi,Shallot N.; Wasswa,William
Background: Chest X-ray (CXR) imaging is important for diagnosing pulmonary and cardiothoracic conditions, but timely access remains limited in many low- and middle-income countries. This study characterized radiographic abnormalities, examined associated clinical factors, evaluated exploratory clinical metadata-based prediction models, and assessed barriers to CXR utilization in southwestern Uganda. Methods: This facility-based observational pilot study included 422 adults undergoing chest radiography for suspected pulmonary or cardiothoracic disease at two healthcare facilities. Prospectively collected demographic, clinical, environmental, and healthcare-access data were linked to routine radiographer reports. Radiographer-reported pneumonia, pleural effusion, and cardiomegaly were summarized descriptively. Associated factors were examined using multivariable logistic regression with complete-case analysis. Exploratory model discrimination was assessed using receiver operating characteristic analysis, while post hoc Stage 1 simulations examined trade-offs between imaging-referral volume and case detection. Results: Complete radiographic-outcome classifications were available for 403 participants. Pneumonia was the most frequently reported abnormality (18.9%), followed by pleural effusion (9.7%) and cardiomegaly (5.5%). Increasing age was independently associated with pneumonia (adjusted odds ratio (aOR) 1.318 per 10-year increase; 95% confidence interval (CI) 1.149-1.523) and cardiomegaly (aOR 1.756 per 10-year increase; 95% CI 1.383-2.240), but not pleural effusion. Higher body mass index was associated with lower odds of pleural effusion and slightly higher odds of cardiomegaly. The exploratory models showed apparent in-sample discrimination for cardiomegaly (AUC 0.881; 95% CI 0.807-0.963), pneumonia (AUC 0.787; 95% CI 0.728-0.849), pleural effusion (AUC 0.699; 95% CI 0.601-0.788), and any reported abnormality (AUC 0.790; 95% CI 0.742-0.836). Effective access declined after accounting for personnel availability, affordability, and willingness to undergo imaging; only 17% of participants reported being both willing and able to complete the diagnostic pathway. The Stage 1 simulations illustrated trade-offs between referral volume and case detection under selected thresholds. Conclusions: Routinely obtainable clinical metadata may contain useful information for preliminary preimaging risk stratification. However, the reported AUCs represent apparent discrimination within the model-development sample, and the referral strategies were neither prospectively implemented nor clinically validated. The findings are exploratory and hypothesis-generating. Independent radiologist verification, validation of the report-classification procedure, model calibration, internal and external validation, and prospective workflow evaluation are required before the proposed approach can support patient-level referral decisions.
Evaluating the Generalisability of Convolutional Neural Networks for Diabetic Retinopathy Detection in Latin America and Sub-Saharan Africa
(MDPI, 2026-06-03) Mwavu,Rogers; Kaggwa,Fred; Arunga,Simon; Wasswa, William
Diabetic retinopathy is a leading cause of vision loss worldwide, particularly impacting individuals in low- and middle-income countries with limited healthcare access. Early detection through automated screening systems is essential for improving outcomes, as timely intervention can prevent severe vision impairment. However, most of the available AI models have not been evaluated in low-resource settings. Hence, this study presents an evaluation of the efficacy of advanced deep learning architectures for detecting rDR across diverse population datasets. A dual-phase validation approach was employed to assess model performance. Internal validation utilised the BrSET dataset to establish baseline performance metrics, while external validation was conducted on the MoDRIA dataset, which encompasses various conditions and demographics, to evaluate model robustness. Key performance metrics, including accuracy, specificity, sensitivity, F1-score, and calibration scores, were systematically recorded and analysed. Internal validation revealed high accuracy across all models, EfficientNetB0 achieved the highest classification accuracy (0.9561; 95% CI 0.9490–0.9630), EfficientNetB3 demonstrated superior overall discriminative performance, achieving the highest AUROC (0.9892; 95% CI 0.9841–0.9934) highest sensitivity (0.9573), and lowest Brier score (0.0168). Meanwhile, DenseNet ex hibited the most balanced clinical screening performance, achieving the highest F1-score (0.7259; 95% CI 0.6797–0.7669) and Youden Index (0.2381), indicating improved balance between sensitivity and specificity. In contrast, external validation revealed substantial deterioration in model performance across all architectures, highlighting major limitations in cross-population generalisability. Although EfficientNetB0 achieved the highest external accuracy (0.8821; 95% CI 0.8746–0.8898), AUROC values declined markedly across models (0.5140–0.6104), accompanied by poor sensitivity, reduced F1-scores, and substantial calibra tion instability. EfficientNetB3 achieved the highest external sensitivity (0.5939), whereas calibration analyses demonstrated unreliable probability estimation under domain-shift conditions. These findings suggest that AI models trained on geographically homogeneous retinal imaging datasets may not generalise reliably across underrepresented populations. Population differences and imaging variability substantially affected external model per formance, highlighting the need for diverse datasets, rigorous external validation, and adaptive recalibration before clinical deployment of AI-driven DR screening systems.
“I Did What I Could Manage”: Caregiver Burden and Rational Triage of Stroke Discharge Instructions in Rural Uganda – A Qualitative Study
(Journal of Multidisciplinary Healthcare, 2026-07-22) Kakongi,Nathan; Owokuhaisa, Judith; Asiimwe,Derrick; Nuwahereza,Amon
Background: In rural Uganda, informal caregivers provide most post-discharge stroke care, often without formal training, written guidance, or community rehabilitation support. This study explored caregiver burden and factors shaping implementation of stroke discharge instructions in rural southwestern Uganda. Methods: We conducted a qualitative study at Mbarara and Kabale Regional Referral Hospitals. Eleven informal caregivers of stroke survivors participated in semi-structured in-depth interviews. Interviews were audio-recorded, transcribed, translated into English, and analysed in Atlas.ti 9 using thematic analysis with a deductive codebook refined inductively during analysis. Results: Four themes were identified: multidimensional caregiving burden; variable discharge instruction content and delivery; barriers to implementing instructions; and facilitators that sustained care. Burden was physical, financial, psychological, and social, often concentrated on one family member. Discharge instructions commonly focused on exercises, medications, diet, and follow-up, but were frequently verbal and lacked clear medication resupply or review plans. A central finding was caregivers’ rational “triage logic”: when resources were insufficient, they prioritized zero-cost instructions such as exercises and delayed or abandoned costly instructions such as medicines, transport, and follow-up visits. Facilitators included love and duty, family support, retained discharge knowledge, resource mobilisation, and occasional home-based health worker visits. Conclusion: Caregivers of stroke survivors in rural southwestern Uganda experience substantial multidimensional burden and implement discharge instructions within severe structural constraints. Written discharge protocols, caregiver screening, clearer follow up pathways, and task-shared community rehabilitation may improve post-stroke continuity of care in similar low
Prophylactic antiseizure medication administration practices in neonates with hypoxic-ischaemic encephalopathy: a scoping review protocol
(BMJ open, 2026) Nalwanga,Lillian Edith; Ayesiga,Innocent; Ediamu,Tom Didimus
Introduction Hypoxic-ischaemic encephalopathy (HIE) is a leading cause of neonatal mortality and long-term neurodevelopmental impairment worldwide. Neonatal seizures are common in HIE and are associated with adverse outcomes. While antiseizure medications are routinely used for treatment, their prophylactic use in the absence of clinically evident seizures remains controversial, with limited consensus and variable clinical practices. Therefore, this review will map the existing evidence concerning the utilisation of antiseizure medication for prophylaxis among neonates with HIE. Methods and analysis The review will follow the Joanna Briggs Institute scoping review methodology and be reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews (PRISMA-ScR) guidelines. A comprehensive search will be conducted across various scientific databases, including PubMed, Embase, Scopus and Web of Science, guided by the Population, Intervention, Concept and Context framework, starting between October and November 2026. The review will consider studies on the prophylactic use of antiseizure medication among neonates with HIE. The collected search results will be organised and uploaded to Covidence, where duplicates will be removed, and titles and abstracts will be screened. Search and screening processes have not yet started. Ethics and dissemination This will be a secondary study using already published information; therefore, ethical approval was not sought. All studies included in this review will be cited accordingly, as attributed to the publication licence. The data gathered will be summarised in a scoping review and submitted for publication in an international peer-reviewed journal. Datasets shall also be made publicly available in a data repository, such as the Open Science Framework
PEDCan: acceptability and feasibility of a mobile application to support primary health-care workers in recognising clinical features of childhood cancer in Uganda
(ecancer, 2026-06-26) Atwiine,Barnabas; Izath,Nura; Ssemitego,James; Nsiimenta,Naome; Birungi,Wilson; Kusaasira,Anitah; Nambogo,Nuriat; Mukama,Martin; Data, Santorino
Introduction: Early diagnosis of childhood cancer depends on a high index of suspicion and timely referral by primary healthcare workers (PHCWs) to tertiary centers. This study aimed to assess the acceptability and feasibility of a mobile application among PHCWs in recognising cancer-related clinical features in children. Methods: We developed a mobile application called (PEDCan) to support PHCWs in rec ognising cancer-related clinical features in children, initiating remote consultations with specialists and referring suspected cases to cancer centers when necessary. In May 2021, we conducted a cross-sectional descriptive study among consenting PHCWs in south western Uganda to collect data on their knowledge of childhood cancer, smartphone ownership and internet usage. After training participants on the use of the application through mock clinical vignettes, we assessed their ease of use and readiness to adopt the application. Descriptive statistics were used to calculate the proportions of participants able and willing to use the application. Ethical approval was obtained from the Research and Ethics Committee of Mbarara University of Science and Technology. Results: A total of 44 PHCWs participated in the study; 25 (56.8%) were nurses or mid wives, 8 (18.2%) were clinical officers, 3 (6.8%) were medical doctors and 8 (18.2%) belonged to other professional cadres. All participants had previously heard about child hood cancer, and 28 (63.6%) had previously encountered a child suspected to have can cer. Thirty-eight (86.4%) participants owned a smartphone and used the internet, while 12 (31.6%) reported having constant internet access. Thirty-eight (86.3%) participants found the PEDCan application very easy or easy to use, and all participants expressed willingness to use it. Conclusion: PHCWs in the participating health facilities were able and willing to use the PEDCan mobile application to recognise childhood cancer clinical features and initiate remote expert consultations using mock vignettes. A pilot study involving real patients would further demonstrate the application’s effectiveness in real-world clinical settings.
Mapping the use of interactive voice response (IVR) in healthcare in Africa: a scoping review protocol
(BMJ open, 2026) Kihumuro,Raymond Bernard; Ochen,Eric; Atimango,Lorna; Kizito,Samuel; Haberer,Jessica; Maling,Samuel; Lydston,Melis; Sundararajan,Radhika; Bebell,Lisa M
ABSTRACT Introduction Africa faces a substantial health burden characterised by a dual burden of communicable and non-communicable diseases compounded by persistent health system and infrastructure constraints. Digital health interventions offer scalable opportunities to strengthen healthcare delivery. Interactive voice response (IVR), an automated telephone-based technology, shows promise across African health settings; however, existing evidence is dispersed across isolated studies and demonstrates variability in outcomes and implementation approaches. This scoping review maps existing evidence on the use of IVR in healthcare in Africa, focusing on applications, outcomes and implementation experiences. Methods and analysis Eligible studies will include all designs that evaluate IVR-based health interventions in African healthcare settings, while non-voice digital interventions and non-empirical publications will be excluded. We will conduct a scoping review guided by the Arksey and O’Malley framework and the Joanna Briggs Institute and report findings using Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) guidelines. We will systematically search bibliographic databases and grey literature from 2000 to the end of 2025. Two independent reviewers will screen studies, extract data and chart evidence on IVR characteristics, implementation strategies, barriers and facilitators and outcomes. A third reviewer will resolve disagreements. We will synthesise findings descriptively and thematically, presenting results in tables, visual maps and an equity-focused analysis guided by PROGRESS-Plus. Ethics and dissemination Ethical approval is not required because this scoping review synthesises previously published, publicly available literature and does not involve human participants or primary data collection. We will share the findings through a peer reviewed publication and conference presentations, and we will discuss the results with digital health stakeholders, clinicians and policymakers to inform future IVR work in Africa
Expression of p53 and p63 proteins and their correlation with histological grades among gastric carcinoma specimens at Mbarara Regional Referral Hospital
(African Journal of Laboratory Medicine, 2026-07-17) Ninsiima,Lydia Jolly; Birungi,Abraham; Kasadha,Richard; Mugisa,Michael Junior; Tibenderana,Leuben; Nuwashaba,Nicholas; Nabaasa,Saphurah; Wasswa,Hassan; Amadile,Lawrence; Ssedyabane,Frank
Background: Gastric carcinoma remains one of the most common and deadly cancers worldwide, especially among older men. In sub-Saharan Africa and Uganda in particular, the diagnostic biomarkers of gastric carcinomas are not well documented; also, limited studies have been conducted in Uganda regarding the prognostic and diagnostic value of p53 and p63 immunohistochemically expression in gastric carcinomas. Objective: We aimed to determine the immunohistochemically expression of p53 and p63 proteins and their correlation with different histological grades among archived gastric carcinoma specimens at Mbarara Regional Referral Hospital. Methods: This was a laboratory-based cross-sectional study. We retrieved 166 archived formalin-fixed paraffin-embedded gastric carcinoma tissue specimens from the histopathology laboratory. They were stained with Harris hematoxylin and eosin stain and subsequently stained with p63 and p53 monoclonal antibody immunohistochemistry. Results: Of the 166 gastric carcinoma specimens retrieved, almost a third of specimens were poorly differentiated (53/166; 32%) although the majority were moderately differentiated (68/166; 41%). P53 expression was 20/45 (44.4%), 27/68 (39.7) and 21/53 (39.6) in well differentiated, moderately differentiated and poorly differentiated gastric carcinoma. Statistically significant positive correlation was observed between p63 and weak immunohistochemically expression (Spearman rho 0.155, p-value 0.047). Conclusion: P53 expression was more than p63 expression in different histological grades. There was no significant correlation of p53 with different histological grades but significant correlation was observed between p63 and moderately differentiated gastric carcinoma specimens. We recommend a prospective study using advanced molecular tests on fresh gastric carcinoma specimens to assess the prognostic usefulness of p53 and p63 biomarkers. What this study adds: This study provides the first Ugandan data on p53 and p63 expression in gastric carcinoma, highlighting p63’s significant correlation with moderately differentiated tumors
Explainable transfer learning-based approach for Mango fruit damage detection and cause identification
(2026 6th International Symposium on Computer Technology and Information Science (ISCTIS), 2026) Safari,Yonasi; Nakibuule,Rose; Joyce Nabende; Nakasi,Rose
This study presents a deep learning approach to detect mango fruit damage and identify its underlying causes using one-stage and two-stage detectors. Unlike existing methods that focus mainly on detection, this approach integrates XAI techniques to provide interpretable, cause-level insights, and is evaluated on a dataset annotated with anthracnose, bacterial black spot, fruit fly, and other damage classes. RetinaNet achieved the best overall performance, attaining a precision of 0.77, recall of 0.75, and F1-score of 0.76. Per-class accuracy exceeded 0.80 in all categories, except for fruitfly, which remained more challenging to detect. XAI techniques that include Grad-CAM, Grad-CAM++, LIME, SHAP, and LRP were further used to highlight vital regions influencing model decisions. A quantitative evaluation of XAI performance was further conducted using faithfulness, robustness, complexity, localization, and randomization. LRP and SHAP results attained are indicative of more faithful and well-localized explanations, while Grad-CAM based methods exhibited limitations in localization, with LIME producing highly complex explanations that result in reduced interpretability
Development of a Simulation-Based Educational Intervention to Improve Nursing Competencies for Stroke Care in Southwestern Uganda
(Advances in Medical Education and Practice, 2026-07-16) Najjuma,Josephine N; Bajunirwe,Francis; Ssedyabane,Frank; Kaddumukasa,Mark; Mbalinda,Scovia N; Moore,Shirley; Tebulwa,Joanita B; Nankunda,Shifra; Atukunda,Esther C
Stroke is one of the leading causes of disability worldwide, and its prevalence is increasing in low- and middle-income countries (LMIC) like Uganda, presenting nursing care challenges. Simulation-based learning has been used successfully to train nurses in general care in LMIC but has not been applied to training for stroke care. In this study, we aimed to develop a simulation based stroke care education intervention for nurses involved in care for patients with stroke at a large hospital in southwestern Uganda. Methods: Between February 2024 and March 2025, we used the iterative design to develop an educational intervention for stroke care. We engaged stroke care content experts through in-depth qualitative interviews (n=14), generated content, and developed the first version of the intervention. Transcripts were derived, reviewed and coded to generate themes and categories using inductive content analytic approach. We then tested, analyzed and refined the intervention through three repetitive exposures of the intervention to final year Bachelor of Nursing students (n=20), with improvements after every exposure. That were grouped in teams of 5 students. Results: Qualitative data from content experts revealed three thematic areas of patient care including admission of a stroke patient, care for a stroke patient during their hospital stay and discharge of a stroke patient as key content for consideration during intervention development. This content was integrated into the “Uganda Simulation-Based Nursing Stroke Care intervention”. Post-exposure qualitative interviews with final year Bachelor of Nursing students revealed great satisfaction with the intervention, improved preparedness and ease to assess stroke patients. There were also high teamwork scores and improved clinical skills. To maintain efficiency with stroke care competencies, they preferred at least three exposures to the intervention per year. Conclusion: With diverse stakeholders, we used an iterative approach to develop a context specific, simulation-based stroke educational intervention. Future work will examine the feasibility, acceptability, and effectiveness of this intervention among nurses working with stroke patients and its effect on the patient outcomes
Cloud computing-based collaborations on health research: A systematic review of applications, barriers to adoption, and migration strategies
(Digital Medicine, 2026) Mugonza,Robert; Ssembatya,Richard; Basaza-Ejiri,Annabella
Background: Modern biomedical research is undergoing a fundamental transformation driven by an exponential increase in data complexity and the integration of artificial intelligence and cloud computing into biomedical and public health infrastructure. While cloud computing offers a pivotal solution to the limitations of on-premise infrastructure, the transition is hindered by a complex set of domain-specific challenges. This systematic review maps the current landscape of cloud adoption, identifies systemic barriers to wider usage, and outlines practical requirements for successful cloud migration. Methods: We conducted a comprehensive systematic search of articles published between 2010 and early 2026 across PubMed and Google Scholar. To capture both clinical and technical literature, the search was expanded to include major computer science and engineering databases, namely IEEE Xplore and Web of Science. We identified 68 primary studies evaluating 50 unique applications and subjected these to a formal assessment using a 5-point informatics quality appraisal (IQA) scale to determine the maturity of evidence. Results: Our analysis revealed a highly mature ecosystem, with 96% of identified applications reaching production-ready or validated implementation stages (IQA scores = 3-5). These applications were categorized into three functional types: task specific tools, integrated platforms, and large-scale archives. Despite a projected market growth exceeding 17%, however, a significant readiness gap persists. Innovation remains concentrated in North America and is characterized by high data-egress costs, regulatory misalignments, and friction associated with technical interoperability. In contrast, initiatives such as the District Health Information System-2 (DHIS-2) demonstrate the potential for cloud applications to drive global health equity in resource limited settings. Conclusion: The cloud has successfully moved from being a conceptual framework to serving as the backbone of modern health research. Nevertheless, a considerable trust deficit coupled with persistent structural barriers related to costs, data sovereignty, and regulatory complexity continue to impede global collaboration. To fully realize the promise of democratizing research, future development must transition from performance-centric models to federated, context-aware architectures that prioritize accessibility, inclusivity and secure data sharing. From a policy and practice perspective, funding agencies, governments, and international research consortia should support the adoption of such architectures while promoting harmonized cross-border data governance frameworks and measures to reduce prohibitive cloud data-egress costs. These efforts are essential for enabling equitable participation in global health research ecosystems, particularly among institutions in low- and middle-income countries.
Barriers and facilitators to linkage to HIV care after hospital discharge among adults living with advanced HIV disease in Uganda: a COM B–guided qualitative study
(HIV Research & Clinical Practice, 2026-07-23) Katuramu,Richard; Semitala,Fred C.; Neumbe,Irene Mildred; Nekaka,Rebecca; Nangendo,Joan; Namuwenge,Proscovia M.; Katahoire,Anne R.; Byakika-Kibwika,Pauline; Kamya,Moses R.; Matovu,Joseph K. B.
Background: Hospital discharge represents a high-risk transition period for people living with HIV, particularly those with advanced HIV disease. In low-resource settings, substantial gaps persist in linking these individuals to outpatient HIV care following hospitalisation. This study explored barriers and facilitators to post-discharge linkage to care from the perspectives of patients and healthcare workers (HCWs) in Eastern Uganda. Methods: We conducted a descriptive qualitative study at Mbale Regional Referral Hospital between May and October 2025. Guided by the capability, opportunity, motivation, and behaviour (COM-B) model, we conducted 20 in-depth interviews with adults with advanced HIV disease who had recently been discharged from the medical ward (10 who successfully linked to care and 10 who did not), alongside 10 key informant interviews with HCWs involved in their care. Data were analysed using a thematic framework approach aligned with the COM-B constructs. Results: Patient-reported barriers included physical weakness (physical capability), limited knowledge of the linkage process (psychological capability), high transport costs and long travel distances (physical opportunity), and stigma or non-disclosure of HIV status (social opportunity). HCWs identified the absence of standardised linkage protocols, post-discharge follow-up mechanisms, and incentive structures as key systemic barriers (physical opportunity). Facilitators reported by patients included initiation or resumption of antiretroviral therapy (ART) before discharge (physical opportunity), support from peer supporters and disclosure to caregivers (social opportunity), and fear of clinical deterioration (reflective motivation). HCWs identified the existing TB–HIV linkage system (reflective motivation) and the presence of ward-based linkage counsellors (physical opportunity) as enabling factors. Conclusion: People living with advanced HIV disease face multiple, interrelated barriers to linkage with HIV care following hospital discharge, including physical debility, limited knowledge of the linkage process, financial and geographic constraints, and stigma. Conversely, pre-discharge ART initiation, peer support, and disclosure facilitate successful linkage. Strengthening post-discharge care will require standardised linkage protocols, structured follow-up mechanisms, and expanded use of peer supporters within health systems.
Build a neuro-fuzzy continuous protocol representation for ventilator adjustments in acute respiratory distress syndrome patients using fuzzy neural network.
(Scientific Report, 2026-08-26) Wanzala,Jimmy Nabende; Atim, Michael Robson; Awal,Moses; Mugabe,Robert
—Acute respiratory distress syndrome (ARDS) is on the increase due to many causes such as: corona virus disease 2019 (COVID-19) which worsens the arterial hypoxemia, smoke inhalation, and injuries that cause fluids to collect in the air sacs of the lungs. Standard clinical criteria have been developed to define ARDS severity, while titration protocols guide ventilator settings. The Berlin definition protocol classifies ARDS diagnostic severity based on the P aO2/F iO2 ratio, whereas positive end expiratory pressure (PEEP) titration is guided by the ARDS Network (ARDSNet) lower and higher PEEP scales. The challenges that however come with physicians manually implementing discrete lookup tables include step-discontinuities at threshold boundaries, performance accuracy, and timely intervention. With the increasing number of patients in the intensive care unit (ICU), physicians may be overwhelmed with the workload, that they may not efficiently control the parameters on the mechanical ventilation. Any slight change or delay in setting the parameters may lead to wrong outcomes. Therefore, the aim of this study was to apply artificial intelligence in guiding physicians during the process of setting ventilator parameters. Fuzzy neural network (FNN) was used in training and conversion of the non-fuzzified ARDS Net PEEP titration tables into a fuzzified continuous inference surface. In this computational study, model performance was evaluated using 5-fold cross validation and tested on unseen intermediate F iO2 set points (F iO2 ∈ {0.35, 0.45, 0.55, 0.65, 0.75, 0.85, 0.95}), achieving a validation Root Mean Squared Error (RMSE) of 0.018 cm H2O, Mean Absolute Error (MAE) of 0.012 cm H2O, and R 2 = 0.9998. The results show that the output of the fuzzified ARDSNet PEEP model is very accurate in comparison with non-fuzzified tables while eliminating abrupt step changes between F iO2 increments. Comparative analysis demonstrates that ANFIS provides smooth C 1 -continuous parameter transitions without the boundary slope discontinuities of linear interpolation or the overshoot oscillations of cubic splines. The potential for reduced clinician workload and error reduction represents a hypothesis for future prospective clinical trials
Wetland-use and wetland-cover change and ecological integrity of wetlands around Kibale National Park, Kamwenge District, Western Uganda
(Journal of Freshwater Ecology, 2026-08-14) Ayebazibwe,Francis; Tumusiime,Julius; Yatuha,Jane; Nyakoojo,Clement; Tinka,John; Ainebyona,Peter
Wetlands around Kibale National Park support biodiversity, regulate water quality and sustain livelihoods but face increasing anthropogenic pressures. This study assessed wetland-use and cover change (between 2016 and 2025), macrozoobenthos diversity and physico-chemical water quality in Bigodi, Kahunge and Businge wetland systems within a 5 km buffer of Kibale National Park, Kamwenge District, Uganda. Sentinel imagery was classified to determine cover change. Sampling sites were classified as high quality, medium quality and low quality using the wetland rapid assessment procedure (WRAP). These categories represented a disturbance gradient based on ecological and anthropogenic characteristics of the wetland sites. Macrozoobenthos were sampled in dry ( January 2025) and wet (April 2025) seasons across all quality sites a scoop net, identified to family level and analyzed for diversity indices and average score per Taxon (ASPT). Physico-chemical parameters were measured in situ and in a laboratory. Wetland-use and wetland-cover (WUWC) change varied: Bigodi wetland system gained 1.17% wetland, Kahunge wetland system lost 3.58% and Businge wetland system lost 8.47% with +7.64% grassland increase. Macrozoobenthos abundance was highest in Bigodi (1189 individuals, 22 families), moderate in Kahunge (538, 21 families) and lowest in Businge (216, 21 families). Seasonal effects were significant in Kahunge (diversity p = 0.002) and Businge (diversity p = 0.033, Average Score Per Taxon p = 0.048), while ASPT differences across disturbance gradients were significant only in Businge (p = 0.004). Physico-chemical parameters generally met National Environment Management Authority standards, with minor nitrite exceedance in isolated low-quality sites. A principal component analysis showed system- and season-specific water chemistry: Businge dry-season linked to higher nitrates (mean = 0.413 ± 0.393 mg/L) and hardness (mean = 175 ± 144.3); Bigodi wet-season to higher Total dissolved solids (mean = 142.5 ± 80.6 ppm) and nitrates (mean = 0.675 ± 0.496 mg/L). Despite acceptable water quality, significant wetland loss, especially in Businge, threatens ecological integrity, warranting targeted habitat protection and stricter wetland-use controls.
Maternal inflammation disrupts transplacental antibody transfer in pregnant individuals with and without HIV
(Frontiers in Immunology, 2026-08-28) Bernshtein,Biana; Butler,Audrey; Roberts,Drucilla J.; Atwine,Raymond; Adong,Julian; Alter,Galit; Ngonzi,Joseph; Correia,Katharine F. B.; Bebell,Lisa M.
Introduction: Maternally transferred immunity plays a critical role in protecting infants from pathogens during the vulnerable period before their own immune system matures. Maternal IgG antibodies cross the placenta and circulate in the infant's body for up to six months, providing protection against infection. Despite the importance of transplacental antibody transfer, the mechanisms regulating this process remain poorly understood, including the maternal and placental factors that influence transfer efficiency. Methods: We analyzed maternal and umbilical cord plasma samples from 354 mother-baby dyads in Uganda, of whom 176 were pregnant individuals with HIV taking antiretroviral therapy. We measured IgG antibody levels, antibody binding to Fc receptors, antibody glycosylation, and cytokine levels in both maternal and umbilical cord compartments, and we evaluated placental inflammatory states with placental histology. We used this comprehensive dataset to define correlations between transplacental antibody transfer efficiency, placental inflammation, and cytokine levels in maternal and umbilical cord plasma. Results: We found that antibody transfer efficiency was not impacted by histologic placental inflammation, but that higher maternal plasma levels of the pro-inflammatory cytokines TNF-a and IL-6 correlated with reduced transplacental antibody transfer efficiency. Discussion: Our findings reveal the impact of maternal systemic inflammation on antibody transfer across the placenta and have implications for maternal immunization strategies.
Simulation of a multi-parameter indoor air purification control system using fuzzy logic–PID hybrid control tuned by genetic algorithms
(Scientific Reports, 2026-07-15) Wanzala,Jimmy Nabende; Awamu,Abdallah; Atim,Michael Robson; Mugabe,Robert; Awal,Moses
Indoor Environmental Quality ( IEQ)critically im- pacts human health, comfort, and productivity yet existing air purification systems often rely on single-sensor data, manual monitoring, or cloud-based processing, resulting in limited intelligence, increased latency, and reduced suitability for resource- constrained environments such as Uganda. To address these challenges, this paper presents the design, mathematical modelling, and simulation of a novel multi-parameter indoor air purification control system. The proposed frame work integrates six environmental sensors (dust, smoke, gas, volatile organic compounds (VOC), temperature, and humidity) with a hybrid Fuzzy Logic–Proportional–Integral–Derivative (PID)controller whose parameters are optimally tuned using a Genetic Algorithm (GA).A first-principles mathematical model of the DC motor-driven purifier actuatoris developed to describe the relationship between the input voltage and volume tric air flow. To enhance the physical realism of the simulations, the system model incorporates practical operating factors, including fan inertia, room-air mixing, pollutant transport, filter dynamics, and environmental sensor noise. The complete system is implemented in the MATLAB/Simulink environment using a Mamdani-type fuzzy inferencesystemwith49rulesandaGA-basedmulti- objective fitness function to minimize over shoot, settling time, and steady-state error. Simulation results demonstrate that the proposed GA-tuned Fuzzy-PID controller achieves a stable and well-damped responsewitharisetimeof53.93s, a settling time of 119.99s, and amaximumovershootof2.14%, reflecting the expected dynamics of practical indoor air purification systems. The proposed frame work provides a scalable, low-latency, and network-independent intelligent control solution that establishes a realistic foundation for future prototype development and experimental validation in low-resource environments.