School of Public Health (Public-Health) Collections

Permanent URI for this collection

http://10.29.1.53:4000/handle/10570/184

Browse

Recent Submissions

Now showing 1 - 5 of 918
  • Item
    Assessing health facilities preparedness in detecting and controlling effectively cholera outbreaks in Kampala City Uganda
    (Makerere University, 2025) Kajumba, Pamela
    Abstract Background Cholera remains a global health challenge. In Kampala City, Uganda, rapid urbanization, poor sanitation, and high population density favor recurrent outbreaks. Evidence on health facility preparedness to detect and respond to cholera is limited. This study assessed preparedness of health facilities in Kampala, focusing on readiness, Water, Sanitation, and Hygiene (WASH) infrastructure, laboratory capacity, and public–private differences. Methods A cross-sectional study was conducted in 52 health facilities across Kampala’s five divisions. Data were collected using questionnaires adapted from international cholera preparedness checklists. Preparedness was defined as the presence of essential components for timely cholera detection and management, including trained staff, surveillance systems, case management protocols, laboratory capacity, and adequate WASH infrastructure. Response capacity was assessed by ability to promptly detect, report, and manage cholera cases in line with national guidelines. Quantitative data were analyzed using STATA version 15 to generate descriptive statistics and compare preparedness between public and private facilities. Results Fifty-two facilities (26 public, 26 private) were evaluated; 45 (86.5%) were prepared for cholera detection and response. Public health facilities showed higher preparedness than private (25 [96.2%] vs 20 [76.9%], p=0.042). Most facilities had adequate WASH infrastructure: 50 (96.2%) had functional toilets and handwashing stations, and 48 (92%) had chlorine for disinfection. However, only 34 (65.4%) had functional laboratories equipped to diagnose cholera. Conclusion Health facilities in Kampala demonstrated good overall preparedness and WASH infrastructure. However, gaps in laboratory diagnostic capacity, particularly in private facilities, may delay timely cholera detection and response. Strengthening laboratory services and engaging private facilities in preparedness initiatives are essential for comprehensive cholera control in urban settings.
  • Item
    Antibiotic prescription practices and associated factors among health workers in managing Pneumonia among children (<5 years) at Mengo and Lubaga Hospitals
    (Makerere University, 2024) Kamoga, Gonzaga Cena
    Introduction: Pneumonia is a major cause of morbidity and mortality in children under five years in low-and middle-income countries like Uganda. Inappropriate antibiotic prescription contributes to antimicrobial resistance and poor treatment outcomes. Objectives: To evaluate the antibiotic prescription practices and associated factors among children under five years admitted with pneumonia in the pediatric wards of Mengo and Lubaga hospitals in Kampala, Uganda. Methods: This was a cross-sectional study that collected data retrospectively from inpatient registers of children under five years with pneumonia admitted at Mengo and Lubaga hospitals. All inpatient records at the pediatric wards in the year 2022 that met the eligibility criteria were reviewed. Data was collected using a data abstraction tool. A data entry screen was developed in EpiData using checks, and data was entered in duplicate. The data was then transferred to STATA version 14 and cleaned prior to the analysis. Data analysis involved univariate, bivariate, and multivariate analyses using modified Poisson regression methods at 95% confidence level. Results: From the 678 files assessed for inappropriate antibiotic prescription, 375 (55.3%) were male, with a median age of 15 months and a median weight of 9.8 kg. Most files (99.7%) had a confirmed pneumonia diagnosis. The majority (82.0%) had two antibiotics prescribed, predominantly ceftriaxone and gentamicin (28.6%). Inappropriate antibiotic prescription prevalence was 39.7%, with 95.3% based on UCG, 48.7% on Lubaga's protocol, and 40.5% on Mengo's protocol. Bivariate analysis indicated age, weight, and hospital stay duration as significant factors. Multivariate analysis revealed that weight (aPR: 0.979, p=0.027) and hospital stay duration (aPR: 0.959, p=0.049) were significantly associated with inappropriate antibiotic prescriptions. Conclusion: This study highlights a significant reliance on dual antibiotic therapy for children under five with pneumonia in Mengo and Lubaga hospitals, with ceftriaxone and gentamicin being the most prescribed combination. High inappropriate antibiotic prescription rates based on UCG guidelines suggest deviations from recommended practices. Notably, child weight and hospital stay duration were linked to inappropriate prescription, emphasizing the need for enforcement of guideline adherence by health workers and continuous healthcare provider training to optimize antibiotic use and combat antimicrobial resistance.
  • Item
    Integrating machine learning into DHIS2 to predict multidrug resistant tuberculosis: a case of the DHIS2's electronic case based surveillance system
    (Makerere University, 2025) Outeke, Mark
    Multidrug-resistant tuberculosis (MDR-TB) remains a critical public health challenge in Uganda, where cases continue to rise. The existing electronic Case-Based Surveillance System (eCBSS) tracks patients and treatment adherence for Drug-Susceptible TB (DS-TB) but lacks integration with machine learning (ML) for predictive analytics and data-driven decision-making. This study developed and tested an ML-powered web application integrated into DHIS2’s eCBSS to predict the likelihood of MDR-TB development among DS-TB patients and identify geographic hotspots. Objective: The study aimed to gather user requirements, develop, and validate an ML-integrated analytical dashboard within DHIS2’s eCBSS to support healthcare workers in predicting MDR-TB risk and visualizing hotspots through a GIS module. Methodology: A Design Science Research approach guided the development, using a User-Centered Design framework. In-depth interviews were conducted with healthcare workers from three high-volume TB treatment facilities in Uganda’s Central Region, all using eCBSS. Requirements informed the development of a one-dimensional deep neural network (1DNN) in Python and a web interface built with JavaScript and Node.js for local DHIS2 deployment. The ML model was validated using accuracy, AUC, and confusion matrices, while usability testing assessed ease of use and satisfaction. Results: Two major themes emerged: user requirements (functional and non-functional) and implementation needs. Functional needs included real-time alerts, risk scoring, and adherence tracking, while non-functional needs emphasized scalability, security, and usability. The 1DNN achieved 92% accuracy and an AUC of 0.96, demonstrating strong predictive performance. The GIS module effectively visualized MDR-TB hotspots, supporting targeted interventions. Usability testing showed high satisfaction (mean scores 4.65–4.87/5), though participants recommended improved search, filtering, and clearer error handling. Conclusion: Integrating ML and GIS into DHIS2’s eCBSS significantly enhanced predictive capacity, spatial analysis, and resource allocation, demonstrating the potential of AI-driven tools to strengthen MDR-TB management in Uganda.
  • Item
    Early diagnosis of ischemic stroke on non-contrasted CT scan using machine learning. A case study of Mulago National Referral Hospital
    (Makerere University, 2025) Kateregga, Michael
    Background Ischemic stroke is a major cause of disability and death, especially in low-resource settings such as Uganda. Timely diagnosis using non-contrast CT scans is essential but often delayed due to the limited number of radiologists and the complexity of image interpretation. This study developed and tested a machine learning model to support faster and more accurate detection of ischemic stroke at Mulago National Referral Hospital. Methods A dataset of 1,000 non-contrast CT scans was collected and annotated by three radiologists. The images were preprocessed through intensity normalization and texture feature extraction. A convolutional neural network (CNN) was trained and validated using five-fold cross-validation. Model performance was measured using sensitivity, specificity, accuracy, and area under the receiver operating characteristic curve (AUC). A web-based platform and RESTful API were also developed to enable clinical testing and deployment. Results The CNN achieved an average accuracy of 99.4%, with 100% sensitivity and 98.9% specificity across all validation folds. The AUC was 0.995, showing excellent discrimination between stroke and non-stroke images. When compared to a senior radiologist, the model detected all stroke cases correctly, producing results within three seconds per scan. The web interface allowed image upload, automatic analysis, and generation of a diagnostic summary report. Conclusion The developed model accurately detects ischemic stroke on CT scans and performs comparably to an experienced radiologist while offering faster results. The accompanying platform demonstrates readiness for integration into hospital workflows. This study highlights the potential of machine learning to improve diagnostic efficiency and stroke care in Uganda and similar healthcare settings.
  • Item
    Enhancing lung cancer diagnosis in Uganda using convolutional neural networks on CT images from Uganda and Iraq
    (Makerere University, 2025) Kamuhanda, Success
    Lung cancer remains the leading cause of cancer-related mortality worldwide, with over 2.2 million new cases and 1.8 million deaths annually. Early detection significantly improves survival rates; however, in low- and middle-income countries (LMICs) such as Uganda, diagnosis often occurs at advanced stages, leading to poor treatment outcomes and high mortality. Computed Tomography (CT) scans are the gold standard for early detection, but their impact is constrained by a severe shortage of radiologists and pulmonologists, Uganda has fewer than 50 radiologists and under ten pulmonologists for over 45 million people. Despite government investments in CT infrastructure, including regional scanner installation and a National Tele-Radiology Centre, specialist shortages continue to delay diagnosis and underutilize imaging capacity. This study developed and validated a lightweight, 10-layer Convolutional Neural Network (CNN) for automated classification of lung CT scans into cancerous and non-cancerous categories. Using a combined dataset from the public Iraq-Oncology Teaching Hospital/National Center for Cancer Diseases (IQ-OTH/NCCD) repository and de-identified CT scans from Mulago National Referral Hospital. The final model achieved a test-set accuracy of 98%, 97.5% sensitivity, 98.0% specificity, and an area under curve (AUC) of 1.00 on the test set. To address the gap between algorithm development and clinical use, the model was integrated into a web-based diagnostic interface built with Flask backend and React.js frontend, enabling real-time predictions from uploaded scans. Manual testing of 20 unseen images via the interface yielded 95% accuracy, confirming operational stability outside controlled environments. By reducing reliance on scarce specialist interpretation, optimizing existing CT infrastructure, and enabling faster, more accurate lung cancer detection, this system has potential to improve early diagnosis rates and reduce mortality in Uganda. Its Central Processing Unit (CPU), only deployment capability ensures suitability for resource-limited settings, supporting scalable integration into national cancer control programs and public health screening initiatives. Key Words: Lung cancer, Convolutional Neural Networks, early detection, CT scans, deep learning, Uganda, artificial intelligence, medical imaging, web-based application