School of Public Health (Public-Health) Collections
Permanent URI for this collection
Browse
Recent Submissions
1 - 5 of 974
-
ItemLevel of access to and utilisation of flood early warning information in Kasese District, Western Uganda(Makarere Unversity, 2026)Background: Flooding is a recurrent and destructive hazard that disproportionately affects low- and middle-income countries. In Kasese District, flood risk has intensified due to climate change, environmental degradation, and the expanding human activities in flood-prone areas. While FEWS are critical for disaster risk reduction, their effectiveness depends on communities’ access to and utilisation of FEWI. Guided by the Protection Motivation and Diffusion of Innovations theories, this study assessed levels of access to and utilisation of FEWI, and the factors influencing their use in Kasese District. Methodology: An explanatory sequential mixed-methods design within a cross-sectional framework was applied across 11 flood-prone sub-counties. Quantitative data were collected from 424 households using KoBoCollect and a multistage sampling approach. Analyses in STATA 14 included descriptive statistics, Chi-square tests, and modified Poisson regression to assess FEWI access and utilisation. Survey results informed a qualitative phase comprising 12 key informant interviews with government officials, NGO representatives, and community leaders. Thematic analyses identified contextual, institutional, and socio-cultural drivers, barriers, and enablers of FEWI utilisation. Integration of quantitative and qualitative findings provided deeper insights into FEWI in Kasese District. Results: Of the 424 respondents (99.5%) reached, the mean age was 43.4 years, and SD 14.2, with 47.6% aged 34-55 years. Just over half were male (52.6%); most lived in rural areas (63.4%), were married (77.6%), and farmed (65.8%). Only 4.9% had a tertiary education. Overall, 70.1% (n=297) accessed FEWI, and among these, 74.1% (n=220) utilised compared to 60.1% (n=255) in the total sample. Although Access and utilisation were consistent across groups, Utilisation varied across sub-counties (16.4% in Bugoye to 96.9% in Nyamwamba). Key predictors of utilisation included organised warning systems (74.0%), trusted intermediaries such as traditional leaders (81.6%), local government (79.3%), disaster committees (79.3%), and messages via meetings or barazas (79.3%) or family/neighbour (74.8%). Also, Household preparedness (74.8%) strengthened responsiveness. Conclusion: While FEWI access in Kasese was relatively high, utilisation remains inconsistent, influenced by trust in local actors, preparedness, and institutional barriers. Structural, socio-cultural, and coordination challenges hinder warning effectiveness, especially for rural and vulnerable groups. Community-driven, inclusive, and contextually tailored FEWI strategies that harness traditional knowledge, enhance local communication, and address gender and sub-county disparities are recommended to strengthen flood risk response. Keywords: Flood warning systems; Access to flood early-warning information; Utilisation of flood early-warning information; Disaster risk reduction; Kasese District, Uganda.
-
ItemReadiness of HIV clinics in Nakaseke District to integrate hypertension care(Makerere University, 2026)Background: With the long lives made possible by the antiretroviral therapy (ART) programs globally, hypertension is on the rise among People Living with HIV (PLWH), especially in sub-Saharan Africa. Integrating hypertension into HIV programs promises opportunities for improved health outcomes. However, there is insufficient evidence on the capacity of rural Ugandan HIV primary healthcare facilities to integrate hypertension care. Objective: To assess the extent of hypertension screening among people living with HIV and the readiness of HIV clinics in Nakaseke District to integrate hypertension care. Methods: We conducted a mixed-methods cross-sectional study among nine HIV clinics (two hospitals, two Health Centre IVs, and five Health Centre IIIs). We assessed the facilities’ readiness to integrate hypertension using a modified WHO PEN tool across five domains: policy and clinical guidelines, human resources, essential medicines, basic equipment, and health information systems and records. We reviewed 411 records of PLWH using a data abstraction tool to establish the proportion of PLWH screened for hypertension. We conducted 16 in-depth interviews with PLWH and 15 key informant interviews with healthcare workers to explore facilitators and barriers to integration. Quantitative data were analyzed using STATA Version 14, while qualitative data were analyzed thematically using NVivo. Results: Readiness of HIV clinics in Nakaseke District to integrate hypertension care was low, with no facility meeting the 70% threshold. Key gaps were in medicines, equipment availability, records, and policies. Only 74.7% of PLWH had documented blood pressure screening. Screening was likely among PLWH diagnosed more than a year (aPR 1.69; 95% CI 1.15-2.49), and receiving care at a hospital (aPR 1.4; 95% CI 1.25-1.57) and less likely among PLWH on repeat visits (aPR 0.56; 95% CI 0.39-0.8), and those receiving care at health centre IVs (aPR 0.49; 95% CI 0.39-0.68). Barriers to integration were inadequate resources, a lack of capacity-building initiatives for the staff, and a varying refill and drug delivery model for hypertension. Facilitators for hypertension and HIV integration were training of the staff, support from implementing partners, and the perceived benefits of integrating care. Conclusion: HIV clinics in Nakaseke District are not ready to integrate hypertension care. They have major gaps in equipment, medicines, policies, and records. To enable integration, more attention should be paid to strengthening the information systems, workforce capacity, and providing essential equipment and hypertension medicine.
-
ItemDeep learning and high image processing for diagnosis of Zoonotic Bovine Tuberculosis (Mycobacterium Bovis) in Cattle(Makerere University, 2025)Background: Zoonotic tuberculosis (zTB), caused by Mycobacterium bovis, remains a persistent public health and economic burden in low- and middle-income countries (LMICs). Conventional diagnostic approaches such as tuberculin skin testing and post-mortem inspection are slow, costly, and often unreliable in rural abattoirs. These limitations lead to underreporting and delayed interventions, increasing the risk of human infection through contaminated meat and milk. Recent advances in deep learning, particularly Convolutional Neural Networks (CNNs), offer opportunities to improve diagnostic speed and accuracy in resource-constrained settings. Aim: This study developed and evaluated a CNN-based diagnostic tool for detecting zTB from high-resolution lymph node images of cattle, designed to enhance abattoir-level and veterinary diagnosis in Uganda. Methods: A dataset of 278 lymph node images (zTB-positive and negative) was compiled from the International Livestock Research Institute (ILRI), European Reference Laboratory, and Makerere University’s College of Veterinary Medicine. Images were preprocessed—resized, normalized, and augmented—to ensure consistency and improve robustness. An EfficientNet-B0 CNN, pre-trained on ImageNet, was fine-tuned and evaluated using 70% training, 15% validation, and 15% testing splits. Model performance was measured using accuracy, precision, recall, F1-score, specificity, and AUC-ROC. A web-based interface was developed to demonstrate real-time diagnostic capability. specificity, F1-score, and AUC-ROC. A web-based interface was also developed to demonstrate the model’s potential for real-time diagnostic application. Results: The model achieved over 90% accuracy, sensitivity, and specificity on the test set, with an AUC-ROC above 0.95, confirming excellent discriminative ability. The web interface allowed users to upload images and obtain predictions instantly, demonstrating field-level feasibility. Conclusion: The CNN-based tool significantly improved the accuracy and efficiency of zTB diagnosis compared to conventional methods. By enabling real-time abattoir-level screening, this innovation enhances food safety, strengthens zoonotic surveillance, and supports the One Health approach in Uganda. Keywords: Zoonotic tuberculosis, Mycobacterium bovis, Convolutional Neural Networks, Deep Learning, Veterinary Diagnostics, Image Processing
-
ItemBarriers, facilitators and effect of using locally formulated therapeutic foods on treatment success among children admitted with Severe Acute Malnutrition in Uganda : a mixed method study(Makerere University, 2026)Background: Severe Acute Malnutrition (SAM), defined as severe wasting or nutritional oedema, remains a major cause of child mortality in Uganda. Frequent stockouts of commercial therapeutic foods (CTFs) in inpatient therapeutic care (ITC) facilities have necessitated reliance on locally formulated therapeutic foods (LTFs). However, evidence on their effect on treatment success, as well as the barriers and facilitators influencing their use, remains limited. Objective: To investigate the association between treating SAM with LTFs versus CTFs on SAM in Uganda. The barriers and facilitators influencing the use of LTFs by health workers were also explored. Methods: A retrospective cohort study was conducted by extracting data from the integrated nutria tion register (INR) of children who received LTFs and CTFs during the stockout of CTFs (January to September 2023) and no stockout of CTFs (October 2023 to June 2024), respectively, across 22 health facilities in Uganda. The association between receiving LTFs versus CTFs and treatment success was tested by modified Poisson regression analysis using STATA 15 to report adjusted risk ratios (aRR) at 95% confidence interval (CI). In addition, in-depth interviews (IDIs) were conducted with 22 purposively selected ITC in-charges to explore facilitators and barriers to the use of LTFs. Data generated from IDI was analyzed using themes. Results: Of the 340 children, 48.5% received LTFs during the stockout periods. Treatment success rates were 80.3% and 71.3% for commercial and locally formulated therapies, respectively, with no statistically significant differences observed (p=0.08). Multivariable analysis identified nutrition status on admission, SAM with oedema (+) aRR1.14 (95%Cl:1.02-1.27, P=0.02), and Weight-for-Height Z-score above -2 standard deviations [aRR 1.13, 95%CI: 1.06-1.10, P<0.001] were likely to achieve treatment success. In contrast, defaulters from ITC were less likely to achieve treatment success, aRR 0.28 (95%CI:0.15-0.55, P<0.001). Qualitative findings showed that affordability, accessibility, cultural acceptability, perceived effectiveness, and caregiver empowerment facilitated LTF use, while labor-intensive preparation, spoilage risks, and added financial burden on facilities were key barriers. Conclusion: Use of locally LTFs during the stockout period of CTFs has an equal potential as CTFs to treat SAM among children in Uganda. Health facility administrators need to support and address facilitators and barriers perceived by ITC health workers to effectively use LTFs, respectively
-
ItemA Machine Learning Model for Prediction of Antibiotic Resistance with Escherichia Coli Infections Using Demographic, Clinical and Microbiological Data.(Makerere University., 2026-01-13)IntroductionIn, low- and middle-income countries like Uganda, there is growing reliance on empirical prescription of broad-spectrum antibiotics which, while targeting a wide range of pathogens, contributes to the development of resistance to common pathogens such as E. coli. This challenge is compounded by the poor selection of antibiotic panels in many laboratories, which often fail to reflect local resistance patterns and patient-specific factors, leading to inefficient use of scarce resources and delayed appropriate treatment. Objectives of the study The objectives of this study were to; 1) identify risk factors for drug resistant E. coli infections using machine learning techniques; 2) evaluate the performance of different machine learning models in predicting the likelihood of drug resistance among patients with E. coli infections using demographic, clinical and microbiological data; and 3) develop a web-based interface to support proper antibiotic prescription and targeted antimicrobial decision-making. MethodologyA retrospective analysis was conducted on 1,552 records of patients diagnosed with E. coli infections in 10 tertiary healthcare facilities in Uganda. These records were analyzed using machine learning models including Lightgbm, xgboost, random forest, gradient boosting, and decision trees. Feature selection was guided by a weighted importance score and frequency count framework. The best performing model was deployed in a streamlit-based web interface.Results Key predictors of resistance included antibiotic type, patient age, hospital site, specimen type, prior antibiotic use, and hospitalization history. XGboost emerged as the top-performing models for prediction of drug resistance, with an accuracy of 82.32%, a precision of 82.36%, recall of 85.37%, F1 score of 83.84%, and ROC AUC of 90.17%. The web-based interface was implemented using python streamlit technology, intergrated with the best performing model to enable real-time resistance prediction.ConclusionThis study demonstrates the potential of machine learning to transform antimicrobial resistance surveillance and clinical decision-making in resource-limited settings.