The Nelson Mandela African Institution of Science and Technology
Institutional Repository
Research Collection
Select to browse Research collections.
Recent Submissions
A Mobile-Based Machine-Learning Model for Predicting Maintenance of Dental Machines for Health Facilities: A Case of Ministry of Health in Uganda
(Springer Nature, 2026-04-01) Ayebare, Famina; Kaijage, Shubi; Mduma, Neema; Kiwoli, Liston; Kaluuma, Hillary; Byarugaba, Lambert; Ssanyu, Lydia
The health sector remains underfunded, falling short of the 15% budget allocation recommended by the Abuja Declaration of which Uganda is a member. Between 2010 and 2016, the budget of the health sector was an average of 7.8% of the national budget. In 2020/2021, it accounted for 5.1% of the national budget, from the previous 7.9% in the previous financial year. Major issues facing the health system include grossly underpaid health practitioners, scarcity of health workers, and necessary equipment in government facilities. The concept of machine learning is ubiquitous, and different algorithms have been used to train and build models to come up with different solution in the health sector. The objective of this project was to come up with a mobile-based machine-learning model for predicting the maintenance of dental machines based on the failure type to determine whether the machine is in good condition to work on patients efficiently or needs repair and maintenance. The project used both quantitative and qualitative methods of data collection. Random Forest and XG boost classifiers were used to train and test the model using structured data. The dataset contained 8 features, which were the product ID, type of machine, UDI (unique identifier), air temperature, process temperature, rotational speed, torque, and tool wear. Heat dissipation, power failure, random failure, overstain, and no failure were labels used to predict the type of failure that could occur, XG boost classifier emerged as the best with accuracy score of 97.3% in comparison with Random Forest after the accuracy score and confusion matrix of the two algorithms.
Anaerobic Digestion and Composting: A Sustainable Solution for Organic Waste Management
(Elsevier Ltd., 2026) Pegu, Ratul; Paul, Sarmistha; Paul, Bishal; Chetry, Akash; Hezron, Elkana
Solid waste production is growing quickly on a global scale every day, which has an influence on the environment in many ways. Organic and inorganic wastes make up solid waste. Due to its putrefying nature, organic waste creates a nuisance in the environment. Nowadays, one of the biggest issues faced in all developing nations is the handling of organic waste. Organic waste originates from various sources such as schools, hospitals, markets, workplaces, restaurants, households, and wastewater treatment plants. The economy, ecology, and public health are all seriously threatened by the rapidly rising volume of organic waste. Organic wastes 150are biodegradable and can be converted to value-added products using suitable techniques. To manage organic waste sustainably, the cause and characteristics of its creation must be identified, followed by appropriate management measures. This chapter discusses the role and effectiveness of anaerobic digestion and composting as long-term sustainable organic waste management solutions. Understanding how organic waste impacts the ecosystem’s health, how anaerobic digestion and composting can manage organic waste, and the benefits of these techniques will be highlighted in this chapter.
Fluoxetine enhances reproductive output without affecting spawning-site selection in turquoise killifish
(Elsevier Ltd., 2026-06) Kimario, Theresia; Goede, Nicole; Buratto, Noémie; Kafula, Yusuph; Missawi, Omayma; Thoré, Eli
Pharmaceutical contamination of aquatic ecosystems is widespread, with mounting evidence that pollutants such as the antidepressant fluoxetine can boost fish offspring production. But such increases may come at a hidden cost if pharmaceuticals also disrupt key reproductive behaviors—like where fish choose to spawn—that determine whether those offspring ultimately survive. To test this possibility, we exposed turquoise killifish (Nothobranchius furzeri) to a field-realistic concentration of fluoxetine (317 ng/L) in a controlled laboratory experiment. Fish could choose among spawning sites of white, orange, or black sand, both in single pairs and in groups. Fluoxetine-exposed fish produced about 50% more eggs in single pairs, but this effect was not observed in groups. Regardless of social context, fish deposited roughly four times more eggs on black sand than on orange or white sand, and fluoxetine did not alter spawning-site preference. In addition, fluoxetine did not affect habitat preference in a non-mating context, with both control and fluoxetine-treated fish spending more time over darker backgrounds. These results show that environmentally relevant fluoxetine concentrations can enhance reproductive output without necessarily affecting spawning-site selection. The broader ecological significance of such changes will depend on how pharmaceutical-induced shifts in reproduction interact with natural social and habitat dynamics in the wild.
Crimean–Congo Hemorrhagic Fever Virus in Africa: Epidemiological Trends, Transmission Ecology, Hotspot Heterogeneity, and Preparedness Challenges— A Narrative Review
(MDPI, 2026) Shao, Elichilia , Pyuza, Jeremia , Kibona, Tito , Shirima, Laura , Mlay, Eliaichi , Andongolile, Alice , Kayaga, Ray , Kilonzo, Semvua , Mmbaga, Blandina and Chilongola, Jaffu
Background: Crimean–Congohemorrhagic fever virus (CCHFV)is animportant tick-borne zoonosis and an emerging public health threat across Africa. Although evidence of viral circulation is mounting, information remains fragmented, limiting a comprehensive understanding of transmission ecology, regional hotspot heterogeneity, and preparedness needs across the continent. Methods: This narrative review critically synthesized published literature on CCHFV in Africa, identified through PubMed, Scopus, and Google Scholar and supplemented by citation tracking and authoritative public health reports.
Evidence from epidemiological, ecological, molecular, surveillance, and One Health studies was integrated to examine transmission dynamics, geographic hotspot distribution, viraldiversity, risk factors, diagnostic and surveillance challenges, and preparedness strategies. Results: Available evidence shows marked geographic heterogeneity in CCHFV transmission across Africa, with hotspot regions shaped by ecological suitability, Hyalomma tick distribution, livestock–human interactions, and health system capacity. Livestock consistently show higher exposure than humans, underscoring their role as key indicators of viral circulation. Diagnostic limitations, passive surveillance, ecological variability, and serological cross-reactivity contribute to substantial under recognition of disease burden, while molecular studies reveal considerable viral diversity and ongoing evolution across African regions. Conclusions: CCHFV remains underdiagnosed and underreported in many African settings because of limited surveillance and diagnostic capacity. Strength ening integrated One Health surveillance, expanding laboratory and genomic capacity, utilizing livestock as sentinel populations, and improving cross-sectoral collaboration are critical for enhancing early detection, outbreak preparedness, and effective public health response across the continent.
Integrating 3D video tracking with the standard WHO tunnel assay: a proof-of-concept to support improving insecticide-treated nets for mosquito control
(Springer Nature, 2026-06-08) Bredt, Beatrice; Fatou, Mathurin; Lugenge, Aidi; Kamande, Dismas; Liechti, Nathalie; Moore, Sarah; Müller, Pie
The World Health Organization (WHO) tunnel test is a standardised laboratory assay used to characterise the biological availability and potency of active ingredients on the surface of an insecticide-treated bed net (ITN) against host-seeking mosquitoes. However, the assay provides only endpoint measurements—the proportions of mosquitoes killed and blood-fed – and therefore offers no insight into how mosquitoes interact with the ITN. Therefore, complementary behavioural data would be highly valuable, for example to reveal the extent to which mosquitoes engage with the net, thereby helping to explain differences in endpoint outcomes, or to determine the minimum assay duration required and thus improve throughput. For capturing mosquito behaviour in detail, automated three-dimensional (3D) video tracking offers a powerful approach. Here, we present a proof-of-concept study combining the WHO tunnel assay with Trackit3D, a versatile tracking system, in a laboratory in Tanzania where tunnel assays are routinely performed. The system successfully tracked multiple mosquitoes simultaneously as they were attracted to a rabbit, including measuring the duration of contact with the net, despite typical fluctuations in power supply and lighting conditions. The ability to obtain high-resolution trajectories within the WHO tunnel assay provides new opportunities to enhance the behavioural evaluation of ITNs and strengthens both the interpretability and utility of the assay.