Abstract
Background: Artificial intelligence (AI) transformed healthcare worldwide and has the potential to address challenges faced in the South African healthcare sector, such as limited public institutional capacity, staff shortages, and variability in skills levels that exacerbate the demand on the healthcare system that can lead to compromised care and patient safety.
Aim: This study aimed to describe how AI, especially machine learning is used in healthcare in South Africa over the last 5 years.
Method: The Joanna Briggs Institute (JBI) methodology for scoping reviews was used. Peer-reviewed articles in English, which were published from 2020 to date were sourced and reviewed using the Population, Concept, Context (PCC) framework.
Results: A total of 35 articles were selected. The results showed a focus on conventional machine learning, a health focus on HIV and/or tuberculosis (TB) and cancer, and a lack of big data in fields other than cancer.
Conclusion: There has been an increase in the use of machine learning in the analysis of health data, but access to big data appears to be a challenge.
Contribution: There is a need to have access to high-quality big data, inclusive policies that promote access to the benefits of using machine learning in healthcare, and AI literacy in the health sector to understand and address ethical implications.
Keywords: artificial intelligence; deep learning; health sector; machine learning; South Africa.
Introduction
South Africa has a quadruple burden of diseases, which severely impacts the delivery of healthcare (De Villiers 2021; Louw et al. 2023; Okeibunor et al. 2023). Healthcare is often inaccessible to people living in remote areas (De Villiers 2021; Louw et al. 2023), and most people depend on limited public resources. Limited public institutional capacity, staff shortages, high staff turnover rates, and variability in skills levels exacerbate the demand on the healthcare system and can lead to compromised care and patient safety (De Villiers 2021).
Artificial intelligence (AI) has been posited to address some of these challenges. Artificial intelligence refers to the imitation of human intelligence by automated processes (Kuziemsky et al. 2019) such as the use of machine learning (i.e., training machines to learn from datasets and performs tasks; Shandhi & Dunn 2022) and the use of generative AI (i.e., the use of large language models to do a range of tasks). Artificial intelligence has been applied in healthcare to assist in the diagnoses of diseases, analyse healthcare plans, monitor health, develop personalised treatment plans, and perform surgical treatment through robotics (Amisha et al. 2019). Artificial intelligence shows potential for personalised medicine as individual risk can be predicted from patient data that enables the development of individualised treatment plans, which can enhance patient outcomes (Amisha et al. 2019; Kuziemsky et al. 2019; Shandhi & Dunn 2022). Artificial intelligence can also streamline healthcare delivery and provide more efficient healthcare, for example, online appointment scheduling, online check-ins, digitalisation of medical records, and automatic reminders for follow-up appointments (Amisha et al. 2019).
Within the field of AI, machine learning (ML) has been used in predictive analytics where predictive models are formed by combining machine learning and traditional statistics and are used for prognosis, optimising healthcare delivery, and individualised treatment (Manlhiot 2018). Machine learning includes conventional machine learning (which focuses on developing algorithms and models that permit computers to learn from information); deep learning (which uses neural networks to simulate the human brain to learn from unstructured data, such as images and speech recognition); and ensemble learning (which combines multiple models to improve accuracy and robustness), and can be applied to both conventional and deep learning models (Sharifani & Amini 2023). Machine learning models are useful to determine which patients are likely to benefit from a particular treatment based on certain characteristics (e.g., genetically informed therapeutic planning) (Dong et al. 2015; Shandhi & Dunn 2022) and to identify biomarkers that can assist in early detection of disease, prediction of treatment response, and provide indicators of the progression of diseases (Shandhi & Dunn 2022).
Review question
The following is the review question: ‘What articles on AI used in healthcare in South Africa were published in the last 5 years?’ The objectives were to determine: (1) what the status of using AI in the South African health system is, and (2) what applications of AI-based innovations are currently in use in the health sector in South Africa. As no generative AI articles were found, the review focused on machine learning.
Materials and methods
Research design
This scoping review followed the guidelines of the Joanna Briggs Institute (JBI) for evidence-synthesis (Aromataris et al. 2024; Peters et al. 2020), and the reporting is done in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for scoping reviews (PRISMA-ScR) checklist (Tricco et al. 2018).
Protocol and registration
The review protocol was registered in Open Science Framework (OSF) prior to conducting the study, and is available at https://doi.org/10.17605/OSF.IO/2SMU4
Eligibility criteria
Given the rate of development of technology, this scoping review collected and analysed peer-reviewed full-text empirical research articles written in English from 2020 to 2024 to ensure that only the latest research articles pertaining to the use of AI in healthcare settings in South Africa are included. All study methods were considered, except non-research studies such as reviews and opinion article. Using the PCC (population, concept, context) framework (Peters et al. 2020), inclusion and exclusion criteria were formulated (Table 1). The common types of machine learning are defined in Appendix 2. Studies that did not comply with the inclusion criteria were excluded.
| TABLE 1: Inclusion and exclusion criteria based on participants, context, and concept. |
Search strategy, selection of sources, and data extraction
We used the key terms of AI, machine learning and deep learning synonyms, South Africa and health sector (see Appendix 1, Table 1-A1 for detailed search strategy), to conduct searches on PubMed, Academic search Complete, and Google Scholar, to retrieve the most relevant data. Descriptors such as titles, abstracts, and full texts of relevant and suitable articles on the topic were used to conduct the search. Relevant studies from the reference lists of retrieved articles were hand-searched from Google Scholar database (‘snowballing’).
Following the literature search, all identified sources were imported to the systematic review management software, namely Covidence (2023), which guided the screening process comprising the screening of titles and abstracts, review of full texts, and data extraction. Two reviewers (the third and fourth authors) screened the sources independently for eligibility. Disagreements between the reviewers at each stage of the selection process were resolved through online discussion by the reviewers and checked by the first author. The PRISMA-ScR flowchart was used to illustrate the search process (the removal of duplicates, the number of qualifying sources included for analysis, and the reason for excluding non-eligible sources). Using a data extraction tool in Excel, data were extracted by the fourth author, and checked by the first author. There was no formal quality assessment of the sources included in the review because a formal assessment of the methodological quality is usually not required for scoping reviews (Peters et al. 2020).
Data analysis
For scoping reviews, frequency counts for the required fields of data are sufficient, although a more in-depth analysis can be appropriate (Peters et al. 2020). Frequency counts and thematic analysis Braun and Clarke (2006) were performed by the first author. Themes of health area, source of data and type of machine learning were analysed.
Ethical considerations
This article followed all ethical standards for research without direct contact with human or animal subjects.
Review findings
Selection of sources of evidence
A total of 810 records were retrieved for the period from 2020 to 2024, with one current unpublished study included. Thirty-five articles (n = 35) were selected for final analysis (Figure 1). Of the 35 articles, 9 (25.7%) were published between 2020 and 2021, and 26 (74.3%) between 2022 and 2025 (Table 2).
 |
FIGURE 1: PRISMA-ScR flowchart illustrating the search process. |
|
| TABLE 2: A summary of articles (N = 35). |
A summary of the studies included in the review
Three articles were on the same study focusing on developing a machine learning algorithm to predict mortality in critically ill children (Pienaar et al. 2022a, 2022b, 2023). Most data were primary data (n = 14, 42.4%) or from established data registries such as the South African Cancer Registry, the National Health Laboratory Services (NHLS) and the World Data registry of COVID-19 cases (Table 2).
Characteristics of the studies
The reviewed studies used a range of conventional machine learning (50%), deep learning (20.2%) and ensemble learning techniques (29.8%) focusing on different health areas (Table 3). Two health areas had the highest number of studies: cancer and COVID-19 (n = 7, 21.2% each) (Table 2). All the studies on cancer used data from the South African Cancer Registry. Separate studies were conducted by groups with overlapping authors who work in the field, with three studies by Achilonu et al. (2021a, 2021b, 2022) and two studies by Olago et al. (2020, 2023) (Table 3). The purpose of using conventional and ensemble machine learning ranged from classifying risk for length of stay (Achilonu et al. 2021a), generating histopathology reports (Olago et al. 2020) and cancer diagnoses related to human immunodeficiency virus (HIV) diagnoses (Olago et al. 2023) and processing free text to classifying benign versus malignant cancer (Achilonu et al. 2021b, 2022). Deep learning was used to generate malignant urine cytology images (McAlpine et al. 2022), and the prediction of the Gleason Grade group in prostate cancer (Mokoatle et al. 2022).
| TABLE 3: Characteristics of the studies. |
Studies on COVID-19 were also common, because of the recent pandemic (Table 3). Data sets were obtained from the National Institute of Communicable Diseases (NICD) (Lieberman et al. 2023), the National Health Laboratory Service (NHLS) (Stevenson et al. 2021) and Our World in Data (Akinola et al. 2023), a publicly available repository on COVID-19 daily case counts in South Africa. The rest of the studies used either primary data (Okonkwo, Amusa & Twinomurinzi 2022), hospital records (Chimbunde et al. 2023) or data from the Hospital and Emergency Centre Tracking Information System application (HECTIS7), which is the emergency department electronic register in the Western Cape (Fuller et al. 2023). Deep learning and machine learning were used to predict ICU mortality (Chimbunde et al. 2023) and forecast new waves of COVID-19 (Akinola et al. 2023; Stevenson et al. 2021), and conventional and ensemble machine learning were used to confirm vaccination status (Okonkwo et al. 2022), classify hot spots (Lieberman et al. 2023) and predict adverse outcomes (Fuller et al. 2023).
The next most common health area targeted with machine learning was HIV and/or tuberculosis (TB) (n = 8, 24.2%) (Table 2). Six of the eight studies used primary data collected from observational studies (Eken et al. 2020; Majam et al. 2023; Maskew et al. 2022; Onywera et al. 2020; Pahar et al. 2021; Turbé et al. 2021), two studies used hospital records of patients receiving anti-retroviral therapy (ART) treatment (Esra et al. 2023a, 2023b), and one study used data on culture positive isolates confirmed by the National Health Laboratory Services (NHLS) (Achilonu et al. 2021a) (Table 3). The purpose of using machine learning ranged from using deep learning to develop movement tracking systems (Eken et al. 2020), classifications of TB-related versus other coughs, assessing the impact of HIV on Human papillomavirus (HPV) (Onywera et al. 2020) to the classification of rapid HIV tests (Turbé et al. 2021). Conventional and ensemble machine learning were used to classify and predict poor outcomes from HIV such as interruptions to ART treatment (Esra et al. 2023a, 2023b), attendance at clinic visit and viral load (Maskew et al. 2022), HIV risk (Majam et al. 2023), and drug resistance (Sibandze et al. 2020).
The rest of the studies (n = 9, 27.2%) focused on a range of different health focus areas (Table 3). Three studies conducted by the same group of authors, specifically focused on paediatric critical illnesses, using primary data and deep learning to develop a model to predict mortality in paediatric ICUs (Pienaar et al. 2022a), followed by a comparison of different models using conventional and ensemble machine learning (Pienaar et al. 2022b). Thereafter, a study was performed by experts to describe the domain knowledge for a machine learning model for paediatric illness (Pienaar et al. 2023). Deep learning was also used to screen digital otoscopic images (44) and a screening tool for severe acute malnutrition (Nel et al. 2022) (Table 3). Conventional and ensemble machine learning were further used to predict mortality in heart failure (Mpanya et al. 2023), risk of readmission in surgical and trauma emergency departments (Tokac et al. 2025a, 2025b), and to identify risk factors for mortality in laparotomy surgery (Smith et al. 2022), gestational diabetes (Kolozali et al. 2024) and provide malaria warnings (Martineau et al. 2022). In one paper, a combination of deep learning, ensemble and conventional learning was used to develop the most effective model for triage of critically ill children presenting to a tertiary hospital (Pienaar et al., 2022b).
Lastly, a range of AI methods used in the reviewed studies fall into three main types based on their characteristics and applications (Table 1-A2). Conventional machine learning was most often used (22, 66.7%) among the articles, usually combined with ensemble learning (12/22) or deep learning (2/22) or in a combination of all three (4/22) (Table 3). Conventional machine learning relies on traditional statistical and mathematical approaches and is used for tasks such as classification and regression, often in simpler datasets and problems where feature extraction is typically performed manually. Ensemble boosting was used in 18 (54.5%) articles, often in combination with deep or conventional machine learning (14/18) (Table 3). Ensemble learning combines multiple models, such as decision trees, to enhance prediction accuracy, robustness, and generalisability, which are useful in managing diverse and imbalanced datasets. Lastly, deep learning, which is characterised by its reliance on neural networks with multiple layers, designed to process large and complex datasets, was used in 14 articles (42.4), eight times on its own (Table 3).
Discussion
South Africa has a history of disparity in access to technology and education, presenting challenges in adopting the use of artificial intelligence in the health sector. Digital transformation is growing at a slow rate in the private health sector, and more so in public sector, compared to other industries such as banking and insurance (Willie 2020). This review reflects the slow uptake of machine learning, similar to a review on AI use in the health sector in Tanzania (Sukums et al. 2023), but does show some increasing emergence of machine learning to predict different health outcomes in South Africa. The high number of studies on HIV and/or TB is possibly linked to HIV and/or TB being one of the key burden of diseases (De Villiers 2021; Louw et al. 2023) and those studies reported issues of health-seeking behaviour and compliance with treatment regimens (Stangl et al. 2019). Similarly, the recent COVID-19 pandemic, with its widespread impact, the national requirement of reporting COVID-19 cases for surveillance, and the need for rigorous clinical and societal responses (Van Der Schaar et al. 2021), is also reflected in the studies examined in this review.
There has been numerous reports on the importance of integrating machine learning techniques into local and national healthcare systems to improve health response and health outcomes (Van Der Schaar et al. 2021). However, access to big data routinely collected in South Africa is limited (Tokac et al. 2025a, 2025b), and the review demonstrates the challenges of integrating machine learning into data in the health sector. Firstly, similar to another scoping report (Sukums et al. 2023), the high usage of primary data highlights the challenges of limited access to or availability of large volumes of high-quality data for training and validating AI-based models. Having the availability of the national registries of the NICD and NHLS enables researchers to access these large data sets to apply deep learning to classify images and results and apply conventional and ensemble machine learning.
Secondly, there is lack of data analysis skills in the health sector (Ngiam & Khor 2019). This is similar to reports of challenges related to human, infrastructure, and financial resources for the design, development, and implementation of AI-based solutions in the health sector as reported in Tanzania (Sukums et al. 2023). This is also crucial, considering ethical and medico-legal implications, health workers’ understanding of machine learning tools, and data privacy and security (Ngiam & Khor 2019).
Implications and recommendations
Overcoming challenges in South Africa requires access to high quality big data, inclusive policies that promote widespread access to the benefits of using AI in healthcare, and AI literacy in the health sector to understand and address ethical and medico-legal implications.
Strengths and limitations
A strength of this scoping review is its focus on machine learning; however, a limitation is that there is a possibility that important articles might have been missed because of a wide range of machine learning techniques.
Conclusion
In South Africa, there has been an increase in the use of machine learning in the analysis of health data, but access to big data appears to be a challenge, and disparities have had an impact on the adoption of AI technologies in the healthcare sector.
Acknowledgements
The authors would like to acknowledge the initial discussion with Dr Felix Sukums and sharing of the scoping review performed in Tanzania.
Competing interests
The authors declare that they have no financial or personal relationships that may have inappropriately influenced them in writing this article.
Authors’ contributions
J.C. contributed to conceptualisation and project administration; J.C., T.S., and A.C. contributed to methodology and software; J.C., A.B. and T.S. were involved in formal analysis; A.C. and J.C. contributed to writing – original draft preparation; J.C., A.B., A.C. and T.S. were involved in writing – review and editing. All authors have read and agreed to the published version of the article.
Funding information
The research reported in this article was supported by the South African Medical Research Council with funds received from the Department of Science and Innovation. The content and findings reported and/or illustrated are the sole deduction, view, and responsibility of the researcher and do not reflect the official position and sentiments of the South African Medical Research Council (SAMRC) or the Department of Science and Innovation.
Data availability
The authors confirm that the data supporting the findings of this study are available within the article and its references.
Disclaimer
The views and opinions expressed in this article are those of the authors and are the product of professional research. It does not necessarily reflect the official policy or position of any affiliated institution, funder, agency, or that of the publisher. The authors are responsible for this article’s results, findings, and content.
References
Achilonu, O.J., Fabian, J., Bebington, B., Singh, E., Nimako, G., Eijkemans, R. et al., 2021a, ‘Use of machine learning and statistical algorithms to predict hospital length of stay following colorectal cancer resection: A South African pilot study’, Frontiers in Oncology 11, 644045. https://doi.org/10.3389/fonc.2021.644045
Achilonu, O.J., Olago, V., Singh, E., Eijkemans, R.M.J.C., Nimako, G. et al., 2021b, ‘A text mining approach in the classification of free-text cancer pathology reports from the South African National Health Laboratory Services’, Information 12(11), 451. https://doi.org/10.3390/info12110451
Achilonu, O.J., Singh, E., Nimako, G., Eijkemans, R. & Musenge, E., 2022, ‘Rule-based information extraction from free-text pathology reports reveals trends in South African female breast cancer molecular subtypes and Ki67 expression’, BioMed Research International 2022, 6157861. https://doi.org/10.1186/s12982-018-0073-y
Akinola, S.O., Qingguo, W., Olukanmi, P. & Tshilidzi, M., 2023, ‘A boosted evolutionary neural architecture search for timeseries forecasting with application to South African COVID-19 cases’, International Journal of Online Engineering 19(14), 107–130. https://doi.org/10.3991/ijoe.v19i14.41291
Amisha, Malik, P., Pathania, M. & Rathaur, V., 2019, ‘Overview of artificial intelligence in medicine’, Journal of Family Medicine and Primary Care 8(7), 2328. https://doi.org/10.4103/jfmpc.jfmpc_440_19
Aromataris, E., Lockwood, C., Porritt, K., Pilla, B. & Jordan, Z. (eds.), 2024, ‘JBI manual for evidence synthesis’, JBI. https://doi.org/10.46658/JBIMES-24-01
Braun, V. & Clarke, V., 2006, ‘Using thematic analysis in psychology’, Qualitative Research in Psychology 3(2), 77–101. https://doi.org/10.1191/1478088706qp063oa
Chimbunde, E., Sigwadhi, L.N., Tamuzi, J.L., Okango, E.L., Daramola, O., Ngah, V.D. et al., 2023, ‘Machine learning algorithms for predicting determinants of COVID-19 mortality in South Africa’, Frontiers in Artificial Intelligence 6, 1171256. https://doi.org/10.3389/frai.2023.1171256
Covidence, 2023, Covidence systematic review software, Veritas Health Innovation, Melbourne, viewed n.d., from https://support.covidence.org/help/how-can-i-cite-covidence.
De Villiers, K., 2021, ‘Bridging the health inequality gap: An examination of South Africa’s social innovation in health landscape’, Infectious Diseases of Poverty 10, 19. https://doi.org/10.1186/s40249-021-00804-9
Dong, Z., Zhang, N., Li, C., Wang, H., Fang, Y., Wang, J. et al., 2015, ‘Anticancer drug sensitivity prediction in cell lines from baseline gene expression through recursive feature selection’, BMC Cancer 15, 489. https://doi.org/10.1186/s12885-015-1492-6
Eken, M., Meyns, P., Lamberts, R.P. & Langerak, N.G., 2020, ‘Markerless movement tracking using a machine-learning algorithm to assess arm movements during gait in children with HIV encephalopathy’, Gait & Posture 81(suppl 1), 85–86. https://doi.org/10.1016/j.gaitpost.2020.07.074
Esra, R., Carstens, J., Le Roux, S., Mabuto, T., Eisenstein, M., Keiser, O. et al., 2023b, ‘Validation and improvement of a machine learning model to predict interruptions in antiretroviral treatment in South Africa’, Journal of Acquired Immune Deficiency Syndromes 92(1), 42–49. https://doi.org/10.1097/QAI.0000000000003108
Esra, R.T., Carstens, J., Estill, J., Stoch, R., Le Roux, S., Mabuto, T. et al., 2023a, ‘Historical visit attendance as predictor of treatment interruption in South African HIV patients: Extension of a validated machine learning model’, PLoS Global Public Health 3(7), e0002105. https://doi.org/10.1371/journal.pgph.0002105
Fuller, G.W., Hasan, M., Hodkinson, P., McAlpine, D., Goodacre, S., Bath, P.A. et al., 2023, ‘Training and testing of a gradient boosted machine learning model to predict adverse outcome in patients presenting to emergency departments with suspected covid-19 infection in a middle-income setting’, PLoS Digital Health 2, e0000309. https://doi.org/10.1371/journal.pdig.0000309
Kolozali, S., White, S.L., Norris, S., Fasli, M. & Van Heerden, A., 2024, ‘Explainable early prediction of gestational diabetes biomarkers by combining medical background and wearable devices: A pilot study with a cohort group in South Africa’, IEEE Journal of Biomedical and Health Informatics 28(4), 1860–1871. https://doi.org/10.1109/JBHI.2024.3361505
Kuziemsky, C., Maeder, A.J., John, O., Gogia, S.B., Basu, A., Meher, S. et al., 2019, ‘Role of artificial intelligence within the telehealth domain: Official 2019 yearbook contribution by the members of IMIA telehealth working group’, Yearbook of Medical Informatics 28, 35–40. https://doi.org/10.1055/s-0039-1677897
Lieberman, B., Kong, J.D., Gusinow, R., Asgary, A., Bragazzi, N.L., Choma, J. et al., 2023, ‘Big data- and artificial intelligence-based hot-spot analysis of COVID-19: Gauteng, South Africa, as a case study’, BMC Medical Informatics & Decision Making 23, 1–15. https://doi.org/10.1186/s12911-023-02098-3
Louw, Q.A., Conradie, T., Xuma-Soyizwapi, N., Davis-Ferguson, M., White, J., Stols, M. et al., 2023, ‘Rehabilitation capacity in South Africa – A situational analysis’, International Journal of Environmental Research and Public Health 20, 3579. https://doi.org/10.3390/ijerph20043579
Majam, M., Segal, B., Fieggen, J., Smith, E., Hermans, L., Singh, L. et al., 2023, ‘Utility of a machine-guided tool for assessing risk behaviour associated with contracting HIV in three sites in South Africa’, Informatics in Medicine Unlocked 37, 101192. https://doi.org/10.1044/2022_AJA-21-00270
Manlhiot, C., 2018, ‘Machine learning for predictive analytics in medicine: Real opportunity or overblown hype?’, European Heart Journal – Cardiovascular Imaging 19(7), 727–728. https://doi.org/10.1093/ehjci/jey041
Martineau, P., Behera, S.K., Nonaka, M., Jayanthi, R., Ikeda, T., Minakawa, N. et al., 2022, ‘Predicting malaria outbreaks from sea surface temperature variability up to 9 months ahead in Limpopo, South Africa, using machine learning’, Frontiers in Public Health 10, 962377. https://doi.org/10.4081/jphr.2019.1677
Maskew, M., Sharpey-Schafer, K., De Voux, L., Crompton, T., Bor, J., Rennick, M. et al., 2022, ‘Applying machine learning and predictive modeling to retention and viral suppression in South African HIV treatment cohorts’, Scientific Reports 12, 1–10. https://doi.org/10.1038/s41598-022-16062-0
McAlpine, E., Michelow, P., Liebenberg, E. & Celik, T., 2022, ‘Is it real or not? Toward artificial intelligence-based realistic synthetic cytology image generation to augment teaching and quality assurance in pathology’, Journal of the American Society of Cytopathology 11(3), 123–132. https://doi.org/10.1016/j.jasc.2022.02.001
Mokoatle, M., Mapiye, D., Marivate, V., Hayes, V.M. & Bornman, R., 2022, ‘Discriminatory gleason grade group signatures of prostate cancer: An application of machine learning methods’, PLoS One 17, e0267714. https://doi.org/10.1016/j.asoc.2021.107161
Mpanya, D., Celik, T., Klug, E. & Ntsinjana, H., 2023, ‘Predicting in-hospital all-cause mortality in heart failure using machine learning’, Frontiers in Cardiovascular Medicine 9, 1032524. https://doi.org/10.3389/fcvm.2022.1032524
Nel, S., Feucht, U.D., Nel, A.L., Becker, P.J. & Wenhold, F.A.M., 2022, ‘A novel screening tool to predict severe acute malnutrition through automated monitoring of weight-for-age growth curves’, Maternal & Child Nutrition 18, 1–10. https://doi.org/10.1111/mcn.13364
Ngiam, K.Y. & Khor, I.W., 2019, ‘Big data and machine learning algorithms for health-care delivery’, Lancet Oncology 20(5), e262–e273. https://doi.org/10.1016/S1470-2045(19)30149-4
Okeibunor, J.C., Jaca, A., Iwu-Jaja, C.J., Idemili-Aronu, N., Ba, H., Zantsi, Z.P. et al., 2023, ‘The use of artificial intelligence for delivery of essential health services across WHO regions: A scoping review’, Frontiers in Public Health 11, 1102185. https://doi.org/10.3389/fpubh.2023.1102185
Okonkwo, C.W., Amusa, L.B. & Twinomurinzi, H., 2022, ‘COVID-Bot, an intelligent system for COVID-19 vaccination screening: Design and development’, JMIR Formative Research 6, e39157. https://doi.org/10.1038/s41598-020-58290-2
Olago, V., Muchengeti, M., Singh, E. & Chen, W.C., 2020, ‘Identification of malignancies from free-text histopathology reports using a multi-model supervised machine learning approach’, Information 11(9), 455. https://doi.org/10.3390/info11090455
Olago, V., Nimako, G., Bartels, L., Bohlius, J., Dhokotera, T., Egger, M. et al., 2023, ‘Cancer diagnostic service use in people with HIV in South Africa: A cross-sectional study’, PLoS One 19(6), e0291897. https://doi.org/10.1371/journal.pone.0291897
Onywera, H., Williamson, A.-L., Cozzuto, L., Bonnin, S., Mbulawa, Z.Z.A., Coetzee, D. et al., 2020, ‘The penile microbiota of Black South African men: Relationship with human papillomavirus and HIV infection’, BMC Microbiology 20, 1–18. https://doi.org/10.1186/s12866-020-01759-x
Pahar, M., Klopper, M., Reeve, B., Warren, R., Theron, G. & Niesler, T., 2021, ‘Automatic cough classification for tuberculosis screening in a real-world environment’, Physiological Measurement 42(10), 105014. https://doi.org/10.1016/j.procs.2021.01.281
Peters, M.D.J., Godfrey, C., McInerney, P., Munn, Z., Tricco, A.C. & Khalil, H., 2024, ‘Scoping reviews (2020)’, in E. Aromataris, C. Lockwood, K. Porritt, B. Pilla & Z. Jordan (eds.), JBI Manual for Evidence Synthesis, pp. 417–476, JBI. https://doi.org/10.46658/JBIMES-24-09
Pienaar, M.A., Sempa, J.B., Luwes, N., George, E.C. & Brown, S.C., 2023, ‘Elicitation of domain knowledge for a machine learning model for paediatric critical illness in South Africa’, Frontiers in Pediatrics 11, 1005579. https://doi.org/10.3389/fped.2023.1005579
Pienaar, M.A., Sempa, J.B., Luwes, N., George, E.C. & Brown, S.C., 2022a, ‘Development of artificial neural network models for paediatric critical illness in South Africa’, Frontiers in Pediatrics 10, 1008840. https://doi.org/10.1016/j.isci.2021.103298
Pienaar, M.A., Sempa, J.B., Luwes, N. & Solomon, L.J., 2022b, ‘An artificial neural network model for pediatric mortality prediction in two tertiary pediatric intensive care units in South Africa. A development study’, Frontiers in Pediatrics 10, 797080. https://doi.org/10.1136/bmjopen-2021-049786
Sandström, J., Myburgh, H., Laurent, C., Swanepoel, W. & Lundberg, T., 2022, ‘A machine learning approach to screen for Otitis Media using digital otoscope images labelled by an expert panel’, Diagnostics (Basel) 12. https://doi.org/10.1186/s12936-022-04146-1
Shandhi, M.M.H. & Dunn, J.P., 2022, ‘AI in medicine: Where are we now and where are we going?’, Cell Reports Medicine 3, 100861. https://doi.org/10.1016/j.xcrm.2022.100861
Sharifani, K. & Amini, M., 2023, ‘Machine learning and deep learning: A review of methods and applications’, World Information Technology and Engineering Journal 10, 3897–3904.
Sibandze, D.B., Magazi, B.T., Malinga, L.A., Maningi, N.E., Shey, B.-A., Pasipanodya, J.G. et al., 2020, ‘Machine learning reveals that Mycobacterium tuberculosis genotypes and anatomic disease site impacts drug resistance and disease transmission among patients with proven extra-pulmonary tuberculosis’, BMC Infectious Diseases 20, 1–15. https://doi.org/10.1186/s12879-020-05256-4
Smith, M.T.D., Bruce, J.L. & Clarke, D.L., 2022, ‘Using machine learning to establish predictors of mortality in patients undergoing laparotomy for emergency general surgical conditions’, World Journal of Surgery 46, 339–346. https://doi.org/10.1007/s00268-021-06360-5
Stangl, A.L., Earnshaw, V.A., Logie, C.H., Van Brakel, W., C. Simbayi, L., Barré, I. et al., 2019, ‘The health stigma and discrimination framework: A global, crosscutting framework to inform research, intervention development, and policy on health-related stigmas’, BMC Medicine 17, 31. https://doi.org/10.1186/s12916-019-1271-3
Stevenson, F., Hayasi, K., Bragazzi, N.L., Kong, J.D., Asgary, A., Lieberman, B. et al., 2021, ‘Development of an early alert system for an additional wave of COVID-19 cases using a recurrent neural network with long short-term memory’, International Journal of Environmental Research and Public Health 18. https://doi.org/10.3390/s22103620
Sukums, F., Mzurikwao, D., Sabas, D., Chaula, R., Mbuke, J., Kabika, T. et al., 2023, ‘The use of artificial intelligence-based innovations in the health sector in Tanzania: A scoping review’, Health Policy and Technology 12(1), 100728. https://doi.org/10.1016/j.hlpt.2023.100728
Tokac, U., Chipps, J., Brysiewicz, P., Bruce, J. & Clarke, D., 2025a, ‘Using machine learning to improve readmission risk in surgical patients in South Africa’, International Journal of Environmental Research and Public Health 22(3), 345. https://doi.org/10.3390/ijerph22030345
Tokac, U., Chipps, J., Brysiewicz, P., Bruce, J. & Clarke, D., 2025b, ‘Using natural language processing in the LACE Index Scoring Tool to predict unplanned trauma and surgical readmissions in South Africa’, World Journal of Surgery 49(4), 1067–1073. https://doi.org/10.1002/wjs.12523
Tricco, A.C., Lillie, E., Zarin, W., O’Brien, K.K., Colquhoun, H., Levac, D. et al., 2018, ‘PRISMA Extension for Scoping Reviews (PRISMA-ScR): Checklist and explanation’, Annals of Internal Medicine 169(7), 467–473. https://doi.org/10.7326/M18-0850
Turbé, V., Herbst, C., Mngomezulu, T., Meshkinfamfard, S., Dlamini, N., Mhlongo, T. et al., 2021, ‘Deep learning of HIV field-based rapid tests’, Nature Medicine 27, 1165–1170. https://doi.org/10.1038/s41591-021-01384-9
Van Der Schaar, M., Alaa, A.M., Floto, A., Gimson, A., Scholtes, S., Wood, A. et al., 2021, ‘How artificial intelligence and machine learning can help healthcare systems respond to COVID-19’, Machine Learning 110, 1–14. https://doi.org/10.1007/s10994-020-05928-x
Van Zyl-Cillié, M.M., Bührmann, J.H., Blignaut, A.J., Demirtas, D. & Coetzee, S.K., 2024, ‘A machine learning model to predict the risk factors causing feelings of burnout and emotional exhaustion amongst nursing staff in South Africa’, BMC Health Services Research 24(1), 1665–1620. https://doi.org/10.1186/s12913-024-12184-5
Willie, M.M., 2020, ‘Digital transformation in healthcare – South African Context’, World Medical Journal 66, 34–38.
Appendices
Appendix 1
Appendix 2
| TABLE 1-A2: Detailed categorisation of machine learning methods by type. |
|
Email this article 
Hide
Show all
