Many organizations are already adopting AI for healthcare in Africa. The following are a few examples of where AI is being used in the continent:
• minoHealth AI Labs in Ghana is automating radiology by applying deep learning and an algorithm known as a convolutional neural network.
• Philips Foundation has successfully implemented AI software, developed by Delft Imaging, in 11 South African hospitals to help triage and monitor COVID-19 patients via X-ray imaging. Delft Imaging’s AI-based CAD4COVID software, which complements existing COVID-19 diagnostic technologies, estimates the severity and progression of COVID-19 disease based on routinely available chest X-rays.
• South Africa is also currently applying a multinomial logistic classifier-based method to individual resource scheduling, particularly in predicting the duration health employees may stay within public service [3].
• In Tanzania and Zambia, Delft Institute’s CAD4TB software has been used to assess the utilization of the computer-aided analysis of pulmonary tuberculosis from the chest radiographs.
• Ilara Health is also offering accurate and affordable diagnostics to people in rural areas via small, AI-powered diagnostic devices incorporated through a proprietary technology policy and correspondingly distributed openly to the primary care doctors.
• Antara Health is using AI-assisted health technology to make healthcare simple for patients and providers.
• XELPHA Health operating Aphya as the sole mobile-first EMR solution that assists in the detection and optimization of specific devices and hence facilitating active contribution and engagement amidst both patients and providers.
Essential building blocks and barriers for a sustainable AI in Africa
To leverage the opportunities for AI in healthcare in Africa, there is a need to address the main building blocks that are essential to delivering a sustainable AI for African healthcare systems. For example, there needs to be:
• Proper digital infrastructure for storage of data from health facilities.
• A strong data culture within health facilities that values data and makes tools and resources accessible to clinicians.
• Suitable regulations and standards for AI and data science, which will enable regulators to examine AI applications within health before their deployment.
• Adoption of local solutions by comprehending and finding suitable solutions while promoting self-reliance and assisting in cultivating the local ecosystem.
• More targeted funding for AI health start-ups in Africa that links entrepreneurs with corresponding financiers and reduces the risk for private investors.
Some of the AI barriers that need to be overcome are associated with digital infrastructure, data culture, regulations and standards, adoption of local solutions and funding. Lack of capacity among healthcare professionals remains a challenge. Training on and adoption of AI is needed to ensure compliance, especially with regulations [4]. Besides the difficulty in acceptance by patients, clinical adoption is still a major barrier, as many healthcare professionals are still not completely comfortable using AI technologies. However, much can be achieved through capacity building.
What to expect and not to expect from AI in healthcare
The emergence of AI in healthcare continues to spur mixed reactions from experts in terms of what to expect and what not to expect.
But what can we not expect from AI in healthcare? There are a lot of conversations around what AI cannot bring to healthcare. While AI has a great potential for healthcare in Africa, particularly discovering unusual associations and forecasting public health crises, it is not likely to replace the human factor. AI can handle repeated tasks based on data, but will not provide the compassion and empathy that patients need during treatments. Elements of human reasoning are also missing in AI. One of the biggest questions since the emergence of AI in healthcare is whether AI will replace medical professionals and how it will affect medical specialties. Well, certain repeated medical procedures will definitely be taken over by AI. However, physicians will not be replaced in their entirety. We cannot rely on AI for mental health, for example. Some of the reasons include that it is never tech vs. human (these two complement one another) and AI cannot make medical decisions alone, as it is fully dependent on data.
References:
[1] Alhashmi, S.F., Alshurideh, M., Al Kurdi, B. and Salloum, S.A., 2020, April. A systematic review of the factors affecting the artificial intelligence implementation in the health care sector. In Joint European-US Workshop on Applications of Invariance in Computer Vision (pp. 37-49). Springer, Cham.
[2] Alhashmi, S.F., Alshurideh, M., Al Kurdi, B. and Salloum, S.A., 2020, April. A systematic review of the factors affecting the artificial intelligence implementation in the health care sector. In Joint European-US Workshop on Applications of Invariance in Computer Vision (pp. 37-49). Springer, Cham.
[3] Alhashmi, S.F., Alshurideh, M., Al Kurdi, B. and Salloum, S.A., 2020, April. A systematic review of the factors affecting the artificial intelligence implementation in the health care sector. In Joint European-US Workshop on Applications of Invariance in Computer Vision (pp. 37-49). Springer, Cham.
[4] Guo, J. and Li, B., 2018. The application of medical artificial intelligence technology in rural areas of developing countries. Health equity, 2(1), pp.174-181.