Application of deep learning and machine learning models to improve healthcare in sub-Saharan Africa: Emerging opportunities, trends and implications

Abstract

Deep learning and machine learning techniques present unmatched opportunities to improve healthcare in sub-Saharan Africa (SSA). However, there is a paucity of literature on AI-based applications deployed to improve care in SSA, which makes it challenging to organise the research contributions in the present and to highlight obstacles and emerging research areas that need to be explored in the future. This study applied the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analysis) model to conduct a comprehensive review of deep learning and machine learning models deployed in SSA to improve access to care while exploring emerging opportunities, trends and implications for integrating AI-based models in SSA healthcare. This study reveals that AI models can analyse and derive inferences from massive health data for early detection, diagnosis, monitoring for chronic disorders, prediction of diseases, monitoring large-scale public health patterns and help limit exposure in contagious environments. AI can facilitate the development of targeted health interventions and improve patient outcomes in all stages of diagnosis, treatment, drug development and monitoring, personalised medicine, patient control and care. Integrating AI models with health applications can tremendously assist health professionals and policymakers in disease diagnosis and making informed decisions. AI algorithms bias, poor access to health data and formats, and lack of policies and frameworks supporting the integration of data-driven AI-based solutions into health systems hinder the integration of AI-based models into health systems. There is a need for transparency and ethical use of AI and crafting policies that support the use of AI in SSA health systems. Utilising AI-based models in healthcare can also assist researchers and healthcare workers to move towards smart care and better comprehend future research needs of AI in smart care.