Artificial Intelligence (AI) is rapidly reshaping the healthcare landscape, offering transformative solutions across a broad spectrum of applications. The integration of AI within healthcare is primarily aimed at analysing clinical data to enhance patient outcomes and tailor medical treatments to individual needs. In practice, AI assists in improving diagnostic accuracy, expediting drug discovery, and providing personalised patient care. It is also leveraged to streamline administrative processes, thus reducing costs, and to support broader public health objectives, including training healthcare professionals and enabling remote medical services.

AI’s capability to process and analyse vast amounts of data is playing a critical role in predicting patient health trajectories, recommending treatment plans, guiding surgeries, and monitoring patients’ well-being. These tools are particularly promising in managing population health by identifying and addressing community-wide health challenges.

The potential of AI in healthcare is vast, especially in developing countries where resources can be scarce and healthcare infrastructure may not be as robust as in developed nations. A recent study published by Nature Medicine on November 20, 2023, exemplify how AI can significantly impact medical diagnostics. The study discusses Pancreatic Ductal Adenocarcinoma (PDAC), which is often detected at a late and typically inoperable stage. The development of a deep learning tool named Pancreatic Cancer Detection with Artificial Intelligence (PANDA), which can accurately detect and classify pancreatic lesions through non-contrast CT scans, is a breakthrough. This tool could be particularly beneficial for hospitals in developing countries. It can enable hospitals to upload the reports of the diagnostic tests that can be reviewed remotely, aiding doctors in government and rural hospitals where access to specialised radiologists may be limited.

Such AI tools can help in early detection, which is crucial for diseases like PDAC. By providing high accuracy in diagnosis, AI like PANDA can reduce the workload on overburdened healthcare systems, minimise the rate of false positives, and allow for large-scale screening that was previously not feasible.

Moreover, the application of AI in healthcare can extend beyond diagnostics to treatment planning, patient monitoring, and even in predicting patient outcomes based on historical data. This leads to more personalised care and efficient use of limited resources. In the context of developing countries, the ability to provide high-quality healthcare with AI assistance can lead to improved health outcomes and a better quality of life for patients.

AI algorithms expedite image acquisition in Magnetic Resonance Imaging (MRI), optimising department productivity and improving patient experiences by reducing exam times while maintaining high-resolution imaging quality, regardless of patient movement or other conditions.

The application of AI in ultrasound measurements has been instrumental in cardiac care, where AI-enabled automatic measurements provide fast and reliable echo quantification, allowing healthcare professionals to concentrate on complex diagnostic and treatment tasks. AI assists radiologists by acting as an auxiliary tool, highlighting areas of potential concern on images that may be overlooked, thereby enhancing diagnostic accuracy and efficiency. Moreover, AI fosters multidisciplinary collaboration in cancer care by integrating diverse clinical data, thereby improving decision-making processes for timely and informed treatment strategies.

Key challenges

In the context of developing countries, key challenges include the scarcity of medical imaging resources and the need for creative application of AI to address healthcare disparities. Initiatives like those by Penn Medicine and RAD-AID International focus on bridging these gaps by fostering global health collaboration, with AI at the forefront of improving radiology services in resource-limited settings.

Throughout Africa, AI is being deployed to tackle healthcare challenges innovatively. In Kenya, AI advancements are crucial for improving health worker-patient communications and aiding in the detection of ocular conditions and serious health threats.technologies like PapsAI are transforming the diagnosis and classification of cervical cancer from pap smears, exemplifying the potential of AI in oncology. South African healthcare facilities utilise AI for human resource management and optimising the schedules of community health workers. Nigerian innovation is prominent with start-ups such as Ubenwa, which employs AI for the diagnosis of birth asphyxia through infant cry analysis, showcasing AI’s role in addressing critical neonatal health issues. In Zambia, AI applications are evident with tools like the Dawa Health Clinic App, enhancing healthcare delivery and diagnostics, pointing to the utility of AI in improving healthcare outcomes.

The successful deployment of AI in these environments hinges on the effective distribution of AI technologies, whether through hardware, desktop applications, cloud services, or Picture Archiving Communication Systems (PACS), each with its own unique set of challenges that must be met with flexibility and collaboration.

Moreover, the deployment of AI in healthcare has raised significant ethical concerns, particularly regarding biases and inequalities. A notable study in this area is Obermeyer et al’s research published in Science, which found that an AI system used in US hospitals to allocate healthcare resources was biased against black patients. This algorithm, intended to help healthcare providers identify patients who would benefit from additional care, inadvertently favoured white patients over black patients, not because of explicit racial programming but because it used healthcare costs as a proxy for health needs. This situation highlights the risks of unintentional bias in AI systems, which can perpetuate existing disparities.

To transcend these ethical concerns, it’s crucial to incorporate robust ethical frameworks in the development and deployment of AI in healthcare. This involves a multidisciplinary approach, engaging ethicists, technologists, healthcare professionals, and diverse patient groups in the design and evaluation process.

The writer is OSD, Research, EAC-PM. Views are personal