Kenya should adopt AI to boost drug safety monitoring

Recent advancements in Artificial Intelligence (AI) have thrust humanity into what is being hailed as the most consequential era in technological development since the Industrial Revolution.

AI applications not only increase efficiency and productivity, but they have also been shown to improve decision-making, as well as result in significant cost savings.  Today, entrepreneurs in many sectors around the globe, from agriculture to education, and even engineering, are grappling to integrate AI into operational frameworks in order to take advantage of the efficiencies afforded by the revolutionary technology.

In the healthcare sector, the pharmaceutical industry is expected to especially benefit from the advances in AI-based technologies. AI not only accelerates the identification of candidates for drug development, but it can also be used to predict the efficacy and toxicity of experimental medicines, streamline clinical trials, and enhance drug manufacturing through automation.

Significantly, AI also has applications in pharmacovigilance (PV). PV is the science that deals with the detection, assessment, and prevention of the adverse effects of health products and technologies (HPTs), with the main goal of protecting patients and consumers.

Some of the main activities of PV include collection of data on adverse events from patients/consumers; analysing the data to identify previously unknown safety concerns (an exercise known as signal detection); evaluating the benefits and risks of products based on all available data (risk assessment); developing and implementing strategies to minimise potential risks associated a product (risk management); and informing stakeholders, including healthcare professionals, regulators, and the public, about drug safety issues and risk minimisation measures.

AI can be applied to most of these PV activities, taking advantage of the benefits of the ground-breaking technology.

For example, AI-based machine learning applications can be used to automate safety data collection and signal detection, as they can efficiently and accurately analyse vast amounts of data from multiple sources, such as electronic health records and social media. AI can be integrated into safety surveillance systems to monitor adverse events in real time, enabling regulators and manufacturers to respond swiftly.

It is notable that in many low-income nations, including Kenya, drug safety monitoring and reporting remains a big challenge, with some scientific reports showing a PV reporting rate of only about 6 per cent in these areas.

Weak regulatory structures, poor funding for pharmacovigilance activities, limited public awareness of drug safety surveillance, and limitations in the availability of competent PV analysts have been cited as some of the main reasons for low reporting. Luckily, some of these constraints can be addressed through AI-based innovations.

Cheap AI applications can be applied to collect and analyse large PV datasets, negating the need for expensive data collection and analysis systems and personnel. Moreover, personalised chatbots can be used to summarise and communicate safety data to patients in a simplified manner.

Automated AI systems can also be used to quickly and accurately process and compile adverse event reports, further enhancing financial feasibility in the conduct of PV activities.

From the foregoing, it would be imprudent for any actor in the PV industry to ignore the tremendous potential of AI in reducing surveillance costs and enhancing the quality and coverage of safety reporting.

The Health Ministry and the Pharmacy and Poisons Board should move with speed to develop necessary frameworks to facilitate the deployment of AI applications in the surveillance of HPTs.
Dr Kivuva is a pharmacovigilance specialist