Can artificial intelligence (AI) predict a drug shortage before it starts?

This was one of the themes explored by a panel of experts in a critical discussion at the recent SAPICS Conference, Africa’s leading event for the supply chain profession. Panellists Matt Christian, Daniel Rosen and Glaudina Loots shared their insights with the 750 supply chain managers from 30 countries who attended this annual gathering in Cape Town.

Christian is the director of Supply Chain Insights at United States Pharmacopeia (USP), which he explained is a non-profit that develops quality resources and standards for medicines, dietary supplements and food ingredients. Rosen is the managing director of new business at Maisha Meds, a digital health organisation whose software and supply chain solutions improve care in private pharmacies and clinics across Africa. Loots is the Health Innovation director at the Department of Science and Innovation. The topic of this important panel discussion was “The Promise of AI for Predicting Drug Shortages and Designing Solutions to Mitigate Supply Chain Risk”.

Drug shortages pose a major threat to public health, and USP’s recently released inaugural Annual Drug Shortages Report reveals that the situation is getting worse. This report notes that the number of year-end drug shortages in the United States has increased over the past decade and that there is a concerning trend of longer-lasting shortages.

The global medicine supply chain is a complex marketplace of manufacturers, suppliers, and distributors from many countries, which makes it susceptible to disruptions that can impact getting critical medicines into the hands of those who need them.​ A key to addressing the challenges and saving lives lies with artificial intelligence (AI). AI can analyse huge volumes of data faster than any human analyst could, Rosen told SAPICS delegates. “It can go upstream and look at all the flows through the supply chain. It can predict irregularities by analysing drug use and overuse. AI can quickly predict if there is going to be a drug shortage, allowing other role players in the system to act quickly to avert a crisis.”

Christian explained that USP’s Medicine Supply Map is a graph-based predictive data model that generates insights into the upstream medicine supply chain. Using machine learning and advanced analytics, the Medicine Supply Map identifies, characterises and quantifies factors linked to supply chain disruptions for drug ingredients and finished drug products. It analyses over 250 million data points and more than 24 data sources from USP, regulatory agencies and the private sector, to develop a more complete and accurate view of potential supply chain disruptors.

The panellists discussed artemisinin-based malaria medicine as an example of a drug that has been impacted by shortages and other challenges over the years, and which underscores the complexity and volatility of global medicine supply chains. Made from Artemisia annua, a weed native to China, artemisinin has long been one of our most effective tools in the fight against malaria. However, securing a stable supply was a big challenge. “Farmers entered and left the market, depending on the price and demand at a particular time. High prices led to an oversupply and then the price crashed, leading to shortages when farmers left the market for more lucrative crops,” Christian explained. “AI could mitigate these risks, with all the data and visibility to inform on volumes and prices, and to enable an early warning system for issues, including shortages.”

Today, as Loots noted, artemisinin-based therapy is losing some effectiveness as the malaria parasite mutates and develops partial resistance. Rosen stressed that in this changed landscape, AI could also play a vital role by improving the base of knowledge on drug resistance.

While AI has the potential to predict and reduce drug shortages and ensure a more reliable and resilient pharmaceutical supply chain, Christian cautioned that the data it delivers must be shared responsibly. “There is a burden on data scientists to not just publish information which could cause panic. We need a global entity tasked with managing this,” he stated.

This year’s 46th SAPICS Conference was hosted by The Professional Body for Supply Chain Management (SAPICS) in association with Southern African Association of Freight Forwarders (SAAFF). Recognising the importance of efficient healthcare supply chains that lives depend on, the event featured a global public health supply chain track that ran throughout the conference programme, with exceptional local and international experts who shared their insights on healthcare supply chain issues.