Navigating the Cancer Drug Shortage: Building Supply Chain Resilience with AI

The White House recently announced actions to strengthen American supply chains, introducing a new Council on Supply Chain Resilience. As part of these efforts, the White House will be dedicating resources toward bolstering the pharmaceutical supply chain in particular.

While it’s too soon to know exactly what types of actions this newly-formed council will take, their attention towards the pharmaceutical supply chain is much needed. Take, for example, the cancer drug shortage that has been acutely impacting the treatment of many American patients this year. In 2023, 90% of hospitals in the US experienced shortages of the 15 chemotherapy drugs most commonly used in cancer treatments. Cancer drug shortages have resulted in doctors and patients having to make impossible decisions that result in potentially delaying or foregoing lifesaving care. This is clearly unacceptable.

These cancer drug shortages have been caused in large part by supply chain disruptions. In November of 2022, FDA inspectors found that a major cancer drug manufacturing plant – responsible for half of the US supply of a widely used generic chemotherapy drug called cisplatin – had major safety issues. This episode led the plant to stop production, and unfortunately, other manufacturers have failed to pick up the slack to meet global demand. 

The circumstances around the cancer drug shortage beg a question: how could just one manufacturing plant’s safety concerns lead to such drastic ripple effects throughout the supply chain? The answer is simple: our supply chains lack resilience. 

Specifically, supply chains that rely upon single points of failure and lack meaningful redundancy are vulnerable to shocks that can result in sudden shortages, price shocks and can even set off panic-buying cycles for critical supplies such as food, fuel and medicine. That vulnerability means that anything – a war, a storm, a factory shutdown – can quickly cascade into a much larger economic impact.

In recent years, many organizations have prioritized making their procurement operations faster and more efficient, which often means consolidating demand around fewer and fewer suppliers. While there have been cost efficiencies returned to the business, the practical impact of these decisions in aggregate across supply networks has been an increase in single points of failure, brittleness, and susceptibility to shocks. This is especially true in the context of JIT (Just In Time) supply chains, where small delays can result in a disruption to a production schedule. 

Up until now, supply chain and procurement professionals have been forced to choose between resilience and efficiency, with efficiency often winning out. But does it have to be that way? What if the forced choice between resiliency and efficiency, between multi-sourcing and economies of scale was a false choice? 

One new approach rests within novel predictive models and AI technology that power decision systems and forecasts for actors across the supply chain. Predictive AI technology, for example, can be employed to assess and predict potential bottlenecks in the supply chain based on historical data, geopolitical factors, weather patterns, and other variables. Predictive models can also uncover the ways that these events can impact potential single points of failure, and highlight vulnerabilities in the supply chain. Additionally, novel generative AI approaches can perform search queries across huge sets of unstructured data, including supplier contracts, price agreements, and historical spend to unearth trends in lead time, availability and unit price.

Predictive AI can also help build resiliency through geographic segmentation and diversification, as well as route optimization. Consider that when disruptions occur, not every location will feel the effects in the same way at the same time, nor will goods and cargo pass through disrupted zones in the same way depending on carrier and ports of entry. 

AI can help pharmaceutical supply chain teams diversify their supply base across a range of geographies, taking into account how quickly each location typically reacts to macroeconomic trends, and how likely they are to react to single points of failure. As a result, even if unexpected issues do arise outside of what a supply chain or procurement team has forecasted, they will have a much better chance of securing the inventory they need. In the case of the cancer drug shortage, diversifying across a wider array of manufacturers could prevent future disruptions, as the US would not be relying so heavily on the production from a single now-closed Indian manufacturing plant.

In addition, AI can do all of this while also helping procurement teams move quickly and with better utilization of existing resources – so there’s no sacrificing resilience for the sake of efficiency, and there’s no need to make the knee-jerk response to a shortage fear to be an increase in safety stocks, which in fact may only exacerbate the shortage by increasing global demand in the short term. 

Finding ways to increase the resilience of the pharmaceutical supply chain will not only have impacts in the short-term – helping patients get access to the urgent treatments they need right now – but in the long-term, too, as we look down the barrel of inevitable future disruptions. By leaning into technology like AI, we can effectively address the complex challenges facing the pharmaceutical supply chain and better navigate critically urgent shortages.