By Dave DeFusco
When the COVID-19 pandemic struck, the world celebrated the arrival of vaccines. But getting those vaccines from factories into people’s arms turned out to be much harder than expected. Vaccines need to be kept at carefully controlled temperatures from the moment they are made until the moment they are given to patients. Any break in this “cold chain” makes doses useless. During the pandemic, those breakdowns, combined with poor planning and uneven access, meant that up to 30 percent of vaccines worldwide were wasted. That’s millions of lost doses at a time when every shot mattered.
This global problem caught the attention of Yeukai Gilmore Magara, a student in the Katz School’s M.S. in Biotechnology Management and Entrepreneurship and a student consultant with DREAM Venture Labs. For one of his research projects, he studied how artificial intelligence can help create smarter, fairer and more resilient vaccine supply chains. His report, “Optimizing Pandemic Vaccine Supply Chains: AI-Driven Strategies & Future Directions,” examines how AI could drastically reduce waste and make vaccines more accessible in future health crises.
“The pandemic exposed just how fragile our supply chains are,” said Magara. “AI tools could enable real-time responses, help predict where doses are needed, keep them safe during transport and improve fairness in access so that everyone, no matter where they live, has a fair chance at getting vaccinated.
Magara’s review of more than 100 studies identified five major weaknesses in how vaccines were distributed during the pandemic:
- Bad forecasting: Planners often relied on outdated demand data. This led to oversupply in some areas and shortages in others.
- Centralized systems: Most countries used large, central warehouses. This slowed down deliveries to rural areas.
- Inequity: Only about 10 percent of models even considered fairness in distribution, leaving vulnerable populations behind.
- Cold chain failures: Lapses in refrigeration ruined millions of doses, especially in low-resource settings.
- Lack of monitoring: Fewer than 20 percent of models used real-time tools, like sensors, to track shipments.
The result was that Canada discarded 2.3 million doses, Africa lost more than 15 million and worldwide wastage reached nearly a third of all vaccines produced. According to Magara’s study, AI-driven tools could make a big difference. Among the most promising strategies include:
- Better forecasting: By combining infection data with AI models, companies like Pfizer and Moderna improved predictions by up to 25 percent.
- Decentralized storage hubs: Instead of relying on one central warehouse, smaller local hubs—and even drones like those used by Zipline in Africa—cut delivery times almost in half.
- Real-time monitoring: Internet-connected sensors track vaccines’ temperature and location, reducing cold chain losses by 15 to 20 percent.
- Fairness-based allocation: AI can help prioritize underserved regions, boosting access by approximately 25 percent.
- Hybrid approaches: Some companies, like Johnson & Johnson, combine AI with advanced math-based planning to make supply chains more flexible.
If adopted widely, these methods could cut vaccine waste from 30 percent to 6 percent, raise on-time deliveries to nearly 99 percent and save more than a million dollars annually, while getting lifesaving doses to more people faster.
Looking ahead, Magara believes vaccine supply chains need to be both high-tech and human-centered. AI can help predict outbreaks, optimize deliveries and minimize waste, but fairness must remain at the heart of any system.
“A vaccine that doesn’t reach the people who need it most is a failure, no matter how advanced the logistics look on paper,” he said. “The next pandemic will come—it’s a matter of when, not if. We need to build supply chains that are flexible, equitable and resilient enough to handle it.”
