Data on supply chain organization and performance should be considered a key component of public health data as we continue to learn from the pandemic and plan for future public health crises.
Dr. Tinglong Dai, a professor of Operations Management and Business Analytics at the Carey School of Business, has dedicated his career to investigating the interplay between supply chains and healthcare. He says the public should not be pessimistic about the U.S. response to COVID-19 given the incredible advances with vaccine development and an unprecedented level of global data collection.
Supply chain management is about matching supply with demand. When demand for vaccination exceeded supply a few months ago, the focus was on generating the most vaccinations from a limited supply of vaccines. Now there is a surplus of vaccines, which comes with a problem every business faces — sustaining and increasing demand, also known as marketing. It’s probably fair to say no business disciplines are more data-driven than marketing. To start with, we can use data to understand what works and what doesn't to reduce friction in the vaccination process.
Data really plays a central role in evaluating effectiveness of different vaccine outreach programs, such as lotteries, as well as ensuring transparency and promoting trust. Oftentimes, people are hesitant not because of their inherent anti-vaccine beliefs, but because of stories they read on social media that don’t come with context. Credible sources are rarely available at a single click, leaving people confused and helpless. I also found certain people fear they risk being labeled as anti-vaccine by sharing a link or raising a question. We need to make sure that credible vaccine data and expert resources are as accessible as the vaccines themselves to help more people make the crucial decision to get vaccinated.
The supply chain consists of flow and stock. Flow concerns how fast things are being consumed where they are, and stock is how much inventory exists at a particular moment. We need to pay as much attention to flow data as we do to stock data. People tend to worry about supply chains when they see empty shelves (i.e. insufficient stock), and by then it is too late to fix supply chain issues. Reliable and timely forecasts of vaccine flow are essential to provide a smooth vaccination effort. We should pay attention not only to inventory level, but also to trends and how supply evolves over time.
In the United States we were not only tracking the number of doses distributed from the CDC, but also how many doses were actually given to people. That complete flow data is what made us more successful in vaccine distribution than with testing kits and personal protective equipment (PPE). We knew we had a lot of test kits and PPE getting distributed, but we had no idea where they were and how many units were being used.
Our primary target should be performance data because people are inevitably driven by performance metrics. The reason Israel was so successful with vaccinations was their motivation to become number one, and they were for quite a while. We should continue to collect, analyze, and rank vaccine supply, distribution, and administration performance across jurisdictions to inform and motivate governments to continue and expand vaccination efforts.
We also must use data to highlight the correlation between vaccination progress and infection control. To sustain and even boost interest in vaccination, we need to demonstrate to the public that the vaccines work. We should be able to do that in a granular way involving demographic data so that people really see the vaccine does work and will work for people just like them.
The beauty of vaccination data is that it is cleaner and more objective compared to testing and other pandemic-related data because it is binary — you either get vaccinated or you don’t. There will be different brands, doses, and timing, but those are minor differences. Compare that with testing. Different countries, regions, and providers have very different diagnostic standards and reporting criteria. In that sense, vaccination data can be a powerful way to inform us about global vaccine distribution strategies and help different countries learn from each other because it is so easily comparable. For example, people are surprised to learn Canada is ahead of the United States, United Kingdom, and Israel in first-dose vaccinations, and we should use that surprise to motivate us to learn from their success and inform our own vaccination efforts.
I do believe AI can help identify and reduce disparities and improve efficiency in healthcare, but only if we have the right kind of data and performance metrics. The possibilities are limitless. We could have AI-based demand forecasting to minimize waste and shortages of healthcare materials. We can also use AI to identify and prioritize critical populations to maximize effectiveness of limited vaccines. AI could help correlate vaccine progress and infection control.
AI is utilized for similar activities in the private sector. We already have most of the data and technology we need to transition AI efforts to healthcare. We have high-quality demographics data, mobility data (travel, distance, frequency), and health data about infection and mortality, but we need to establish connections between different databases so that they can talk to each other.
If you rank the vaccination rates of the states for this pandemic and compare the result with the next most recent pandemic, H1N1, the top and bottom states are basically the same. It just shows that this pandemic is not as unique as we think it is, and we have a lot to learn from history and from each other. Hopefully we are reaching the end of this pandemic, but we should make sure that we are actually learning and preparing for future public health crises. We have never before had the quantity and quality data like what we have today, and we can leverage it to radically change the trajectory of the next pandemic before it even starts.