The rapid development of the vaccines from companies like Pfizer and Moderna certainly represent amazing achievements. These companies leveraged cutting-edge science, such as messenger RNA (mRNA), as well as smart approaches for the approval process and the manufacturing infrastructure.
But unfortunately, the rollout of the vaccines has been far from encouraging. The fact is that the U.S. healthcare supply chain is extremely complex and not particularly amenable to quick distribution at scale.
Yet it seems that technologies like AI can help with the challenges. “Humans don’t have the capacity to consider thousands of competing and evolving factors,” said Arijit Sengupta, who is the Founder and CEO of Aible. “This is precisely what AI does best—that is, complex scenario planning and hypothesis testing that’s flexible enough to adjust quickly to new information so that decisions can be made based on the best available evidence.”
So then what are some AI approaches, recommendations and strategies that can make a difference? Let’s take a look:
Aditya Sriram, the AI Portfolio Lead and CoE Program Manager at ibi, a TIBCO company:
The clinical trials have portrayed the vaccination as a positive solution against COVID-19. But in practice, the testing is on a sample population. As the vaccination rolls out to larger populations, the possible adverse effects caused by the COVID-19 vaccinations will be more apparent. As opposed to a reactive indication on patient symptoms, AI can provide insights on patients that are likely to react to the vaccination, given their medical history and demographics information.
MORE FOR YOU
Cheryl Rodenfels, the healthcare CTO at Nutanix:
AI can be useful with the vaccine reporting process. This process requires multiple steps, and hinges on accurate reporting of data. AI allows for less human inspection and condenses the data onto one plane, so healthcare organizations do not need to build reports for multiple electronic health records or pharmacy systems.
Nigel Duffy, an EY Global AI Leader:
AI could be used to target messages and communications to those eligible to receive a vaccine. These messages could be customized by AI in content, form, and medium to motivate citizens to take the vaccine by highlighting the personal benefits to them (e.g., depending on their profession, vulnerability, etc.) and by countering any misconceptions about the risks associated with the vaccine (e.g., certain audiences might be persuaded by science, others by role models from their community). These messages could also be customized to best inform citizens about how they can access the vaccine, e.g., register via phone call, website, or text.
Ted Kwartler, the Vice President of Trusted AI at DataRobot:
Vaccine distribution is often discussed at the macro-level, such as the national or state, but practically, it also requires a much smaller scale. Within a state, what locations and subpopulations should be allocated vaccines in what order? AI-augmented simulations can, at a granular level, factor in hundreds of inputs, like mobility data, hospital utilization, and current infection rates, to forecast thousands of possible futures for dozens of locations within a state. This enables decision-makers to send vaccines to those communities and people with the greatest need right when it helps the most. However, with AI, patterns in the data need to be understood to avoid bias. An AI-enabled vaccine rollout could reinforce societal inequalities if it fails to account for racial, age, and economic risks, as the burden of sickness has been unevenly distributed.
Tony Bates, the CEO of Genesys:
Chatbots are a great example of an AI-infused technology that can address COVID-19 questions via a customer’s existing voice, chat or social channel. Genesys recently released COVID-19 Vaccine Rapid Response, a new digital solution that will enable companies to use AI-driven chatbots to help address the wave of expected requests for information from more than 330 million people in the U.S.
Dr. Steve Kearney, who is the Medical Director at SAS:
Many times when you see AI bias, it’s because you are not incorporating all of the appropriate inputs or outputs to a data model. Simply put, bias happens when you don’t see something in your data set. For example, public health experts may be basing priority vaccine decisions on county hospital paid claims data, but they may not realize that all of that data is based on people who have health insurance and access to hospital care. It largely ignores vulnerable populations without health insurance or a medical history in a given location. Public health populations must recognize social determinants of health, education, job status, and where people live by bringing in government databases and private databases alike. When you bring in all of that data you get a completely different look at the population that drives insights about where to prioritize distribution.
Chris Hale, the Founder and CEO of Kountable:
One of the key missing ingredients is a government-private partnership rooted in quality data acquisition at the last mile that aligns that data acquisition and utilization process with the data sovereignty of the individual patient. This is key to a successful government-private partnership that will have the staying power to deliver over time by building trust in the US population. In addition, the partnership would need to consider interoperability across federal and state data governance and health care portability policies. It sounds complex, but the good news is that the tech exists to do this today and there are companies that are built to deliver it and that have the values needed to align with the patient’s desire to own and control their own data.
Mike Beckley, the Chief Technology Officer and Founder of Appian:
AI can and should help with a vaccine rollout but probably not in the way people are imagining it. It doesn’t take AI to figure out who and where people are at risk from COVID-19. Instead, AI solutions like Appian Intelligent Document Processing should be used to validate vaccine eligibility. Keep people in control but use AI to read doctor’s notes validating pre-existing health conditions or utility bills to confirm residency.
The trouble now is mostly how does your local government handle the crush of calls and emails and questions—and there are Conversational AI providers like ContactEngine that can help automate appointment scheduling far better than the very expensive and bloated CRM products being heavily advertised these days.
Tom (@ttaulli) is an advisor/board member to startups and the author of Artificial Intelligence Basics: A Non-Technical Introduction, The Robotic Process Automation Handbook: A Guide to Implementing RPA Systems and Implementing AI Systems: Transform Your Business in 6 Steps. He also has developed various online courses, such as for the COBOL and Python programming languages.