Notes from a CSIS Virtual Event: Artificial Intelligence Applications for Healthcare

By Elizabeth Benke
 
On April 1, 2022, CSIS hosted an event about how artificial intelligence (AI) can be applied to reshape the healthcare research field, featuring Wendy Nilsen (Program Director for Smart and Connected Health at the National Science Foundation (NSF)) and Caoimhe Vallely-Gilroy (Global Digital Health at the Healthcare Business of Merck KGaA in Darmstadt, Germany). The panel was moderated by James A. Lewis, SVP and Director of the CSIS Strategic Technologies Program.

When defining what AI is, both Nilsen and Vallely-Gilroy agreed that it is a broad term. Nilsen articulated that AI is an ever-changing constant that creates analytical methods to process heterogenous data at a faster rate than before. Within the field of healthcare, however, Vallely-Gilroy added that the use of AI is centered around finding ways to use rich health data to make treatments faster and better.

Training an AI to support that quest requires troves of data. Nilsen recognized there is a vast amount of data available from public and private sources, however there are two overarching challenges: (1) the sheer amount of work that comes from cleaning the data sources, and (2), the burden that befalls upon the researcher to understand how the data has been managed and measured. For Vallely-Gilroy, the main challenge to the researcher is parsing out which data is relevant and useful to their project. Diving into an example, the panelists discussed the difficulties and drawbacks of wearables such as smart watches as sources of data. Vallely-Gilroy found that there are an extremely limited number publicly available datasets that come from wearables, given the from the consumer and the company to share such data. Nilsen added that there is a question of ethics as it relates to using consumer data as health data.

Further complicating the question, Nilsen still sees wearables as a problematic source of data. Since a wearable device is an expensive purchase, the data extracted is a biased sample of a restricted demographic. Wearables cannot be approached as a data source that describes the general public. Vallely-Gilroy argued that a paradigm shift on how the public understands the value of these devices is needed: wearables could be used as a method of empowering underserved communities to understand their own health in greater detail.

So, what will the value of AI be to healthcare? It is “completely transformative”, according to both panelists. According to Nilsen, AI allows for shorter clinical trials because the data is gathered at a greater speed and the long-term data collection quality of AI allows researchers to understand subjects outside of a single slice of time. Vallely-Gilroy lauded the benefits of longitudinal data collection. She claimed that AI will shorten clinical trials and consequently lower the prices of drugs.

AI will also benefit healthcare research: Nilsen described how AI is being used in precision medicine centers to incorporate research on genomics in their clinical decisions. Vallely-Gilroy discussed how AI in the sphere of augmented reality can be used as a treatment method for patients with phobias as well as an incredible teaching tool. They agreed that AI is a powerful innovation for biological modeling. Nilsen described how AI modeling is also a strong motivator for patients to see the value of treatment. She cited the example of a patient being able to see how treatment overtime could decrease their statistical likelihood of heart attack. This would provide data that would empower patients to create behavioral changes in their lifestyle and seek treatment.

The panel concluded with a brief discussion on the research industry itself. Nilsen argued increased investment in research projects was vital. Vallely-Gilroy echoed this statement and added that investment in understanding the value of this research is vital as it solidifies the long-term investment in this field. Additionally, panelists discussed what skills are necessary for researchers entering this industry, taking into account how automation can streamline these skillsets. Vallely-Gilroy claimed that upcoming researchers need to learn how to read data and be able to communicate the value of it. For Nilsen, automation could alleviate burdens of healthcare researchers. However, she noted that it will be critical to approach this transformation ethically.

What does the future of AI in healthcare look like? Nilsen foresees more predictive models that will help patients and doctors plan treatments, and an increased use of AI in clinical trials. Vallely-Gilroy predicted more AI-supported prescriptions that model themselves off the patients' needs as well as AI-supported programs for patient behavioral change outside of prescription drug usage.
 
The full conversation was recorded and is available here.
 
Elizabeth Benke is a research intern with the Strategic Technologies Program at the Center for Strategic and International Studies in Washington, DC.

The Strategic Technologies Blog is produced by the Strategic Technologies Program at the Center for Strategic and International Studies (CSIS), a private, tax-exempt institution focusing on international public policy issues. Its research is nonpartisan and nonproprietary. CSIS does not take specific policy positions. Accordingly, all views, positions, and conclusions expressed in this publication should be understood to be solely those of the author(s).