4 healthcare AI trends: CVS Health data scientist shares details

Eugenio Zuccarelli, an innovation data scientist at Rhode Island-based CVS Health, didn’t set out on a career in AI because he wanted to work in healthcare. He simply wanted to do work that benefits people. 

“For the past few years, I’ve been working on human-centric AI – using and analyzing data to help impact people’s lives and society,” he said. “I’m not specifically only interested in healthcare, but I think it’s a great way for us to have a strong impact on people’s lives.” 

Luckily, this is exactly how he sees his role at CVS Health, a company that over the past few years has prioritized using data and analytics to transform its retail legacy and move towards broader services that develop into a ‘one-stop-shop’ experience for healthcare consumers. 

The Genoa, Italy native joined CVS Health – which owns CVS Pharmacy, a retail pharmacy chain, CVS Caremark, a pharmacy benefits manager, and health insurer Aetna – in October 2020. These days, he leads a team that spearheads healthcare AI innovation efforts across the company, particularly in complex chronic care – including diseases such as diabetes and hypertension. 

He is currently working on a predictive model that can identify the onset and course of these chronic diseases, and is already in use by dozens of large companies and some federal organizations. The solution helps understand patient risks, monitors therapies and improves medical decision-making. 

“Everything we do on the innovation side is about creating products,” he said. “We commercialize these innovative AI solutions and sell them to some of the biggest companies in the world.”

VentureBeat asked Zuccarelli for his thoughts on some of the top trends he sees in the healthcare AI space. These are four he considers top-of-mind across the industry: 

AI-driven personalized healthcare solutions 

There has long been growing interest in the healthcare industry to move toward personalized medicine, which uses genetic or other biomarker information to determine patient treatment decisions. 

But organizations are using AI to personalize solutions based on everything from behavior to lifestyle.

“It’s about how we can develop personalized solutions that are unique for each individual,” said Zuccarelli. “For example, one person might be at a higher risk of worsening of diabetes or hypertension because of the wrong diet. Other people might not be taking the right medicine.”

An AI-driven solution can drive personalized recommendations to each person, depending on where there are gaps in care. Zuccarelli compares it to Netflix, as a solution that offers tailored recommendations to specific users.

Overall, Zuccarelli maintains that personalized medicine will completely revolutionize healthcare – and AI and data will be the tools that support that shift. 

“I strongly think that doctors will always be the most important component in healthcare and the patient journey, but it’s not going to be a ‘one size fits all’ – it’s going to be about tailored solutions,” he said. 

Ethical healthcare AI applications

While the focus on AI throughout the past few years has often been around creating better-performing, more-accurate models, Zuccarelli says healthcare organizations are increasingly focused on how applications relate to the people that use and are affected by them.

“That’s why ethical AI is such a huge component, especially in healthcare, where privacy, fairness and ethics are so important,” he said. “Every day, we see AI applications that could be done better if we thought about how powerful they are as tools and how they should have guidelines and an understanding of the cost of bias.”  

Even in a regulated industry such as healthcare, there is still much to work on with data from an ethical standpoint, he adds, pointing out that many companies and research institutions can (and do) use models without any oversight around bias or discrimination that may exist, related to ethnicity, age or other issues. 

“Figuring out these nuances and how to make the system better will be one of the most important components for the [healthcare AI] industry going forward,” he said.  

Interoperability of healthcare data

Hospitals, companies and individuals have a tremendous amount of health-related data, but that data lacks interoperability. That is, data often lives in different types of systems, through a variety of vendors, or even simply exists on pieces of paper, and therefore the data cannot communicate with each other, said Zuccarelli. 

“I think that’s one of the most important components that still needs to be leveraged and resolved,” he explained. “Once there is a solution to interoperability of data, healthcare AI will get a significant boost.” 

AI can help with resolving these issues – not the algorithms themselves — but how they are used and created. Federated learning, for example, is a technique that trains an algorithm across multiple decentralized edge devices or servers holding local data samples. 

“Instead of moving the data around and fitting it into the model, you’re taking the model and moving it into all the locations where data is stored,” he said. “So the data can stay where it is, and the model can still have access to all the data.” 

Impact of big tech players getting into healthcare

“It’s going to be very interesting to see,” said Zuccarelli responding to Amazon’s announcement last week that it will acquire One Medical, a network of more than 180 clinics, for nearly $4 billion. After its push into the pharmacy space in 2020, it’s clear that the company is trying to become a leader in the health space, he explained. 

“The more players in this industry we can have, especially tech players, it will probably move the whole sector forward when it comes to finding healthcare technology solutions,” he said. 

Healthcare, he pointed out, is an industry historically interested in privacy and data ownership, he added.

“So it’s going to be very interesting to see how Amazon will be able to guarantee privacy, perhaps help interoperability and also allow the healthcare system to thrive.” 

A healthcare AI ‘holy grail‘

Zuccarelli says if he could choose his own ‘holy grail’ project, he would tackle the data interoperability issues in healthcare.

“I would probably try to solve that because there is a huge amount of data just in CVS Health,” he said. “It’s certainly one of the biggest healthcare companies in the world, but it’s one single private company. If we were able to make a single data framework where we had the ability to share data from hospitals across the world with privacy intact, we would have the ability to improve the health and wellbeing of so many people.”