Blog /

The 2023 Future of Health Data Summit: Key Learnings from Datavant’s Annual Gathering of Healthcare Ecosystem Leaders

Publish Date
Read Time
October 17, 2023

The Future of Health Data Summit brings together over 400 leaders, thinkers, researchers and policymakers every year from across the public and private health sectors for conversation around the data challenges our industry faces. Held in Washington, D.C., the event offers a pulse check on those issues that are newly top of mind (generative AI) and those issues that persistently present themselves within the ecosystem (interoperability).

While conversation at the Summit grows from within the framework of the healthcare system, there is a general recognition that these are some of the most pressing challenges for American society writ large. And while the challenges within the healthcare space are enormous, and in many cases deeply entrenched, the tone of the day was one of optimism and a firm belief that we will see substantive change on many fronts in the relative near term.

Here are just a few of the takeaways that emerged at the 2023 Future of Health Data Summit.

The future of generative AI for health data may look dystopian, and maybe that’s not a bad thing

Although there was only one session specifically devoted to AI, the potential (and hope) for AI to address a wide variety of challenges in the health data space came up throughout the day.

Healthcare advisor and investor Dr. Gaurav Singal, who was moderating the session on AI opportunities, risks, and regulations, began by pointing out just how novel this discussion was. “The context of the conversation feels completely new. A year ago I wouldn’t have imagined we would be talking about AI in this context.”

David Sontag, Professor of Electrical Engineering and Computer Science at MIT, painted a particularly vivid picture of an AI-driven near future (~5 years from now) that he described as, “dystopian but not necessarily bad.” An example of how AI might be deployed in this vision would start with using clinical data to create appropriate justifications for claims. Providers would then use AI bots to manage claim submissions and claim denials, sending claims to payers who would use AI bots to manage claims submitted and either approve or deny them. While this sounds like it has the potential to spiral into a never ending claims-bot-war, it could also lead to an equilibrium game where AI fills the gap between the data offered by providers and the data payers need to justify a care expense. If the entire cycle is managed automatically with AI, then we could have the potential both to take the right actions and reduce costs for everyone.

Other applications of AI that were discussed included:

  • Using chatbots to generate greater trust between patients and providers, as chatbots have already scored higher on “empathy” than human providers in several studies.
  • Relieve administrative burden on clinicians, for example in the form of ambient note takers, and allow clinicians to focus the precious time they have with patients on treatment.
  • Offer second opinions to diagnoses.
  • Create seamless integration within provider workflows.
  • Play the role of data analyst in clinical trials.
  • Provide Expert Determination reports.
  • Clean up and standardize inconsistent data in health records, improving data quality.

When discussing the difficulties in regulating the vast field of generative AI tools, FDA Commissioner Robert Califf posed two guiding questions:

  • Does the algorithm reliably produce operating characteristics (i.e. well-defined outputs, unambiguity, finiteness, and language independence)?
  • Can you describe it?

He noted that generative AI is a very different phenomenon from predictive AI, and algorithms left to themselves will deteriorate over time if they are not regularly tended and updated. Commissioner Califf sees the field as much too vast to be regulated by traditional regulators and believes that successful regulation of generative AI must be in the form of community regulation.

Health equity is not being applied equitably

Like AI, the issue of health equity emerged again and again over the course of the day.

“The access issues ARE the issues,” said Dr. Timothy Huerta, CRIO and Associate Dean of the College of Medicine at Ohio State University, while discussing the opioid crisis.

“The future is already here, it’s just not evenly distributed,” Dr. Alice Chen, Chief Health Officer of Centene, declared while describing the need to think systematically about embedding equity principles across all layers of healthcare.

Commissioner Califf also put a fine point on the matter. “Technological solutions drift to society’s penthouses. Diseases sink to the poor…Right now, if you don’t generate profit with your healthcare, you’re in bad shape in this country.”

Dr. Chris Boone, Vice President and Global Head of Health Economics & Outcomes Research at AbbVie, echoed Dr. Chen in describing the challenges of serving underserved communities, where there often is a lack of good data, “There is so much focus on the ‘digital divide,’ I would call it the “data divide.”

Creating federated data models is more effective than attempting to create universal solutions

Sheenu (Jalla) Kachru, CEO of Optum Life Sciences, pointed out that “Innovation inevitably leads to fragmentation,” and this fragmentation directly challenges the notion of universal solutions. The idea of a federated approach to working with health data, one that is decentralized and collaborative, was most pronounced in the discussion around patient registries because, as Dr. John Halamka, President of Mayo Clinic Platform, pointed out, “It is hard to centralize anything at a national and global scale.”

Dr. Boone described the opportunity to agree on a methodology for interoperability and to think about the importance of data sharing as the silver lining of the COVID crisis. Dr. Halamka agreed, pointing out that “COVID created a cultural change that reduced a lot of competitive barriers,” and went on to discuss the need to develop a standardized schema to work with local regulations. “If we de-identify and tokenize [patient data] and put it in a standard schema, the data never leaves their country and collaborators work in a controlled environment to draw value from that data.”

As data gathering and sharing becomes increasingly global, the role of aggregators, collective agreements, and the use of AI as analysts will become increasingly important.

The end game of privacy frameworks should be trust.

Do patients trust providers? Do providers trust payers? Do consumers trust companies to use their data in a way that benefits the consumer?

There was strong agreement that privacy regulation should be in the service of building trust, not hobbling research initiatives or slowing the pace of innovation. While HIPAA is generally regarded as the guiding framework for patient privacy, Melanie Fontes Rainer, Director of the Office for Civil Rights at HHS, pointed out that the majority of HIPAA-related complaints her office fields are not about data breaches, but about patients’ right to access their own health records. And while it may be true that complaints regarding smaller providers arise simply due to lack of administrative capacity, these situations nonetheless erode trust with patients, who then become less willing to seek care.

This is compounded by privacy issues in the technological space. Ann Waldo, Principal of Waldo Law Offices, pointed out that HHS sent a warning letter to 130 providers about the use of tracking pixels on their websites. Andrew Kopelman, SVP, DGC and Chief Privacy Counsel at Medidata Solutions, pointed out that a major privacy failure would significantly damage the trust being built around many initiatives, discouraging people from utilizing the healthcare system and therefore decreasing the amount of health data available to improve patient care.

In discussing the difficulties of enrolling large populations in clinical trials, James Carroll, Head of Real World Evidence and Analytics for Walgreens Clinical Trials, pointed out that Walgreens has seen better than average levels of participation among Black and Hispanic customers who want more clinical research that will directly improve their lives. He believes this is because Walgreens is so deeply embedded across the U.S., particularly in areas that are considered socially vulnerable, Walgreens has taken on the role of “provider” in those spaces. Gadi Lachman of TriNetX echoed this sentiment: “As long as people trust your ability to protect patient privacy, they are willing to go along with the project. Doing massive collaboration is possible as long as you respect privacy.”

Director Fontes Rainer also noted how fraught the patient-provider relationship has become since the Dobbs decision. Her office has seen instances of states going on “fishing expeditions” in other states, looking for patients who have sought healthcare that is illegal in their home jurisdiction. Who then, should patients trust, and how do they know a given entity is trustworthy? A similar sentiment was shared by Ann Johnson, corporate Vice President at Microsoft, who pointed out that one of the current problems with generative AI is that tech companies have not done enough to instill confidence that data is being handled properly. As Deven McGraw, Lead for Data Stewardship & Data Sharing at Invitae, summed it up, “If we want people to trust what we’re doing, we have to tell them what we’re doing.”

Connecting available data is a bigger problem than a lack of data…except in the cases when a lack of data is the problem

Dr. Darshak Sanghavi, Program Manager of Advanced Research Projects at the Agency for Health, pointed out that health data is more and more a commodity: “The problem isn’t a lack of data. The problem is how do we extract intelligence from all of the data around us?” Datavant’s Chief Product Officer Shannon West got even more specific, “We can figure out how to technically exchange data, but that doesn’t mean we have semantic interoperability.”

Dr. Chen summed up this common thread across panels: “No one has a birds eye view of the data.” And Ann Johnson noted, “Healthcare is a big data problem. There is enough data in the system to do predictive analysis and have better patient outcomes, better cost structure for payers and providers, but we’re not optimizing any of that data.” Dr. Jennifer Layden, Director for the Office of Public Health Data, Surveillance, and Technology, CDC pointed out that during the COVID pandemic, data was not able to flow quickly and timely within a “complex landscape of jurisdictions.”

The opposite viewpoint about data quantity was presented during the panel on the digital landscape for mental health. Dr. Huerta discussed a project in radical data transparency at Ohio State. They discovered that communities generating mental health data did not have access to the data they were generating. After releasing all of the available data back to the communities generating it, the response they received was, “This is great data, but it’s not the data I need,” to which Huerta responded, “This is ALL the data you give us!” Dr. Huerta pointed out that without any data, people believed there was data to be had, but that it was inaccessible. Being given the available data, clinicians were able to drive better data collection and make more informed decisions about where to spend their research dollars.

The healthcare system is created by the people, for the people

Two of the more poignant moments in the day came during opening and closing remarks.

Dr. Peter Slavin, Former President of Massachusetts General Hospital, began the day with a heartfelt call to industry leaders. While describing several manifestations of dysfunction in our healthcare system (including nurses being attacked by patients on average twice a day), Dr. Slavin pointed out that for every hour of face-to-face time clinicians spend with a patient, they spend another hour on administrative work. Dr. Slavin described this growing administrative burden as the unintended consequence of piling systems on top of systems. He called on industry leaders to ask themselves before launching a new system, how it will integrate into already existing larger systems so as to reduce rather than exacerbate clinicians’ administrative workload.

And at the close of the day, FDA Commissioner Robert Califf, challenged the large room of industry leaders working to improve patient outcomes: “The most patient-centric thing you can do right now is create a longitudinal record.”

Datavant would like to extend our sincerest gratitude to all who gathered at the Future of Health Data Summit to share their expertise. We look forward to helping our partners realize their vision of a more data-driven future over the next year, and to meeting again in 2024 to discuss our collective progress toward improving patient outcomes.


Connect to the Nation's Largest Health Data Ecosystem

Request a demo

Achieve your boldest ambitions

Explore how Datavant can be your health data logistics partner.

Contact us