Micky Tripathi’s future may be uncertain, but he says the agency he’s been charged with running should stay put.
Tripathi, assistant secretary for technology policy at the Health and Human Services Department, is winding down his four years overseeing the federal government’s health IT policies. Like other Biden administration officials, he’ll leave his role in January. He said he plans to take some time before deciding on his next career move.
Related: What's behind ONC's name change?
While the incoming Trump administration has promised to significantly cut the size of the federal government, Tripathi doesn’t think any reductions would — or should — include the agency formerly known as the Office of the National Coordinator for Health Information Technology. ONC became the Office of the Assistant Secretary for Technology Policy in September, a move giving it oversight of HHS' department-level and interagency technology, data and artificial intelligence functions. This month also represents the one-year anniversary of when the agency made the Trusted Exchange Framework and Common Agreement, meant to create a baseline for health information-sharing across different provider systems, fully operational.
Tripathi spoke about the future of the ASTP/ONC, as well as his proudest accomplishments and biggest regrets during his four years heading the agency. The interview has been edited for length and clarity.
The Trump administration has promised to reduce the size of the federal government. Are you concerned your office could be cut or folded into another area of HHS?
I don't see any reason for it. We need technology to be looked at from a strategic perspective. Our reorganization [to ASTP/ONC] was HHS itself saying, “Technology is going to be is more and more strategic. We need to have an organization that has dedicated resources toward that ... with a dedicated set of leaders who are focused on that.” It allows us to attract top talent from industry and attract more resources in general.
In future discussions with Congress, they will fully understand what it is we're doing, and we can give them greater transparency and accountability with respect to the dollars spent on technology.
We need to have some top-down coordination to work together across the department in the use of these critical technologies.
So no, I don't expect it to go away. I haven’t talked to anyone from the incoming administration about this, but I very much anticipate the incoming administration is going to see this as well: This is what every organization has been doing or ought to be doing. I would think they would very much agree with it … and then maybe be grateful that we did a lot of the hard work of the reorganization.
What is your proudest moment of the last four years in the federal government?
Bringing a sense of urgency in advancing interoperability. In my very first public meeting, I promised I was going to be incredibly impatient with the status quo and work hard to push all of us to have public-private collaboration and help us move the dial towards our interoperability goals.
What about your biggest regret?
I would have liked to have gotten the HTI-2 rule out a little bit earlier, so that we could have a little bit more industry participation and feedback on that.
There are some areas with the Trusted Exchange Framework and Common Agreement where I wish we would be further along. For instance, payer-provider interoperability. On the other hand, I had lower expectations for public health in TEFCA and we now have over 50 jurisdictions participating in public health for TEFCA. When I came in, we had a version of TEFCA that was in draft mode. Now, we have seven Qualified Health Information Networks that are live in December and an additional three that are about to go live. So, in some ways, I think we exceeded expectations.
What are your predictions for the future of government regulation of artificial intelligence?
There’s still a lot to figure out. We tend to think of AI as one big thing, when it’s really 10 gazillion specific things. We need to recognize our understanding of AI, as people start to use it, is going to get much more nuanced and sophisticated over the coming years.
ONC’s initial regulation focused on transparency. As we start to develop industry conventions around what transparency looks like, we’ll get a better understanding of what we need to do from a federal government perspective. You also have the Food and Drug Administration’s regulatory framework that's in place [that regulates AI as a medical device]. Our approach is focused on transparency, not about approval and disapproval. We think that's a good balance right now between FDA and ASTP.
There seems to be a general recognition that we need some type of nutrition label that can tell you if the AI-based technology in your hands is safe to use. The public and the private sector will have to work together on how we get that.
Organizations like the Coalition for Health AI are building out different types of metrics and approaches to what an AI assurance lab might look like. Other organizations are doing the same thing, such as Microsoft’s Trustworthy and Responsible AI Network. The critical question for us to figure out, both in the public and private sector, is what will compel someone to feel the need to get their system certified or approved.
The question at the end of the day is: Why would Mass General Brigham, Iowa University Health System, OpenAI or anyone feel the need to go and get their systems certified? I’m confident we can get the structure in place. It’s just a matter of how it gets enforced.
