Final HTI-1 rule now in effect, with AI in mind – and more to come, says ONC
ORLANDO – Wednesday on HIMSS24, officials from the Office of the National Coordinator for Health IT offered a series of reminders and refreshers on an important final rule that went into effect for many healthcare organizations just over a month ago, on February 8.
And ONC also provides a preview of what we can expect for the next related regulations.
Last January, the agency published the final rule for (deep breath) health data, technology and interoperability: updates to certification programs, transparency of algorithms and information sharing. HTI-1) to the federal registerwith an effective date of March 11, 2024.
Read more about our coverage of HTI-1 here and here.
The rule implements more provisions of the historic 21st century Cures Act and provides updates around standards, implementation specifications and other criteria to the long-standing ONC Health IT Certification Program.
Among them: new interoperability-specific reporting metrics, revised information blocking rules, promotion of USCDI v.3 as a basic standard and, most importantly, transparency of algorithms – setting new requirements for how artificial intelligence and other predictive algorithms are used in certified healthcare IT products.
At HIMSS24 this week, Jeffery Smith, associate director of certification and testing at ONC, offered some thoughts on the agency’s approach to HTI-1 and its approach to AI-enabled clinical decision support.
“When we update certification criteria, sometimes it’s because new standards are available,” says Smith. “Sometimes we want new functionality to be available, and sometimes that’s because a deep policy driver is needed.”
Three years ago, ONC was tasked by HHS Secretary Xavier Becerra to investigate biases perpetrated and propagated by algorithms in healthcare.
“Our starting point was really looking at: How can this technology help and hurt?” said Smith.
“You can walk around the show floor and see that there are literally hundreds, if not thousands, of different applications that can help physicians do their jobs better, and help nurses and other physicians do their jobs more efficiently,” he said at HIMSS24. “But we also understand that there is a potential for harm to enter the equation.”
And so ONC spoke to stakeholders and examined peer-reviewed literature to explore the different ways AI can cause problems in healthcare.
The technology “can do a lot of good, but it can also do a lot of bad – and on a large scale it can do bad,” Smith said. “We are not talking about one patient being injured, but hundreds, thousands and in some cases millions. There are very well-known peer-reviewed studies that look at the use of technology used in different populations to determine who will become sicker. predicting who will need more care in the future – they were systematically biased, and that impacted literally millions of patients.”
ONC is not a safety regulator like the FDA exploring approaches to AI-based software as a medical device. But it does have certification authority for IT products.
“Looked at through the prism of our certification program, we thought, ‘Okay, what can we do?’” Smith said. “How can we help people understand the quality of information or the quality of the predictive model, and what kind of information would be needed to understand that? That’s why we came up with an acronym.”
That handy acronym? FAVORITES.
High-quality predictive algorithms and decision support interventions should have five characteristics: according to ONC. They should be:
- Honestly.
- Appropriate.
- Valid.
- Effective.
- Safe.
“That’s our guiding principle, our quality framework for predictive algorithms in healthcare,” Smith said. “It was something we thought about early on, and it really drove the rest of our work with HTI-1, and specifically with the decision support and intervention criteria.”
Since then, the five FAVE criteria have been used in other cases, such as the Biden administration’s recent announcement that it had secured commitments on AI safety and ethics from some healthcare heavyweights – Boston Children’s, Geisinger, UC San Diego, Wellspan and two dozen other providers, payers and consumer health companies – promising that their AI and machine learning models will focus on fairness, appropriateness, validity, effectiveness and safety.
With the HTI-1 rule, “we thought we could bring more transparency to the data used as part of the decision support intervention, or we could bring more transparency to the implementation of the decision support intervention, and that we would bring more transparency in the implementation of the decision support intervention. By bringing transparency to the organizational competencies of the organizations developing these things, we can build more trust and help optimize the use of these algorithms,” Smith said.
“Because at the end of the day we are AI optimists,” he said. “We are the ONC. We believe in technology and its ability to improve patient care. And so we try to think about how we can lower the barriers and how we can increase the quality and use of these things.”
‘We are not slowing down’
That same day, National Coordinator Micky Tripathi posted a blog on the ONC site it provides an outlook on what stakeholders can expect from the rules still under development for HTI-2.
That stands for (again, deep breath) the proposed rule for Health Data, Technology, and Interoperability: Patient Engagement, Information Sharing, and Public Health Interoperability.
“This will be the second rule in ONC’s set of rules for health data, technology and interoperability,” Tripathi wrote. “In HTI-2 you can expect a strong focus on interoperability and a specific focus on how the adoption of strategic standards can drive interoperability.
“In fact, we see the impact of standards adoption every day in our ongoing work United States Core Interoperability Data (USCDI). We are thrilled to see the increase in USCDI use across the healthcare industry and among our federal partners.”
With HTI-2, ONC also seeks to address new public health developments through certified IT requirements, he said, and expects to propose new provisions related to API certification focused on use cases including electronic prior authentication, patient engagement and care coordination.
“We are not slowing down,” Tripathi said. “Our work to support patients, help providers deliver care, and set clear expectations across the continuum of care when it comes to healthcare IT continues.”
Mike Miliard is editor-in-chief of Healthcare IT News
Email the writer: mike.miliard@himssmedia.com
Healthcare IT News is a HIMSS publication.