Healthcare AI Transparency – ONC Finalises Sweeping Rule


In a recent development, the HHS’ Office of the National Coordinator for Health Information Technology- ONC has gone on to finalize a sweeping rule that is aimed at boosting data interoperability along with patient access, which includes a provision to create transparency requirements for AI when it comes to health software.

Under this rule, developers related to clinical decision support as well as predictive tools certified by the ONC should make a limited set of identified users to get an access to information like the value of the intervention, how it ought to be used, known risks as well as inappropriate uses, and how the tool gets maintained as well as updated. Notably, the requirements will take shape and go into effect at the end of next year.

The final regulation also went on to include a number of information blocking updates as well as exceptions, such as actors sharing data under the TEFCA framework that, by the way, went live recently.

AI’s role when it comes to the healthcare sector has become a very significant issue for the sector as lawmakers along with the experts anticipate the potential benefits in terms of drug discovery, medical documentation, as well as data analysis.

There are some who are concerned about the rapid rollout of new tools, which include generative AI products, the objective of which is to decrease the number of healthcare workers’ administrative tasks, which could perpetuate inaccuracies or bias.

The new regulations that surround the health clinical decision support software along with the predictive tools look to boost transparency and at the same time decrease the risk in the industry, opined Steven Posnack, who happens to be the deputy national coordinator for health information technology, in one of the media briefings.

It has indeed been a long time in the making to get an update to this certification criterion to be in coherence with the technical evolution in industry and also the response to some of the new issues that are posed by AI and ML as well as other types of large language models, said Posnack. 

Algorithm Transparency and Information Sharing, The Health Data, Technology, and Interoperability: Certification Program Updates, or HTI-1, went on to add a number of provisions for interoperability as well as data sharing. 

The rule takes into account an updated standard when it comes to interoperability, the United States Core Data for Interoperability Version 3, at the beginning of 2026. The industry is at present making use of Version 1, and the regulation goes on to create a predictable step for the sector so as to move toward as well as build new services along with connectivity, Posnack said. 

The standard has in it updates that broaden the data elements as well as classes that can be exchanged, which, as per the ONC, will go on to help promote equity, decrease disparities, and also promote public health data interoperability. 

The rule rolls out some exceptions to data and information blocking regulations as well. Actors who do not adhere to the information sharing requests in any of the matters besides the Trusted Exchange Framework and Common Agreement, or TEFCA, will not be considered to be information blocking.

TEFCA, which happens to set a governance framework along with a technical requirement for nationwide data exchange, was live with five Qualified Health Information Networks. It went to mark a prominent milestone in an effort that went over yearlong, as per the leaders at the HHS. But the framework indeed needs industry buy-in so as to succeed.

Posnack said that they have rolled out a clear set of guidelines and approaches that are associated with this information blocking exception for those actors to become responsive and either to make the electronic health information accessible, exchangeable, or usable by way of TEFCA, so as to promote usage and also provide convenient incentives for participation in TEFCA.