Dir, Computational Health Informatics Program @Bos_CHIP | Harvard Med Prof & MD | Nat Acad Med Leadership Council & Co-Chair Digital Health Action Collab #AI
It took a village. New HHS rule ensures an app written once will run anywhere in the healthcare system. & patients can access a computable copy of their medical record. This journey began > a decade ago when @zakkohane and I proposed a health care API in the @NEJM@SMARTHealthIT
Perspective by Kenneth D. Mandl, MD, MPH (@mandl): Unseen Commercial Forces Could Undermine Artificial Intelligence Decision Support https://t.co/dZMZutlGdc
#AIinMedicine
My Perspective in @NEJM_AI. AI could distort clinical decision-making in ways that prioritize profit over patient care. Oversight & regulation must go beyond performance metrics alone to address hidden commercial forces that could shape decision support. https://t.co/iw27zcuUcZ
If #AI is General Purpose Technology (the other GPT 🙂), akin to steam power, electricity and information technology, it’s poised to reshape the workforce in profound ways @LHSummers@nberpubs
https://t.co/4GDqXEdlsW
🚨 Faculty Opportunity at Harvard Medical School 🚨
@Bos_CHIP at Boston Children’s Hospital is hiring research faculty to advance AI & computational approaches in healthcare.
🗓️ Apply by Jan 13, 2025
🌐 https://t.co/k2Hf0ilAWV
House Taskforce on AI: "If AI tools cannot easily connect with all relevant medical systems, their adoption and use could be impeded." But is
well-addressed by @HHS_TechPolicy 21st Century Cures Act rule, as I have described in @NatureMedicine with @gotdan and @JoshCMandel
🚨 No Embargo (was 12 months) This is a big deal change to NIH’s Public Access Policy.
Starting 12/31/2025, all NIH-funded research must be freely available on PubMed Central upon publication.
Enhancing transparency & public access to taxpayer-funded research. 🌐📚 #OpenAccess
🚨 Faculty Opportunity at Harvard Medical School 🚨
@Bos_CHIP at Boston Children’s Hospital is hiring research faculty to advance AI & computational approaches in healthcare.
🗓️ Apply by Jan 13, 2025
🌐 https://t.co/k2Hf0ilAWV
🔬 Harvard Medical School Postdoc in AI & Computational Health
Join CHIP at Boston Children’s Hospital to advance healthcare with multimodal real-world data
🌟 Help shape the future of medicine. Apply now:
https://t.co/T1aWMjlaTY
Here's how today's HTI-2 Final Rule stacks up against @SMARTHealthIT comments (thanks Gemini Experimental!)
---
Here's a breakdown, organized by the topics you raised in your comments:
1. FHIR Bulk Data Enhancements
Export Performance Parity:
Status: NOT ADDRESSED in this final rule. This was one of the more ambitious proposals, and it seems ASTP/ONC has decided to defer it for later consideration. This topic falls under the broader category of proposals from HTI-2 that were not addressed in this particular rule.
Assessment: Feedback not heeded in this final rule; likely deferred.
Server Support for Patient Group Creation:
Status: NOT ADDRESSED in this final rule. Similar to export performance parity, this was likely deferred. ASTP/ONC has acknowledged the early work in the Argonaut Project but has not yet taken action to incorporate it into the regulations.
Assessment: Feedback not heeded in this final rule; likely deferred.
Server Support for _since and _type:
Status: NOT ADDRESSED in this final rule. This proposal, while valuable, was likely deemed less urgent than other priorities addressed in this partial final rule.
Assessment: Feedback not heeded in this final rule; likely deferred.
2. EHI Export
Status: NOT ADDRESSED in this final rule. Your recommendations for enhancing EHI Export, including download functionality in patient portals, third-party app access, and handling of non-computable formats, were not part of this partial final rule. These are significant proposals that likely require more consideration and potentially a separate rulemaking process.
Assessment: Feedback not heeded in this final rule; likely deferred.
3. Image Links in API Responses
Status: NOT ADDRESSED in this final rule. The specifics of image access via API were not covered in this partial final rule. Your concerns about the shareability and programmatic accessibility of "imaging links" remain unaddressed for now.
Assessment: Feedback not heeded in this final rule; likely deferred.
4. Subscriptions for US Core Data
Status: NOT ADDRESSED in this final rule. The proposal to require subscription capabilities for US Core data was not included in this partial final rule. Your recommendations regarding a phased approach, alignment with Argonaut's work, and specific technical details were not acted upon at this time.
Assessment: Feedback not heeded in this final rule; likely deferred.
5. Dynamic Registration outside of Trust Frameworks
Status: NOT ADDRESSED in this final rule. The proposal to adopt dynamic registration, along with your recommendation to support self-signed certificates, was deferred.
Assessment: Feedback not heeded in this final rule; likely deferred.
6. CDS Hooks
Status: NOT ADDRESSED in this final rule. The inclusion of CDS Hooks 2.0 in the certification requirements was not part of this partial final rule.
Assessment: Feedback not heeded in this final rule; likely deferred.
Overall Assessment
This final rule primarily focused on establishing the regulatory framework for TEFCA and making some limited updates to the ONC Health IT Certification Program. It did not address the majority of the substantive proposals in your comments, which were largely related to:
Enhancements to FHIR Bulk Data Access
Improvements to EHI Export
Specific API requirements for image access
Implementation of subscriptions for US Core data
Details of dynamic registration
Adoption of CDS Hooks
It's important to reiterate that the absence of these topics in this final rule does not necessarily mean they have been rejected. ASTP/ONC has explicitly stated that the remaining proposals from HTI-2 are still under consideration for future rulemaking. It's likely that your comments will be taken into account as they develop those future rules.
It may be worthwhile to follow up with ASTP/ONC during future comment periods or through other engagement channels to reiterate your recommendations and track their progress.
Excited to join my friend @mandl to chat with the amazing crew at Boston Children's about *all the things.* I'm an open book as you all know. If you want to join us, Zoom link here: https://t.co/8zPsHsEZ7v
Excited to join my friend @mandl to chat with the amazing crew at Boston Children's about *all the things.* I'm an open book as you all know. If you want to join us, Zoom link here: https://t.co/8zPsHsEZ7v
EHI Export insight: over 100 individual users (i.e., provider org staff) contributed to some aspect of my EHI in the past 6 years of ambulatory care.
This diagram shows only users & tasks mentioned at least 10 times. (Task labels generated by GPT-4, based on Epic's EHI schema.)
A message from Quinn Sutherland aka @ReelQuinn, who is a treasure:
“Hi everyone. Due to some setbacks in my health I’ve been struggling to make ends meet. If you can help out in any way, I’d appreciate it. Thank you!” https://t.co/pxmpgSKYFo