On February 23, 2026, the LTPAC HIT Collaborative submitted comments to HHS in response to the Request for Information on accelerating AI adoption in clinical care. Our response centers on a critical warning: 

• AI tools being developed and deployed across healthcare are systematically trained on data that excludes or underrepresents the very population LTPAC providers serve — older adults, particularly those over 85, with multimorbidity, polypharmacy, cognitive impairment, and complex social needs. 
• Without intentional policy intervention, AI tools validated on younger, healthier, acute-care populations will produce biased, potentially harmful recommendations when applied to nursing home residents, home health patients, and hospice populations. 
• The convergence of 10,000 Americans turning 65 daily, a persistent LTPAC workforce crisis, and a 15-year federal health IT investment gap creates an urgent imperative for strategic federal action.

AI Has Promise, Limitations Must be Addressed:
Our comments outline a broad set of barriers unique to LTPAC settings, including infrastructure and resource constraints, regulatory uncertainty, limited interoperability with acute care EHRs, and fee-for-service payment structures that provide no incentive to invest in AI even when outcomes could improve. We highlight both promising applications — fall risk prediction, sepsis detection, remote patient monitoring, documentation automation, and polypharmacy optimization — and areas where AI has fallen short, including biased algorithms, alert fatigue, and lack of EHR integration. We also advocate strongly for LTPAC providers and patients, noting that family caregivers and residents need transparency, human connection, and meaningful consent — not black-box tools that shift liability risk entirely to under-resourced facilities.

Recommendations: Key recommendations from our response call on HHS to mandate age-stratified validation and bias testing for AI tools intended for Medicare populations; expand interoperability standards to include geriatric-specific data elements through the PACIO Project FHIR implementation guides; align Medicare and Medicaid payment policy to incentivize high-value AI adoption; establish clear regulatory frameworks for liability, privacy, and algorithmic transparency for non-medical device AI tools; create AI validation testbeds that include LTPAC settings and geriatric populations; and prioritize research funding for multimorbidity management, functional decline prediction, delirium detection, and caregiver support tools. The Administration's commitment to accelerating AI in healthcare is a historic opportunity — but only if LTPAC providers are at the table from the start.

Read the full comment letter below: