ACCESS: ACCESS tests outcome-aligned payment for technology-supported chronic care in Original Medicare.
Boundary: Model outcomes are not yet mature; participation does not prove clinical effectiveness of every participant.
Evidence Explorer
Know exactly what the evidence supports—and what it does not.
Records are claim-specific. Evidence type, authority, directness, uncertainty and decision relevance remain visible instead of being flattened into a citation count.
ACCESS scale: CMS reports more than 150 organizations accepted for ACCESS launch.
Boundary: Accepted status can change and does not equal final participation or endorsement.
TEMPO: TEMPO uses real-world data in connection with ACCESS to evaluate selected digital-health-device intended uses.
Boundary: Pilot participation is not FDA approval or proof of effectiveness for the evaluated intended use.
TEMPO participants: FDA currently lists Cadence Solutions' HypertensionOS and Dexcom's Glucose Health Program in TEMPO.
Boundary: Participant list is dynamic; intended uses remain under evaluation.
PROSPR: PROSPR funds seven teams to build healthspan measures and accelerated trials, with up to $144M over five years.
Boundary: Program success depends on milestone achievement; endpoints and interventions remain investigational.
Healthspan measurement: THRIVE is developing and testing a PROSPR-IC healthspan score using harmonized longitudinal data and a one-year intervention.
Boundary: Validation, calibration, and regulatory utility are not yet established.
Healthspan therapeutics: VITAL-H plans a phase 3 hybrid trial of repurposed FDA-approved drugs to establish a path for testing aging therapeutics.
Boundary: The trial is prospective and does not establish benefit before results are available.
RPM/RTM: CY 2027 PFS proposes new RPM/RTM eligibility, initiating-visit, staffing, valuation, and possible bundling changes.
Boundary: This is a proposal, not final policy; provisions may change before the final rule.
AI real-world performance: FDA identifies post-deployment AI performance, drift, clinical outcomes, infrastructure, and human-AI interaction as important real-world evaluation concerns.
Boundary: The referenced request for comment is not final guidance and does not itself establish enforceable requirements.
Geroscience agenda: NIA will convene the Fifth Geroscience Summit in December 2026 to revisit progress and knowledge gaps with a focus on health.
Boundary: Future agenda-setting event; does not establish scientific consensus in advance.
Medicare App Library: CMS's Medicare App Library includes conversational AI and diabetes/obesity prevention-management use cases and defines participation requirements.
Boundary: The library's presence and criteria do not establish clinical effectiveness of any individual listed product.
CMS Health Tech Ecosystem: CMS reports general availability in July 2026 across multiple Health Tech Ecosystem categories including networks, providers, EHRs, payers, and patient-facing apps.
Boundary: General availability does not mean every pledged organization is fully deployed or that all integrations have equivalent maturity.
USCDI v6: ASTP/ONC approved USCDI v6 through the 2026 SVAP process.
Boundary: Approval through SVAP does not guarantee immediate implementation across all certified health IT products.
DHTs in drug development: FDA opened a 2026 funding opportunity to explore digital health technologies such as actigraphy, photography, and contactless sensors in drug development.
Boundary: A funding opportunity is not evidence that any DHT measure is validated or accepted as a regulatory endpoint.
Real-time clinical trials: FDA announced real-time clinical-trial proofs of concept using AI and data science to improve data flow, safety monitoring, and decision speed.
Boundary: Proofs of concept do not establish that real-time approaches will be appropriate or superior for every trial.
ADVOCATE: ADVOCATE seeks clinical AI agents that can act on cardiovascular care tasks and a supervisory AI overseer that monitors deployed agents.
Boundary: Program objectives are prospective and do not establish that such systems are yet clinically validated or FDA authorized.
IGoR: IGoR aims to build AI-enabled research ecosystems that generate hypotheses, conduct experiments, and accelerate biological discovery.
Boundary: The stated 10× acceleration is a program target, not an achieved result until validated.
ARPA-H small business awards: ARPA-H announced up to $28 million in small-business awards spanning wearables, AI, diagnostics, and gene therapies.
Boundary: Awards indicate funded development, not clinical validation, regulatory clearance, or commercial adoption.
PROSPR GPER agonist: ARPA-H awarded Linnaeus Therapeutics up to $21.9 million to test a GPER agonist for improvement in aging outcomes.
Boundary: The project is investigational; benefit on aging or healthspan outcomes has not been established.
AI in drug development: FDA reports significant growth in drug application submissions that include AI components across the drug-development lifecycle.
Boundary: FDA's statement describes submission trends, not a blanket endorsement of specific AI methods or outputs.
Medicare telehealth: CMS states that starting in CY 2026 it will add services to the Medicare telehealth services list only on a permanent basis.
Boundary: This does not mean all existing telehealth flexibilities are permanent or that every telehealth use case is covered.
AI-enabled medical devices: FDA maintains a public list identifying AI-enabled medical devices authorized for marketing in the United States.
Boundary: Inclusion on the list reflects marketing authorization; it does not establish comparative effectiveness or widespread clinical adoption.
HTI-5: HTI-5 proposes an updated FHIR-oriented certification foundation and discusses automated access to electronic health information, potentially including autonomous AI systems.
Boundary: HTI-5 is proposed policy; final language and implementation may differ materially.
RPM oversight: CMS highlighted OIG findings that about 43% of Medicare enrollees receiving RPM did not receive all three components of remote patient monitoring.
Boundary: The finding is based on the reviewed period and does not mean 43% of all current RPM is inappropriate or ineffective.