For Medical Device & SaMD Companies
The gap between authorization and reimbursement is where device companies lose time, revenue, and market share. Glacis sits at your device’s inference layer and captures every clinical decision in real time — generating continuous, tamper-evident AI performance evidence as a byproduct of normal operation.
No separate study. No parallel registry. No manual data pull. That evidence feeds directly into reimbursement pathways, indication expansions, and payer negotiations simultaneously.
Your post-market surveillance team is collecting data for FDA. Your clinical affairs team is assembling evidence for payer coverage. Your regulatory team is documenting PCCP compliance. Your commercial team needs real-world performance data for health system and payer negotiations.
These are four separate efforts, using four separate workflows, often with four separate data sources — all trying to answer variations of the same question: does your algorithm perform consistently in the real world?
The answer is already in your inference layer. Nobody is capturing it systematically.
Glacis captures what your algorithm does — structured, continuous, and tamper-evident — without changing how it operates.
Glacis sits at the inference layer and records what your algorithm did for every clinical decision — inputs, outputs, confidence scores, thresholds applied, population drift signals, safety enforcement actions. Structured, continuous, tamper-evident. No separate study design required.
Every record is cryptographically signed via Notary. Not a log file. Not an export. A tamper-proof attestation that a specific inference happened, with specific inputs and outputs, at a specific time. This is what 21 CFR Part 11 compliance looks like when it’s built into the system instead of bolted on.
Performance data is automatically stratified by demographics, geography, prescriber type, and clinical setting. When your MAC asks whether the algorithm generalizes across the Medicare population, you don’t commission a study — you pull the data.
Continuous population drift alerts and PCCP threshold monitoring. When the real-world patient population shifts before it affects algorithm performance, you know. PCCP compliance becomes an automated process, not a retroactive documentation exercise.
AI performance evidence isn’t a nice-to-have — it’s the decisive input across every reimbursement and coverage pathway your device will encounter. Glacis generates it once, continuously, and it feeds all six.
FDA’s pilot launched December 2025 in coordination with CMS ACCESS. Manufacturers must collect, monitor, and share structured device performance data with FDA throughout the pilot. Glacis is that data collection infrastructure — continuous, tamper-evident, and built for exactly what TEMPO requires.
CMS’s 10-year outcome-aligned payment model begins July 2026. Recurring payments are tied to measurable patient health improvements. The AI performance evidence layer — what the device did, how consistently — is the necessary foundation for linking device use to outcomes.
Conditional coverage while you generate evidence. Without systematic, continuous performance capture, you risk the coverage period ending without the evidence needed for permanent NCD conversion.
Local and national coverage decisions require population-stratified evidence of real-world algorithmic consistency. Glacis generates it automatically, mapped to the clinical endpoints MACs evaluate.
For Breakthrough Device-designated hardware devices with AI components. Glacis provides the AI performance evidence within the Evidence Development Plan. Note: digital-only SaMD is generally excluded from TCET — TEMPO/ACCESS, CED, and LCD/NCD are the relevant pathways.
Commercial payers require demonstrated performance in their specific member population. Continuous, independently-attested AI performance data — stratified by demographics, geography, and prescriber type — is increasingly the decisive factor in negotiations.
The AI performance data Glacis generates for post-market surveillance is also the foundation for indication expansions — broader patient populations, new clinical settings, expanded age ranges, new clinical use cases.
Prospectively collected, structured performance data combined with clinical outcome linkage can replace or substantially reduce the need for a new clinical trial. One evidence infrastructure. Multiple regulatory and commercial returns.
Glacis captures what the algorithm did. A complete real-world evidence program also requires what happened to the patient. The following categories complete the full evidence stack.
Veradigm, TriNetX
The clinical outcome layer that links to Glacis performance records for complete RWE.
Datavant
PHI-safe linkage of Glacis records to EHR and claims data. Directly responsive to FDA’s December 2025 RWE guidance recommending PPRL.
Greenlight Guru, Veeva, MasterControl
Glacis feeds AI performance data directly into the quality record — PCCP monitoring, post-market surveillance, and audit trails appear automatically.
Glacis makes their work faster and more defensible. Manual data assembly becomes automated evidence generation.
Most device companies treat post-market evidence as a compliance cost. Glacis turns it into market access infrastructure.
The AI performance evidence that satisfies FDA’s PCCP requirements also forms the foundation of your payer coverage dossier, supports indication expansions, and differentiates your product commercially. One evidence infrastructure. Every regulatory and commercial conversation.
Integrations
Glacis integrates at the inference layer — your models, your pipeline, your environment.
Book 30 minutes. We’ll walk through how Glacis captures inference-layer data, what the attestation record looks like, and how it maps to your specific regulatory and reimbursement pathways. No procurement, no SOW.
Book a walkthrough