SafeMesh
The platform that produces the governed record of every responsibility transfer across healthcare boundaries. Who was responsible, under what authority, from what moment, and what transferred it to the next party.
Start with one governed boundary. One live Responsibility Ledger. One measurable Clearing Metric. Prove governance there. Then extend across the network.
Where SafeMesh Works
SafeMesh is deployed across five clinical use cases. Choose the one closest to your challenge.
Neighbourhood Health
Eight organisations in one building. Twelve boundaries between them. SafeMesh makes every boundary visible, scored, and governed.
Learn more →Hospital Discharge
A patient leaves hospital. Who is responsible? SafeMesh makes responsibility transfer bilateral and confirmed.
Learn more →Multi-Provider Networks
Insurers route patients across provider networks. Every referral crosses a boundary. SafeMesh governs the crossings.
Learn more →Clinical Safety Across Boundaries
Safety cases compiled within organisations miss the risks that live between them. SafeMesh maintains hazards as living infrastructure.
Learn more →Remote Monitoring
A reading taken at home is only useful if someone is responsible for acting on it. SafeMesh extends governance to the patient's home.
Learn more →What Inference Clinical Is Not
Before explaining what SafeMesh produces, it is worth being precise about what Inference Clinical is not — and why the categories most people reach for do not solve the governance problem at care boundaries.
| Category | What it does | Why it does not solve the governance problem |
|---|---|---|
| EPR / Clinical System | Records clinical data, diagnoses, medications, letters. | Records what happened clinically. Does not record who was responsible under what authority, whether handovers completed, or whether consent for cross-boundary access was obtained and proved. |
| Workflow / Referral System | Routes tasks and documents between organisations. | Routes the work. Does not create a legally defensible chain of custody for responsibility. A referral sent is not a responsibility accepted. |
| Audit Log / Compliance Tool | Records system events for retrospective review. | Records events sequentially. Does not model responsibility state — who holds it, under what authority, whether it transferred correctly. |
| Document Exchange (MESH / NRL) | Moves clinical documents between organisations. | Moves documents reliably. Does not answer whether the receiving organisation accepted responsibility for acting on the document. |
| Policy and Governance Framework | Describes how organisations should behave. | Describes governance. Does not execute it. A policy that says "explicit consent is required" does not produce the consent record. |
Inference Clinical is not competing with any of these categories. It is the governance layer that makes them defensible when care crosses organisational boundaries. It sits beneath clinical systems, not alongside them.
The Responsibility Ledger
SafeMesh produces the Responsibility Ledger — the append-only, tamper-evident system of record for responsibility state across a patient pathway. It records who held responsibility, under what authority, from what moment, and what transferred it to the next party.
The Responsibility Ledger is not a clinical record. It does not record what clinicians decided. It records the chain of custody that makes every transition auditable. When something goes wrong at a care boundary — a referral not acknowledged, a discharge not followed up, a consent not properly obtained — the Ledger provides the foundation of fact. Not a reconstruction from partial records. A governed, timestamped, tamper-evident record of what happened.
The three lifecycle disciplines — Responsibility Transfer, Clinical Safety, and Clinical Data — are how SafeMesh works. The Responsibility Ledger is what it produces.
The Evidence Fabric
The Responsibility Ledger records responsibility state. The Evidence Fabric makes every element of that state demonstrable. It weaves seven threads — one for each of the Seven Flows — into a structure where consent is linked to identity, identity to provenance, provenance to responsibility transfer, each transfer to clinical intent, each closed transfer to an outcome.
When a regulator, an insurer, or a court asks what happened, the answer is not a reconstruction from partial records. It is a fabric of independently corroborated evidence.
The Clearing Metric
Stripe does not move money. It clears transactions. SafeMesh does not move patients. It clears responsibility.
No commissioner, CIO, or Clinical Safety Officer in the NHS can currently answer: "Right now, how many patients are between responsible clinicians, and how long has each one been waiting?" The Clearing Metric provides that answer in real time.
Clearing Volume
Across all governed boundaries, how many responsibility transfers are open right now — a consent not confirmed, a referral not accepted, a handover not completed?
Clearing Age
Which open transfers are within the safe window, and which have exceeded the clinical threshold and need immediate action? The tail matters more than the average.
Clearing Rate
The headline metric. A Clearing Rate of 99.8% means 2 in every 1,000 responsibility transfers exceeded their clinical threshold. Trackable over time. Comparable across pathways. Demonstrable to regulators.
Clinical Intent and Outcome
The intelligence backbone that transforms the Responsibility Ledger from a compliance record into a management intelligence system.
Clinical Intent
Every responsibility transfer carries a Clinical Intent record: clinical domain, scope, urgency, and expected action. Clinical Intent sets the clearing threshold. A same-day urgent referral has a different clearing clock than a routine six-week pathway.
Outcome
When a clinical action completes, the Outcome flow records it: action taken, patient state, time to resolution, and whether the outcome matched the intent. Outcome closes the governance chain and is the source of every piece of management intelligence the platform produces.
Clinical Intent and Outcome are why the Evidence Fabric is a fabric, not a log. Each thread is woven into the others. That is the intelligence layer that makes the platform valuable beyond compliance.
How SafeMesh Works
SafeMesh delivers three lifecycle disciplines. Each runs as a continuous loop from boundary identification through production enforcement to operational learning.
Responsibility Transfer
The core of SafeMesh. Every material boundary crossing is identified, governed, enforced in production, and measured. Responsibility is bilateral and confirmed. Failure is a governed condition with escalation, not silent degradation.
Clinical Safety
DCB 0129 covers the manufacturer. DCB 0160 covers the deploying organisation. SafeMesh covers the crossing. Safety cases run in production with continuous hazard visibility, LFPSE integration, and evidence generated as a by-product of normal operation.
Clinical Data
Data enters from clinical systems and devices, validated and transformed to FHIR R4 UK Core with full provenance. When it crosses a boundary, governance context travels with it. Outcome data closes the loop so every intervention can be evaluated.
How SafeMesh Is Different
| Category | What exists today | SafeMesh |
|---|---|---|
| Scope | Governance within organisations | Governance across organisational boundaries |
| Responsibility | Assumed at handover | Transferred bilaterally with confirmation |
| Clinical safety | Document-based, compiled for audit | Infrastructure-based, maintained in production |
| Consent | Captured once at registration | Evaluated at point of action, portable |
| Failure mode | Silent degradation | Governed condition with escalation |
| Evidence | Compiled retrospectively | Generated continuously as by-product of operation |
| Interoperability | Data moves between systems | Data and governance context move together |
Why Governance Cannot Be Bypassed
SafeMesh enforces governance through structural constraints, not administrative controls. The Constitutional Spine separates the platform into three layers: Facts (records reality, cannot make governance decisions), Evaluation (interprets facts against rules, cannot confer authority), and Acceptance (responsibility exists only when explicitly accepted by a named party).
No component spans all three layers. No service can silently grant permission. This separation is enforced at build time and runtime. Where today's systems inform and hope, SafeMesh constrains and verifies.
For Clinical Safety Officers
CSOs today sign off on risks mitigated by training and alerts. SafeMesh provides engineering controls that strengthen the safety case. The CSO's exposure is reduced. The evidence is always current.
"The system prevents wrong-patient referral through bilateral identity verification at the boundary"
is a stronger mitigation than "clinicians are trained to check patient identity."
Regulatory Alignment
The Data (Use and Access) Act 2025 sharpens three requirements that SafeMesh addresses directly: mandatory information standards for NHS IT suppliers (Section 121), explicit consent for private providers who cannot rely on the NHS's public task as lawful basis (Article 6(1)(e)), and human acceptance before automated decisions on health data (Article 22A).
Built on AWS
SafeMesh is deployed exclusively on AWS. Every deployment starts with a secure, Well-Architected landing zone defined and operated by Inference Clinical. Governance invariants — obligations under DUAA, DSPT, DCB 0129/0160, DTAC, and UK GDPR — are structural properties of the environment.
For organisations migrating to AWS, SafeMesh drives both clinical compliance and cloud adoption. Governance and migration run in parallel from the landing zone onwards.
AWS partnership