My Health Record holds the clinical histories of 26 million Australians. Every discharge summary, every pathology result, every medication prescription, every specialist letter — stored, indexed, accessible.
And almost none of it talks to any of the rest of it.
A GP in Broken Hill treats a patient presenting with a rare autoimmune marker. Somewhere in the My Health Record system, 340 other Australians have presented with the same marker in the last 18 months. Their treating clinicians made decisions. Some of those decisions worked. Some did not. That outcome intelligence exists. It was generated by the system. It is sitting in the system right now.
None of it reached the GP in Broken Hill before she made her decision.
This is not a criticism of My Health Record. It is a description of what every national health record system in the world currently is: a passive store of clinical data that generates intelligence as a byproduct and then discards it. The intelligence never circulates. It never compounds. It never reaches the clinician who needs it at the moment they need it.
The question is not whether My Health Record is working. It is. The question is what the architecture for the next phase looks like — the phase where stored records become a living intelligence network.
What My Health Record Does, and What It Does Not Do
My Health Record is a shared health summary system. Its primary function is accessibility: ensuring that a treating clinician in an emergency department can see what medications a patient is on, what allergies they carry, what conditions have been diagnosed. This is genuinely valuable. It eliminates dangerous gaps. It saves lives in acute situations.
What it does not do — what it was never designed to do — is synthesise across sites.
When a patient in Perth presents with a condition, and a patient in Cairns presents with the same condition two weeks later, those two clinical events are stored as separate records. They are not connected. No intelligence is extracted from the co-occurrence. No signal is generated about the pattern. No routing mechanism delivers the Perth clinician's outcome to the Cairns clinician's context.
The architecture is a store-and-retrieve architecture. Records go in. Records come out when requested. The synthesis step — the step where patterns across records generate new intelligence — does not exist.
This is the gap. It is structural, not operational. You cannot fix it by adding more storage, more bandwidth, or more staff. It requires a different layer of infrastructure.
The Three Gaps in Any National Health Record System
Every national health record system operating today faces the same three structural gaps. My Health Record is not unique in this — it is simply one of the most mature systems in the world, which means the gaps are more visible.
Gap One: Real-Time Synthesis Across Sites
Clinical events happen continuously across thousands of sites. Patterns emerge across those events — disease clusters, treatment response trends, adverse event correlations — but the synthesis that would surface those patterns does not happen in real time. It happens retrospectively, in research papers, months or years later. By the time the pattern is published, the clinical moment has passed.
Gap Two: Cross-Site Learning That Reaches Clinicians
Even when synthesis does happen — in research, in audit reports, in clinical guidelines — the mechanism for delivering that learning to the treating clinician at the point of care is fragmented. Guidelines are updated. Alerts are issued. But the specific outcome intelligence from a patient three hospital systems away, who presented with the same profile two weeks ago, never arrives.
Gap Three: Rare Condition Intelligence
Rare conditions are rare by definition. A single site may see one or two cases in a decade. The intelligence required to treat those cases effectively is scattered across the entire system — in small case clusters at dozens of sites. Without a mechanism to aggregate and route that intelligence, each clinician treating a rare condition is effectively starting from scratch. The collective knowledge of the system is invisible to them.
These three gaps are not unsolvable. They require a routing layer — a mechanism that extracts distilled intelligence from clinical events, routes it by semantic similarity to where it is relevant, and delivers it without moving raw patient data.
How Outcome Routing Works: The Complete Loop
The Quadratic Intelligence Swarm (QIS) protocol, discovered by Christopher Thomas Trevethan on June 16, 2025, describes a complete routing architecture for distributed intelligence. It is not a single component. The breakthrough is the complete loop — every step working together to enable intelligence to circulate across a network without raw data ever leaving its origin point.
The loop works as follows.
A clinical event occurs. A local processing layer at the originating site distils the event into an outcome packet — approximately 512 bytes. This packet contains no protected health information. It contains no raw clinical data. It contains only distilled outcome intelligence: what was observed, what was done, what resulted, expressed as a structured summary.
That outcome packet is semantically fingerprinted — a representation of its meaning that characterizes the problem class. The fingerprint is used to generate a deterministic routing address: a location in the network that corresponds to similar outcome intelligence. The packet is routed to that address using a distributed routing mechanism — such as DHT-based routing — at a cost of at most O(log N) across N participating nodes.
At the destination, agents holding similar outcome packets synthesise across them locally. N agents produce N(N-1)/2 synthesis pairs — quadratic synthesis depth at logarithmic routing cost. The results of that synthesis are available to any clinician or system that queries the relevant semantic neighbourhood.
When the GP in Broken Hill queries for outcome intelligence on her patient's autoimmune marker, the routing layer finds the semantic neighbourhood that corresponds to that marker, retrieves the synthesised intelligence from the relevant outcome packets, and delivers it — without any of the originating patients' records ever leaving their originating sites.
The raw data never moves. Only distilled outcome intelligence circulates. Privacy is enforced by architecture, not by policy.
What This Means for Australian Healthcare Specifically
Australia's health geography makes outcome routing particularly consequential.
Rural and Remote Care
Australia has one of the world's most dispersed rural populations. Clinicians in remote communities regularly encounter presentations for which local case volume is insufficient to build clinical intuition. Outcome routing means those clinicians are connected to the aggregate intelligence of the entire system — not just their local case history. A nurse practitioner in the Kimberley has access to the synthesised outcome intelligence from every similar presentation across the country.
Indigenous Health
Indigenous Australians experience disproportionate burden of chronic disease. Many of the conditions involved — chronic kidney disease, type 2 diabetes, rheumatic heart disease — have presentation patterns and treatment response profiles that differ from the broader population. Outcome routing across a sufficient case volume can surface those patterns and make them available to treating clinicians in real time, without requiring years of retrospective research to translate into practice.
Rare Disease Diagnosis
Australia's rare disease population is estimated at 2 million people. Average time to diagnosis for a rare disease is 4.8 years. Outcome routing collapses the effective case volume available to any treating clinician. A condition that appears twice at one site appears hundreds of times across the network. Pattern recognition accelerates. Diagnosis timelines compress.
Pandemic and Biosurveillance
A distributed outcome routing layer over My Health Record would provide real-time signal on emerging disease patterns at national scale. Clusters of novel presentations would generate outcome packets that route into shared semantic neighbourhoods, surfacing signals days or weeks before traditional surveillance mechanisms would detect them. The infrastructure for early warning is already implicit in the data. It requires only a routing layer to activate it.
How QIS Maps to Existing FHIR and HL7 Infrastructure
The most common concern raised about new health intelligence architectures is integration cost. Replacing existing systems is not feasible. Schema changes require years of coordination. Clinical workflows cannot be disrupted.
QIS outcome routing does not require any of this.
Outcome packets are FHIR-compatible by design. The distillation layer that generates them operates on existing FHIR resources — AllergyIntolerance, Condition, MedicationRequest, Observation — and produces output expressible as a FHIR DocumentReference or Communication resource. No schema change is required at the source system.
The routing layer operates above the transport layer. It is indifferent to whether the underlying transport is REST API, HL7 v2 messaging, direct database query, or any other mechanism. QIS is transport-agnostic. It routes outcome packets, not raw records. The existing FHIR infrastructure of My Health Record becomes the source layer. The routing layer sits above it.
Implementation does not require rearchitecting My Health Record. It requires adding an outcome distillation and routing capability as an additional layer — one that reads from existing data, generates outcome packets, and participates in the routing network.
My Health Record Today vs. With QIS Outcome Routing
| Capability | My Health Record Today | With QIS Outcome Routing |
|---|---|---|
| Record accessibility | Full, on request | Full, on request |
| Cross-site pattern detection | Retrospective, research-lag | Real-time, continuous |
| Rare condition intelligence | Invisible below case threshold | Aggregated across all sites |
| Rural clinician intelligence access | Local case history only | National outcome network |
| PHI exposure surface | Controlled, centralised | Architecture-enforced zero raw movement |
| Biosurveillance sensitivity | Days to weeks lag | Near real-time signal |
| FHIR compatibility | Native | Native, no schema change |
| Synthesis depth at scale | Linear (per-record retrieval) | N(N-1)/2 pairs at O(log N) routing cost |
The Routing Layer Is the Missing Piece
My Health Record is one of the most significant digital health infrastructure achievements in the world. Twenty-six million Australians have a shared health record. The data is there. The coverage is there. The trust framework — built painstakingly over a decade — is there.
The intelligence that record system generates every day, in every clinical event across every participating site, is currently evaporating. It is generated and discarded. The system that could route it, synthesise it, and deliver it to the clinician who needs it at the moment they need it — that system does not yet exist.
The architecture for it does.
The infrastructure is already there. The routing layer is the missing piece.
Christopher Thomas Trevethan discovered the Quadratic Intelligence Swarm (QIS) protocol on June 16, 2025. 39 provisional patents filed. IP protection is in place. qisprotocol.com
Top comments (0)