Aequus Health
Equitable health is an information problem.
Before it is a technology problem. We build the context layer that turns health data into information people can trust, understand and act on.
Built on standards
- HL7 FHIR R4
- SMART on FHIR
- SNOMED CT
- GS1 Digital Link
- ISO 27269
Read only. Nothing duplicated, nothing leaves.
The problem is not scarcity.
There is no shortage of health data. Every system we work with is drowning in it: screenings, claims, notes, registries, wearables, referrals.
What is missing is context. Without it, data never becomes information anyone can trust, understand or act on. The gap does not close. It widens fastest for the people already furthest from care.
Where information breaks down
- Hard to find
- Too complex
- Not meaningful
- Untimely
- Unactionable
- Not trusted
Our thesis
Equitable health rests on three things held together.
Information, data that carries its context, and a systems approach that holds patients, clinicians and administrators in the same frame. Solve for one alone and you shift the burden onto another.
PILLAR I
Information
Can people trust and use what they are given?
High quality information is not the same as accurate information. It has to survive the journey to the person who needs it, in the moment they need it. Most of what fails in health equity fails here, after the evidence is already correct.
- Trustworthy
- credible source, accurate content
- Meaningful
- relevant to this person, relatable to them
- Usable
- obtainable, timely, actionable
PILLAR II
Data with context
Does the data carry what it means?
A screening score with no history. A note with no neighbourhood. A flag that fires in one clinic and means nothing in the next. The same score carrying who was screened, under what conditions, against what local baseline, is a different object entirely.
- Provenance
- where this came from and how much to trust it
- Population
- who it describes and who it leaves out
- Place
- the local baseline that makes a number mean something
- Point of care
- the moment and the workflow it has to land in
- Purpose
- the decision it is supposed to change
PILLAR III
Systems approach
Who does this serve, and who helped design it?
Solve for one and you shift the burden onto another. We hold all three in tension every time we design. That tension is not a constraint on the work, it is the work.
- Patients and communities
- information they can act on, in a language and format that fits their life
- Clinicians and care teams
- the right thing surfaced at the right moment, without one more system to log into
- Administrators and systems
- outcomes, capacity and cost moving in a direction the board can see
How we build
Closed-loop AI inside the human workflow.
Nothing we build is designed to replace a clinician’s judgement. Everything we build is designed to arrive before it is needed.
1
Sense
Pull signal from what already exists. Screenings, notes, claims, community data. Nothing new to collect.
2
Contextualise
Attach provenance, population, place, point of care and purpose. This is the step everyone skips.
3
Surface
Deliver it inside the workflow the person is already in, not beside it. Not one more system to log into.
4
Learn
Capture what the human did with it and feed that back into the loop. The human stays in it at every turn.
The loop returns to step one.
The Halo Suite
It augments the EMR. It does not replace it.
A secure, standards-based clinical AI platform that reads the record live over HL7 FHIR and extends it into four domains the EMR was never built to reach: the patient, the panel, the community and research. Your EMR stays the system of record. Every query and every model run stays inside your boundary.
For patients
Halo Companion
Record-grounded guidance in the patient's own language. Answers come from their own chart and your protocols, not a general model's impression of medicine. Portable IPS summaries.
For the panel
Halo Beacon
Proactive outpatient risk stratification and care-gap closure between visits. Whole-panel visibility, so the list is managed before it arrives at the door.
For the community
Halo Bridge
Culturally-fit guidance and SDOH follow-through, closing the loop from clinic to community. Built with FQHCs and local clinics in mind.
For research
Halo Prism
Cohort discovery and real-world evidence without data leaving the institution's walls. Audit trails and standards throughout.
Buy the rails once. Each application closes a gap that every major EMR shares, and the Suite grows as new gaps appear, with no new integration to buy.
Adoption on your terms
Start where you are comfortable. Stop where the value is.
Three phases, each one a complete deployment. Nothing built in one phase is discarded in the next, and your EMR relationship never changes. Halo reads it over HL7 FHIR at every phase.
Phase 0
Grounded
An open-weight clinical model on your cloud, grounded by a versioned knowledge base of your own protocols.
- Where it runs
- Your cloud, in your jurisdiction
- What you own
- Your knowledge base
- What the audit says
- Which knowledge version answered
Phase 1
Tuned
A supervised fine-tune on your de-identified corpus. The tuned model is your artefact, on your endpoint.
- Where it runs
- Your cloud, in your jurisdiction
- What you own
- Your model weights
- What the audit says
- Dataset version, job and artefact
Phase X
Edge
The same tuned model on an on-premise appliance inside the hospital. Works disconnected.
- Where it runs
- Your building, offline capable
- What you own
- Weights and the hardware path
- What the audit says
- Phase 1, plus physical control
Start at Phase 0. It proves value without a training run, satisfies in-network requirements immediately, and commits you to nothing further.
AqTrueSource
One barcode. The right answer for who is holding it.
A multi-tenant GS1 Digital Link resolver for regulated healthcare products and patients. Scan the 2D code on a medicine pack or a patient wristband and the identifier resolves in real time to the correct approved destination.
The same barcode resolves differently depending on context: product family, exact GTIN, batch, serial, or the scanner’s region and language. All governed by an auditable rules engine, with strict per-tenant isolation so no manufacturer ever sees another’s data.
Standard leaflet
Scan a product, get the current approved ePIL for that GTIN.
Recall or superseded batch
A unit from a flagged batch resolves to a superseded notice instead of the default.
Suspected counterfeit
A flagged serial resolves to an unverified-product safety page that echoes every identifier scanned, for the call centre.
Last-mile localisation
The same GTIN scanned in the Gulf with an Arabic locale resolves to the Arabic leaflet.
Patient wristband
A GSRN scan resolves to a gated flow that challenges for date of birth before revealing anything.
Next steps
A co-design workshop, before anything is built.
We map one or two priority areas already under real pressure, identify the high-yield use cases inside your bottlenecks, and assess your informatics readiness against what AI actually requires.
What we bring
- A design thinking facilitator
- A prototyping and interface designer
- Policy and regulatory experts
- Four to five weeks, at our cost
What you leave with
- A scoped Phase 0 proposal with named use cases
- A readiness and gap analysis you own outright
- Yours to act on, with us or without us
The workshop is ours to fund.
A contract begins at Phase 0, and only if the workshop earns it. What we ask in return is a named executive sponsor, access to your clinical and informatics leads, and first right to scope Phase 0 if the workshop finds one worth doing.
Aequus Health Inc