Working today
- Emme, in pilot. An agentic patient intake, clinician documentation and sign-off, and AI-drafted care recommendations.
Masters Media
Connecting a patient’s record to the world’s evidence.
Emme is a patient-owned record and care-plan tool, now in pilot. Masie is a research platform we are building to show how far the evidence behind each claim can be trusted. Linking the two will let a care-plan draft show the clinician what the newest evidence says, alongside the guideline. The clinician makes every final decision.
Two separate systems, both from Masters Media. Emme is in pilot. Masie is in research and development.
Why this exists
The person is the only constant in their own care.
Doctors change. Clinics close. People move cities, and countries. Every system holds a fragment of a person’s history and none of them holds the whole thing, least of all the person. Meanwhile the research that should inform their care is published faster than anyone can read, and reaches the consulting room years later, if at all.
Both problems are the same problem. The information exists. It is just spread across many systems, and rarely in one place. Emme puts the history with the person. Masie is being built to keep the evidence current. Linking the two brings both to the clinician at the moment a decision is being made.
1 / 7
primary care visits had important clinical information missing, and clinicians believed it could harm care.
$27–78B
lost every year to failed care coordination in the United States alone.
42%
of OECD countries let patients both access and interact with all of their records.
Smith PC et al., JAMA 2005;293(5):565–571 · Shrank WH, Rogstad TL, Parekh N, JAMA 2019;322(15):1501–1509 · OECD, Health at a Glance 2023, digital health chapter, 2021 survey, self-reported by countries.
Where it stands
What works today, and what we are connecting.
The idea is simple to say and takes time to build. So here is the honest version, split into what exists, what we are building now, and what comes after.
The link, and the Masie research
- The link. Emme querying Masie through an API, so care-plan drafts are grounded in current evidence.
- Masie. Research into scoring how far the evidence behind each claim can be trusted, kept current as new papers, retractions and replications arrive.
- Review flags. Telling the clinician when the evidence behind a signed care plan has moved.
The outcomes loop
- Masie. Taking in de-identified real-world outcomes from Emme patients, weighted against trial evidence with their biases made explicit.
Until the link is live
Today the evidence behind an Emme care-plan draft is AI-synthesised, and the clinician sees a warning on the evidence screen to verify it before any clinical use. Governed evidence from Masie replaces it once the link is live.
The two systems
Separate on purpose.
One is for clinicians and patients and holds a person’s record. The other is for researchers and holds no patient identifiers. They are built, run and governed as two systems.
For clinicians and patients
Pilot · test data only
Emme
EMME — Evidence-based Multidisciplinary Medicine for Everyone
A patient-owned health record and care-plan tool. It starts with an agentic patient intake. Clinicians use that intake to document visits, AI agents draft evidence-informed care recommendations, and the clinician makes every final decision.
It is built for the whole care team, every qualified professional who adds value to a patient’s plan, medical doctors included.
Told once. Never again.
The patient gives their full health history one time. Every visit after that, and every clinician on the team, builds on it, confirming what changed instead of starting from a blank form.
- The patient owns the record and can take it with them.
- Clinicians keep their own copy of what they documented.
- Every care-plan draft is screened against the patient’s own medications, conditions, allergies and pregnancy.
- Sign-off is timestamped, and activity is written to an audit log.
- AI drafts and a person reviews and signs. It informs the decision; it never makes it.
- Linked to Masie, so each draft rests on current evidence. Being built.
Holds the patient’s record, identified and owned by the patient.
For researchers
In development
Masie
Multi-Agent Synthesis Intelligence Engine
A separate biomedical research platform, in development. It is being built to find out whether the credibility of the evidence behind each scientific claim can be computed reliably enough to help the people who write and apply clinical guidance.
A systematic review is a photograph, accurate the day it is published. Masie is being built to keep the picture current.
Being researched now.
- A calibrated credibility score for each claim, drawing on retraction status, replication, funding and publication bias, and statistical quality.
- An evidence knowledge graph that records where claims agree, conflict or remain uncertain, updated as new papers arrive.
- Showing guideline committees where new evidence diverges from current recommendations, across jurisdictions.
- Making explicit who was studied, so it is clear when evidence may not apply to a particular patient.
- Designed to hold no patient identifiers.
Will hold evidence only. No patient identifiers.
The connection
A record that will link to current evidence.
Emme will query Masie through an API. When a clinician reviews a care-plan draft, they will see what the newest evidence says alongside the guideline, and how far it can be trusted.
Clinicians and patients
Emme
Drafts a care plan. For each clinical question it raises, it asks Masie what the evidence says now.
Holds the patient’s record. Identified, and owned by the patient.
API · being built
A clinical question, with no patient identifiers
Scored evidence, with how far it can be trusted
Later
De-identified outcomes, with the patient’s consent
Researchers
Masie
Will score the evidence behind each claim, and answer with what it says and how far it can be trusted.
Will hold evidence only. No patient identifiers.
Solid arrows are being built now. The dashed arrow comes later.
Why it needs its own engine
Showing how far evidence can be trusted means scoring every claim against retractions, replications and bias, and keeping that score current. That is a research problem in its own right, which is why Masie is a separate system.
Only the question crosses
Masie is for researchers and holds no patient identifiers. Emme is for clinicians and patients. By design, a de-identified clinical question goes one way and scored evidence comes back.
Between guideline updates
Guidelines are reviewed every few years, by design, so the work can be done carefully. Evidence keeps arriving in between. The link is meant to show the clinician what has been published since, and how far it can be trusted, while the plan is being drafted.
Being built
When the evidence moves, the plan is flagged
Emme records the evidence a care plan rested on at the moment the clinician signs it. Once the link is live, it will check that evidence against Masie and tell the clinician when a conclusion has moved, so they can review and re-sign. A decision made last year does not quietly stay wrong.
Where it goes
Then the loop closes.
Later, Masie will take in de-identified real-world outcomes from Emme patients, with their consent. It will weight them against trial evidence and make the biases of each explicit. Its agents can then reason across both.
“This is what the trials say about type 2 diabetes management, and this is what happened in Emme patients with a similar clinical profile.”
The kind of answer this will make possible. Illustrative, not a live result.IllustrativeHow a result will read · not live output
Type 2 diabetes management
What the trials say
Trial evidence, graded for certainty, with its risk of bias shown.
What happened in Emme patients
De-identified outcomes for patients with a similar clinical profile, with their own biases made explicit.
Result. A care plan matched to this patient’s comorbidities, history, and what actually worked for people like them.
Medicine is moving from population averages toward the individual patient. A care plan matched to a person, rather than to an average, is what we are building toward.
The roadmap
Four products, one loop.
Emme and Masie are the first layers of a larger platform. This is the whole shape, listed in build order, which is not the order the data moves in.
Emme Record
Masie
Masie Guidelines
Emme Careplan
outcomes
back into Masie
Outcomes will flow back into Masie, and the loop closes. Evidence improves care, care generates evidence.
Masie
Stage 1 · In developmentUniversities · research hospitals · guideline bodies · HTA and payers
Claim-level evidence credibility. A multi-agent system being built to score the evidence behind each scientific claim, keep that score current as new papers, retractions and replications arrive, and show where guidelines and evidence diverge.
- Designed to draw on public sources such as PubMed, retraction registries, replication databases and trial registries.
- No patient data, by design.
- The first research question is whether the scoring can be calibrated and reproduced well enough to rely on.
Masie Guidelines
Stage 2 · NextGuideline councils and committees
Living clinical guidelines, licensed to the bodies that write them. When the underlying evidence moves, Masie will draft the proposed update for the committee to ratify, rather than pushing a change unilaterally.
- A reconciliation layer will flag each guideline as concordant, discordant, or in a gap against current evidence.
- No patient data, but because it drafts clinical recommendations it will carry clinical decision support exposure.
Emme Careplan
Stage 3 · Later, regulatedClinicians
Personalised, living care plans generated at the point of care from the intersection of one patient’s record, current evidence and current guidelines, recomputed whenever any of the three changes.
- Where today’s care plans are built from population averages, this is built for the specific person in front of the clinician.
- It would be regulated as a medical device, so it follows clinical validation and clinician sign-off.
- Sequenced after the evidence engine has earned trust.
In the pilot. An early draft-and-sign-off version is in the Emme pilot.
Emme Record
Stage 4 · Later, isolatedPatients
The self-sovereign health record. A patient-owned lifetime record, birth to death, held by the person and shared with their care team. It is the data origin that makes the whole loop possible, and the product every other layer exists to serve.
- Because it holds identifiable data it will be walled off from the other products, with its own deployment and its own certifications.
- Built on open health data standards, with a patient-held consent ledger.
- It will contribute only tokenised, consent-gated signal into Masie.
In the pilot. An early patient-held timeline with consent-based sharing is in the Emme pilot. The encrypted record is being built.
Why the order looks backwards
The products that touch no patient data are built first, so the heaviest obligations, identifiable records and medical-device regulation, are taken on last and in walled-off layers.
Get in touch
If this is your problem too, we would like to hear from you.
We are talking to clinicians, researchers, universities, research hospitals, guideline bodies and investors. Tell us which side you came for.