Equity by construction
Intuitive Pathos builds physiological monitoring that measures each person against their own baseline — not against a population mean that was never built to include them. Fairness is a property of the mathematics, not a promise in a policy document.
See it for yourself
A resting heart rate of 92 bpm sits comfortably inside the textbook adult range. For one of the two people below, it is the clearest warning of the day. Drag the reading, or jump to a scenario.
The problem
Nearly every physiological threshold in routine use is a population statistic: a range derived from some cohort of people, at some point in time, under some set of conditions. When the person in front of you resembles that cohort, the range works. When they do not, the range quietly stops being a measurement and starts being an assumption.
The failure is not that clinicians are careless. It is that the instrument itself encodes a comparison the clinician did not choose and often cannot see. A patient whose resting physiology sits at one end of a distribution can deteriorate substantially and still read “normal.” Another can be flagged, worked up, and labeled for being exactly who they have always been.
Both are errors of the same kind: a person compared against a distribution they were never part of.
This is not a hypothetical: reference-range and sensor bias are documented in the peer-reviewed literature — see the disparity case.
A number can be normal and still be wrong.
Normal for the population. Wrong for the person in front of you. Every reference range in routine use is capable of both at once.
The framework
These are not values we aspire to. They are constraints the system is constructed from — which is why we call it equity by construction.
01 — the founding constraint
Statistical distance from a population mean is not, by itself, pathology. A body that runs differently from the average is reporting a fact about a distribution, not a defect in a person. Our systems are built so that distance-from-the-mean is never sufficient grounds for an alert.
02 — your reference range is you
Every subject accumulates a within-subject reference distribution: what their physiology does across time, rest, activity and recovery. Change is then measured against that, so the question the system answers is always “what is different for you?” rather than “how unlike other people are you?”
03 — equal and opposite
Departures above baseline (Yang) and departures below it () are treated as equal and opposite. Monitoring systems have a long habit of instrumenting one direction well and the other poorly — agitation over withdrawal, hyperactivity over shutdown, the loud failure over the quiet one. Symmetric treatment removes that asymmetry at the level of the model, before it reaches a threshold or an alarm.
04 — some harms should be unrepresentable
A protection enforced by policy can be waived, misconfigured or forgotten. A protection enforced by architecture cannot. Where a harm is serious enough — comparing a person to a demographic cohort, moving identifiable physiology off a device, letting a non-clinical product emit a clinical conclusion — we make it something the system has no way to express, rather than something it is asked not to do.
Symmetry
Most monitoring is asymmetric by accident. The signals that get instrumented are the ones that are easy to see, easy to chart, and historically the ones that disrupted a ward. The person who becomes withdrawn, slowed, flat or unresponsive generates less noise — and so, in a system tuned by noise, less attention.
Symmetric fairness means the model does not have a preferred direction. A given magnitude of departure from a person’s own baseline carries the same weight whether it points up or down.
End to end
Continuous physiological signal from whatever sensing a deployment already has — bedside monitors, wearables, veterinary telemetry, field kit. No new hardware requirement is baked into the approach.
Build a within-subject reference distribution over time. This is the comparison set. It belongs to one subject and is never swapped for a cohort average.
The Wu mathematics primitives — memory coefficient, critical slowing down, λ* criticality, sliding-window regime detection — look for a system changing state, which is a different and earlier question than “has a number crossed a line?”
Output is expressed as departure from the subject’s own baseline, with elevated and suppressed weighted equally, and with the direction preserved rather than collapsed into a single severity number.
Model improvement is shared across deployments through a federation contract that is enforced in the type system. What may cross the boundary and what may never cross it is a structural property of the client, not a configuration setting.
Product family
The same primitives serve a critical-care unit, a veterinary clinic, a field team and a living room — because “compare this subject to themselves” is not a clinical idea. It is a measurement idea.
The regulated flagship: continuous level-of-consciousness and deterioration monitoring for acute and critical care. In FDA pre-submission preparation; not cleared or approved.
Your own physiology, on your own terms. A general-wellness application built on the per-subject baseline — with no clinical claim, by construction.
Pain and distress monitoring for animals, who cannot self-report and for whom the population-average problem is severe across breed, size and species.
Operator-state monitoring for high-consequence work, where the baseline that matters is the operator’s own and the quiet failure mode is the dangerous one.
Longitudinal wellbeing in the residential setting, oriented to gradual change against a personal baseline rather than to emergencies after the fact.
Occupational strain in industrial settings, and non-invasive stress monitoring for conservation and animal welfare research.
Where to go next
What a per-subject baseline changes at the bedside, what the regulatory posture is today, and what we will and will not claim.
Clinical overviewThe architecture of the moat, the product family, the layering strategy, and an honest statement of stage.
Platform thesisCapability statement, operator-state monitoring, data-sovereignty posture, and how to contract with us.
Capability statementThe idea in plain language: why being different from average is not the same as being unwell, and why that distinction is worth building.
The ideaWhy this exists
Every nurse has watched someone be told they were fine because their numbers matched a stranger’s. And watched someone else be treated as a problem for being exactly who they had always been.
Intuitive Pathos was founded by a bedside neuroscience nurse, not a lab. The framework came out of the specific, repeated experience of watching measurement instruments answer the wrong question — and of noticing that the error was systematic, directional, and fixable in the mathematics rather than in the vigilance of tired people.
Talk to us
Clinical partnerships, veterinary and research deployments, government contracting, or investment — the conversation starts the same way.
Regulatory notice. Products described on this site are in development. Pathos Vigil is investigational and has not been cleared or approved by the U.S. Food and Drug Administration or any other regulatory authority, and is not available for sale or clinical use. Non-clinical products, including Intuitive‑Me, are general-wellness offerings and make no diagnostic or therapeutic claim.
Not medical advice. Nothing on this website is medical advice, a diagnosis, or a treatment recommendation. If you have a health concern, contact a qualified clinician. In an emergency, call your local emergency number.