Skip to main content
Minerva Clinical Intelligence
Minerva Clinical Intelligence
AI empowered clinical treatment

Treatment optimisation, not prediction

Diagnosis, triage and alerting tools are useful — but they stop before the treatment decision, and that market is now commoditised. A Minerva app recommends the next treatment action, continuously, as physiology changes. Reinforcement learning is literally the mathematics of agents: systems that choose actions over time to reach an outcome, rather than assigning a probability to a moment.

We pioneered off-line reinforcement learning and off-policy Policy Estimation methods for medical treatment enabling us to learn from operational real-world data. We developed grading based methods for privacy-preserving transfer learning enabling us to use a multitude of multi-modal data, differing from ward to ward, hospital to hospital, region by region, country by country. Our methods realise transform each patient's longitudinal data stream into digital twin, that we can interrogate and perform counterfactual and what-if questions to optimise decision making across a patient's entire trajectory.


Building one such agent is a research problem that we pioneered. Building a family of them, each safe enough to deploy in a hospital and defensible enough to clear a regulator, is a platform problem. That is what the Minerva AI Clinician Platform is for.


How the agentic loop works:
  1. Ingest clinical data 
    Vitals, laboratory results, prior fluids and vasopressors — 48 time-varying variables from routine care.
  2. Update the model & infer the patient state
    A trajectory-aware model of the patient with calibrated uncertainty, decomposed into what is noisy and what is unknown.
  3. Recommend the next action
    For Sepsis fluid management we need to prop up the patients circulatory system for them to survive: the AI, delivers an intravenous fluid rate and a noradrenaline-equivalent vasopressor dose recommendation — an action, not a probability.
  4. Clinician reviews
    The bedside clinician accepts, modifies or rejects. Advisory decision support, but designed for safety and intepreability at closed-loop standard.
  5. Audit and learn
    Feedback, safety monitoring and governance. Every AI recommendation and action takeb by clinicians is captured, the system learns on the job and provide accountability.

Risk-aware reinforcement learning under uncertainty

The policy inside each app is a risk-aware reinforcement-learning agent. For Minerva AI Clinician Sepsis it was trained on 96,000 patient records and externally validated on independent data from 128 US intensive care units. It is the first system in its class to combine a risk-aware policy with structural action-masking under uncertainty: the model decomposes aleatoric uncertainty (irreducible noise in the data) from epistemic uncertainty (what the model has not seen), and is structurally prevented from recommending when it should not speak.

Predefined safety constraints refine the policy’s behaviour without degrading its performance, and off-policy evaluation quantifies how the recommended strategy would have performed against observed care — the methodological backbone regulators expect for treatment-recommendation AI. Because these methods live in the core rather than in any one app, each new disease inherits them.

A medical device from the first commit

Our software is designed to IEC 62304 Class B with full requirements traceability, a maintained Design History File, calibration verification, a cybersecurity baseline, and summative usability evidence. The regulatory position — UK Class IIb, EU Class IIa, and a US pathway via De Novo. The quality system, safety case architecture and technical file structure are built once at the platform level. Each subsequent app extends an existing regulatory file rather than opening a new one from scratch — which is what makes a catalogue of regulated clinical apps tractable at all.

One integration. Many apps.

Each app addresses a treatment problem with the same shape: repeated, time-critical dosing decisions made under uncertainty, where the right action depends on where the patient has been, not just where they are now. That shape recurs across specialty after specialty — which is why the platform, not just the first app, is our company. The platform unifies 10 years of experience in building an AI Sandbox of our technology, testing it in multiple hospitals with different EHR systems (Cerner/Oracle and EPIC). 


Where the platform goes next

The first apps stay close to critical care, because that is where the data is richest, the decisions are most frequent and the evidence base already exists. But the underlying method — a risk-aware agent choosing a dose from a patient’s trajectory — is not specific to the ICU.

Later waves move into chronic disease, where the same sequential-dosing problem plays out over months rather than hours, and where personalised support between appointments barely exists today. The clinical question changes; the mathematics does not. What changes with it is the time horizon over which a decision’s consequences appear — and reinforcement learning is precisely the tool built for delayed consequences.

Each app is a distinct clinical problem needing its own evidence, its own validation and its own regulatory approval. What they share is the platform beneath them — and that is the difference between building one medical device and building a category.