Virtual wards scaled to more than 1 million patients.
ISO 13485 certified, SaaMD compliant.
A regulated AI platform that predicts patient risk and flags escalation, re-architected from a limited prototype into infrastructure for millions of patients and thousands of clinicians.
supported
using the platform
active patients
achieved
Client
A UK healthcare group running remote patient monitoring (virtual ward) services.
Goal
To improve and scale virtual wards during and after the Covid pandemic, so the platform could securely support millions of patients and thousands of clinicians.
The prototype worked for a handful of patients. It had to work for a million.
The client had an advanced prototype for virtual wards — remote monitoring that lets patients recover at home instead of in hospital. It worked, but only for a limited number of patients and clinicians, and it could not scale to meet the demand the Covid pandemic created.
At the core of the platform sat highly regulated, commercially sensitive algorithms driving triage and escalation decisions. Scaling them meant protecting the IP while proving, to a medical device regulator, that the software was safe. The organisation needed to win and maintain ISO 13485 certification and meet Software as a Medical Device (SaaMD) requirements for its data and technology — while accelerating delivery in the middle of a public health emergency.
Three constraints were non-negotiable:
- Prototype couldn’t scale beyond a limited number of patients and clinicians
- Core algorithms were commercially sensitive and highly regulated
- ISO 13485 and SaaMD compliance had to be proven, not assumed
Re-architected for scale, re-engineered for regulation.
We re-architected the platform as a secure, cloud-native system with AI decision-making at its core. The algorithms predict deterioration risk, recommend monitoring frequency and flag escalation pathways — with the clinician always making the final call. Data ingestion, AI inference and clinician-facing workflows were separated, with role-based access control and encryption in transit and at rest protecting both patient data and the commercially sensitive models.
The algorithms decide what needs attention. The clinician always decides what happens next.
Compliance was built into the engineering process, not added afterwards. We defined a secure product development lifecycle with formal verification and validation against ISO 13485 and SaaMD requirements, and introduced automated testing for AI integrity: unit and integration tests, synthetic and anonymised clinical datasets for regression testing, continuous model drift monitoring, and automated end-to-end tests for safety-critical decision pathways. Releases were gated behind sign-off criteria and full change traceability.
To handle scale, the platform ran on stateless microservices and managed cloud services, able to expand elastically from thousands of concurrent patient sessions to millions of patients at peak demand. Real-time telemetry and secure audit logging supported clinical governance and gave regulators the evidence trail they needed, while company processes were redesigned across product, engineering and clinical teams to embed change control and post-market surveillance.
A million patients later, still audit-ready.
The re-architected platform achieved ISO 13485 certification and met SaaMD compliance requirements, with automated testing and verification and validation pipelines giving demonstrable reliability and integrity at scale. It scaled from a limited prototype to more than a million patients and thousands of clinicians, handling thousands of concurrent active patients and expanding elastically for peak demand. On the strength of that transformation, the organisation was awarded further significant programmes to deploy virtual wards across multiple regions, benefiting thousands more patients with earlier intervention and reduced hospital pressure.
* Case studies reflect work undertaken by our Heads of AI either during their tenure with Head of AI or in prior roles before they were part of the Head of AI network; they are provided for illustrative purposes only and are based on conversations with our Heads of AI.
More case studies.
Your biggest pain point.
Fixed in 14 days. 50% off.
This started with one conversation. Book a 30 minute brainstorm call — we’ll plan your first AI project together and issue your 50% discount code. No payment today.
*Case studies reflect work undertaken by our Heads of AI either during their tenure with Head of AI or in prior roles before they were part of the Head of AI network; they are provided for illustrative purposes only and are based on conversations with our Heads of AI.