A French healthcare engineering company offers a modular organisational digital twin engine that converts existing operational data into executable models of patient flows, resources, capacities and costs. Combining process mining, discrete-event simulation and explainable machine learning, it compares scenarios, anticipates bottlenecks and supports evidence-based decisions. Partners are sought for pilots, integration, co-development and commercial deployment.
TOFR20260730011
Company and context : A French small company specialising in healthcare data science, industrial engineering and decision-support software offers a modular organisational digital twin engine for hospitals, clinics, healthcare manufacturers and public health organisations.
Problem addressed : Healthcare managers often have to decide whether to open or redesign a service, change staffing, invest in equipment, modify schedules or deploy a new medical technology without being able to test the consequences beforehand. Conventional dashboards describe what has already happened. Predictive tools usually estimate one outcome but do not represent the interactions between patients, staff, rooms, beds, equipment, schedules, costs and operating rules. Most healthcare digital twins focus on organs or individual patients rather than care delivery organisations.
Technology. The engine creates an executable virtual model of a care organisation from available operational data and expert knowledge. It combines process mining to reconstruct actual workflows; a monitoring layer that can be synchronised periodically or in near real time; discrete-event simulation to reproduce queues, resource use, delays and patient flows; explainable machine learning and statistical fitting to calibrate durations, probabilities and missing parameters; and simulation-optimisation to explore configurations and rank scenarios against user-defined indicators.
Deployment : The solution can first be used as a stand-alone impact model based on historical data, then evolve into a digital shadow or a connected digital twin through database or application programming interface connectors. It does not require a complete sensor infrastructure: partial synchronisation and periodic data updates are possible.
Applications : Use cases include emergency departments, operating theatres, bed and workforce capacity planning, outpatient units, maternity services, telemedicine, new buildings, territorial care networks, and the organisational and budget impact of medical technologies. Outputs may include waiting times, throughput, occupancy, staff workload, cost, return on investment, quality-of-service indicators and environmental impacts. An interactive web interface supports scenario definition, comparison, monitoring and export.
Cooperation sought : International healthcare organisations, technology manufacturers, software integrators and research institutions are sought for commercial agreements with technical assistance and/or research and development cooperation. A typical partnership starts with a clearly defined decision problem, data assessment and model calibration, followed by a pilot, user validation and optional integration into the partner information system. The desired outcome is a validated deployment in a new healthcare context, followed where relevant by replication, licensing or joint market development.
Advantages and innovations:
- Organisation-level focus: models the delivery system rather than only an organ, device or individual patient.
- Decision-oriented modelling: the model is built around a concrete management question and compares counterfactual scenarios before real-world implementation.
- Explainable hybrid architecture: process maps, rules, statistical distributions and simulation outputs can be reviewed with clinical and operational teams rather than relying on an opaque prediction model.
- Frugal use of existing data: historical extracts and partial synchronisation can be used without mandatory deployment of new sensors.
- Progressive deployment: the same core supports a one-off organisational and budget impact model, a periodically updated digital shadow, or a near-real-time digital twin.
- Multi-criteria evaluation: quality, waiting times, capacity, workforce, costs, return on investment and sustainability indicators can be analysed together.
- Modular software: generic simulation, calibration, optimisation and visualisation components are adapted to each workflow, reducing development time while preserving local relevance.
- Scientific traceability: assumptions, data transformations, parameter calibration, random seeds and scenario definitions can be documented for reproducibility and sensitivity analysis.
Expected role of a partner:
Healthcare organisations : Define a concrete use case, appoint an operational sponsor, provide pseudonymised event, resource and cost data, co-define scenarios and indicators, validate the model with frontline teams and host a 6- to 12-month pilot.
Software or integration partners : Support secure data extraction, database or HL7/FHIR connectivity, hosting and integration into existing systems.
Research institutions : Contribute independent validation, comparative studies, publications and joint funding applications.
Cooperation model : Under a commercial agreement, the provider configures and calibrates the engine, trains users and supports deployment. Under a research and development cooperation agreement, the partners jointly define the protocol and demonstrator and agree intellectual property and exploitation rights before the project.