- €17 million to transform hospital prescriptions: OPTIMABIO brings together AP-HM, Hospices Civils de Lyon, Limoges University Hospital and Kiro in order to integrate AI directly into the heart of hospital software.
- A potentially broader impact than diagnosis: the algorithm analyzes clinical data in real time to recommend the most relevant biological tests, limit unnecessary prescriptions and improve the quality of care.
- AI is becoming a native function of the information system: rather than an additional application, OPTIMABIO develops hospital software into decision-making assistants integrated into doctors’ daily lives.
- University hospitals become co-producers of artificial intelligence: their clinical expertise, protocols and repositories now constitute a strategic asset as important as the AI models themselves.
- A new generation of digital hospital: By optimizing millions of prescribing decisions, AI could generate more value for the health system than the automation of a few specialized medical procedures.
The OPTIMABIO project aims to improve one of the most frequent hospital decisions, namely the prescription of biology exams. This is a project supported by the Assistance Publique-Hôpitaux de Marseille (AP-HM), the Hospices Civils de Lyon, the Limoges University Hospital and the French startup Kiro, and has just received funding of more than 17 million euros as part of the France 2030 iDémo program.
At first glance, the project seems less spectacular than artificial intelligence capable of detecting a rare pathology or interpreting medical imaging, yet its impact could be much more significant. Because behind the optimization of biological prescriptions, a major evolution of the digital hospital is taking shape: artificial intelligence is no longer a specialized tool, used occasionally, to become a function integrated into the heart of hospital software. The challenge is no longer just to produce a better diagnosis, but to improve, in real time, thousands of clinical decisions made every day by caregivers.
Prescription, a new area of conquest for artificial intelligence
Unlike other specialties, medical biology concentrates several characteristics which make it a particularly favorable field for artificial intelligence. The data is massively structured, the medical standards are largely codified and the best practice recommendations are regularly updated. Above all, biology is involved in nearly 70% of diagnostic decisions, which gives it a central place in the care pathway.
In this context, the objective is to improve the relevance of each prescription. The algorithm developed as part of OPTIMABIO will analyze the available biological and clinical data in real time in order to identify the examinations most suited to the patient’s situation. He may indicate that an identical test has already been carried out, recommend a more relevant examination or draw attention to an inconsistency with medical standards. The AI therefore acts upstream of the decision, without ever replacing the clinician.
This approach addresses a major economic problem. The OECD estimates that around 20% of health spending is linked to inadequate care. An unnecessary prescription is not just the cost of an additional test. It mobilizes hospital resources, extends treatment times, makes unnecessary demands on laboratories and can lead to a cascade of additional tests. Reducing these prescriptions therefore amounts to acting simultaneously on the quality of care and the efficiency of the hospital system.
AI ceases to be an application and becomes a function of hospital software
The real change introduced by OPTIMABIO is mainly due to its mode of integration.
For the past ten years, most digital health startups have been marketing applications or platforms in addition to existing tools. This overlapping of software has often constituted one of the main obstacles to their adoption: multiplication of interfaces, breaks in the user journey and poor integration with hospital information systems.
OPTIMABIO adopts a radically different logic where the recommendations generated by artificial intelligence are intended to be directly integrated into the software used daily by doctors. AI no longer constitutes additional software but becomes a native capability of the hospital information system.
This evolution is reminiscent of that observed in the world of enterprise software. After offering autonomous assistants, major publishers have gradually integrated artificial intelligence directly into their ERP, CRM or office suites.
University hospitals become co-producers of artificial intelligence
The project also reveals a more discreet transformation of the role of university hospitals. For a long time, healthcare establishments were mainly seen as testing grounds for startups. They provided data, validated clinical protocols and tested solutions developed by manufacturers.
OPTIMABIO reflects a different evolution, the three partner university hospitals do not only participate in an experiment, but bring their medical expertise, their biological standards, their prescription protocols and the knowledge accumulated by their teams of biologists. In other words, they contribute directly to the construction of the algorithm.
And this evolution is far from trivial, because as artificial intelligence advances, the value no longer lies only in the models themselves, but in the data, business rules and clinical expertise that allows them to be specialized.
Hospitals thus become producers of digital assets, capable of transforming their medical knowledge into software intelligence.
France 2030 finances infrastructure more than a startup
The 17 million euros allocated to the project as part of France 2030 support industrial capacity intended to be widely deployed in the French hospital system.
The objective is to build an infrastructure capable of sustainably improving medical practices.
A strategy that circumvents the main obstacles of medical AI
The positioning adopted by Kiro is also of strategic interest.
Most medical artificial intelligence startups have historically focused on uses directly linked to diagnosis. These applications involve high levels of responsibility, particularly demanding regulatory validations and strong reluctance from practitioners when they appear to compete with their expertise.
By intervening on the relevance of the prescriptions, Kiro chooses a significantly different terrain.
Artificial intelligence does not decide between two diagnoses. It offers contextualized recommendations, based on medical standards and results already available. The final decision remains entirely in the hands of the clinician.
This approach reduces medico-legal risks while facilitating acceptance by healthcare professionals. AI becomes a co-pilot of clinical decisions rather than a substitute for the doctor.
The Next Battle of Hospital AI
OPTIMABIO probably opens up a much larger market than medical biology alone. If artificial intelligence demonstrates its ability to improve the relevance of biological prescriptions, the same logic could be applied to other dimensions of the care pathway: imaging prescriptions, therapeutic choices, patient orientation, prevention of redundant examinations, planning of specialized consultations or optimization of hospital discharges.
In other words, hospital AI could gradually invade the field of clinical organization. Organizational decisions are infinitely more numerous than complex diagnoses, they also represent a significant part of hospital costs and room for improvement in the quality of care.
Artificial intelligence could thus produce more economic value by optimizing millions of daily decisions than by automating a few very specialized medical procedures.
A new generation of digital hospital
Several challenges remain, starting with data governance between establishments, interoperability with the numerous existing hospital software, the explainability of recommendations and the economic model after the public funding phase, all decisive steps.
However, these questions do not detract from the paradigm shift that OPTIMABIO reveals. Hospital artificial intelligence is entering a standardization phase. Phase during which it seeks to silently improve the decisions that structure each working day of caregivers.
Perhaps this is where its most transformative potential lies. No longer in the ability to compete occasionally with a specialist on a complex diagnosis, but in that of assisting, thousands of times a day, in the ordinary decisions which determine the quality, speed and cost of patient care. OPTIMABIO thus illustrates a fundamental trend: medical AI ceases to be an exceptional tool to become an invisible, but essential, component of the daily functioning of the hospital.
