BESAFE · Project

AI Software That Spots Risky Surgeon Behavior Before Accidents Happen

healthPrototypeTRL 3Thin data (2/5)

besafe-h2020.eu

Imagine a co-pilot for surgeons — not one that flies planes, but one that watches how a surgical team interacts with their equipment and flags when someone is about to make a dangerous move. The team took AI algorithms originally built for studying group behavior and adapted them to learn the difference between safe and risky patterns during surgery. They ran a feasibility study with a touchscreen demo showing the AI can sort users into risk groups just by watching how they tap and swipe. The goal is clinical decision support software that catches mistakes before they cause nerve damage or other surgical accidents.

By the numbers

EUR 100,000

EU contribution for feasibility study

consortium partners

1 year

project duration (May 2019 – April 2020)

total deliverables produced

The business problem

What needed solving

Surgical accidents caused by human error remain a major source of patient harm and hospital costs. Nerve damage during surgery is particularly devastating — it can leave patients with permanent disability and expose hospitals to costly malpractice claims. Current surgical monitoring tools track physiology but not the behavioral patterns of the surgical team that lead to mistakes.

The solution

What was built

The project produced two deliverables: a detailed business plan with financials, market analysis, and a roadmap of remaining technical milestones, plus a demonstration GUI showing the IBSEN-derived AI algorithms can analyze touchscreen interaction data and group users by behavioral risk patterns using unsupervised machine learning.

Audience

Who needs this

Intraoperative nerve monitoring equipment manufacturersHospital chains investing in surgical safety and quality programsHealth IT companies building clinical decision support for operating roomsMedical device startups looking for AI-powered differentiationSurgical training simulation companies

Business applications

Who can put this to work

Surgical technology and medical devices

SME

Target: Companies developing intraoperative nerve monitoring equipment or surgical navigation systems

If you are a surgical device manufacturer dealing with the challenge of reducing intraoperative accidents — this project developed an AI algorithm that learns from how surgical teams interact with equipment and flags high-risk behavioral patterns in real time. Built on unsupervised machine learning, it groups users by behavior without needing pre-labeled training data. The coordinator already has expertise in preventing accidental surgical nerve damage using electrophysiological technology.

Hospital management and patient safety

enterprise

Target: Hospital groups and surgical centers investing in quality improvement and risk reduction

If you are a hospital group dealing with costly surgical complications and malpractice claims — this project demonstrated AI that can detect high-risk actions by surgical team members through their interaction patterns with digital interfaces. The system uses unsupervised machine learning, meaning it adapts to your facility without extensive manual configuration. The 2-partner consortium included both an industry SME and a university to bridge clinical and technical expertise.

Clinical decision support software

any

Target: Health IT companies building OR workflow and decision support platforms

If you are a health IT company looking to add a behavioral safety layer to your operating room software — this project built and demonstrated machine learning algorithms that recognize high-risk vs low-risk behavior patterns from machine-user interaction data. The demo used touchscreen interaction to show unsupervised grouping of users by behavioral patterns. Integrating this AI module could differentiate your platform in a market increasingly focused on surgical safety.

Frequently asked

Quick answers

What would it cost to license or integrate this AI technology?

The project received EUR 100,000 in EU funding as a Coordination and Support Action, primarily for business planning and a technology demo. Pricing and licensing terms would need to be negotiated directly with Afferent Technologies SL, the coordinating SME. Based on available project data, the business plan deliverable includes detailed financials and market analysis.

Is this ready to deploy at industrial scale in a hospital?

Not yet. This was a 1-year feasibility and business planning project (CSA), not a product development effort. The demonstration used touchscreen interaction data to show the AI can group users by behavioral patterns, but clinical validation and medical device certification are still required. The objective explicitly mentions estimating remaining technical milestones to reach certifiable software.

What is the IP situation and who owns the technology?

The underlying AI algorithms come from the IBSEN FET project. Afferent Technologies SL, a Spanish SME, coordinated BESAFE to explore commercial exploitation of those results. IP ownership and licensing terms would need clarification from the coordinator, as the technology builds on prior EU-funded research.

What regulatory approvals would be needed?

The objective specifically mentions progressing to 'final certifiable software,' indicating medical device certification (likely CE marking under EU MDR) is required. Based on available project data, the business plan deliverable includes an estimate of remaining technical milestones needed to reach that certification stage.

How long before this could be used in real surgeries?

The project ran from May 2019 to April 2020 and produced a business plan and algorithm demonstration. The objective describes the remaining path to certifiable software as a key unknown that the business plan was meant to quantify. Clinical trials, regulatory approval, and product engineering would all be needed before deployment.

Can this integrate with existing surgical systems?

The coordinator, Afferent Technologies, already works with electrophysiological technology for preventing surgical nerve damage. The AI component was designed as an adjunct to existing surgical software, suggesting it could layer onto current systems. However, the demo was on touchscreen interaction data, so integration with actual surgical hardware remains to be developed.

Consortium

Who built it

This is a compact 2-partner consortium, both based in Spain — one industry SME (Afferent Technologies SL) and one university. The small team and EUR 100,000 budget reflect the project's nature as a Coordination and Support Action focused on business planning rather than heavy R&D. Afferent Technologies brings direct domain expertise in preventing accidental surgical nerve damage through electrophysiological technology, which grounds the AI work in a real clinical problem. The 50% industry ratio is healthy, but the single-country consortium and small scale mean limited market validation across European healthcare systems. For a potential business partner, this means the technology concept is credible but early-stage — the team was exploring whether the IBSEN AI results could become a viable product, not building one yet.

UNIVERSIDAD CARLOS III DE MADRIDparticipant · ES

How to reach the team

Afferent Technologies SL (Spain) — surgical nerve monitoring SME and BESAFE coordinator

Order a report on this consortium

Next steps

Talk to the team behind this work.

Want to explore licensing this surgical AI technology or connecting with the BESAFE team? SciTransfer can arrange an introduction and provide a detailed technology brief.

Order a report on this topic More in Health & Biomedical