“AI in healthcare isn’t magic. We see it as a practical and increasingly important building block for healthcare today and tomorrow,” said Martijn Franken, a clinical physicist at Bravis Hospital, during his keynote speech at Zorg & Facility. Three SAZ hospitals demonstrated there how they, in collaboration with the Healthcare Algorithms Expertise Center, are developing predictive AI for the ER, inpatient units, and outpatient clinics. What works, what doesn’t, and what are the key lessons?
Based on collaboration among hospitals, the keynote provided an honest look at what AI really means in practice. What became clear: AI is not used as a standalone technology, but as a practical tool that helps healthcare providers do their work better, more efficiently, and in a more patient-centered way. It should not be used as a passing fad, but as part of a collective movement toward future-proof healthcare.
While the focus was initially on medical applications, there is now much more emphasis on capacity planning and predicting patient flows. Think of applications that predict how busy the emergency department will be or what the likelihood is of a hospital admission or a no-show at an outpatient clinic. This can truly support healthcare providers in their decision-making.
See also: AI model predicts emergency room congestion
Learning and Developing Together
The keynote presentation at Zorg & Facility was delivered by representatives from three hospitals: Bravis (Martijn Franken), Elkerliek (Tessa Warmink), and Nij Smellinghe (Mirjam Vrieling), with the Expertise Center for Healthcare Algorithms (EZA) playing a key role. EZA was founded by the Association of General Hospitals (SAZ) with the goal of developing algorithms tailored to real-world practice, while ensuring that control over data and applications remains with the healthcare sector.
Within the SAZ, 28 hospitals collaborate and share knowledge through channels such as the AI knowledge network. This collaborative approach makes it possible to learn faster, scale up, and keep costs as low as possible.
Hospitals are working together to develop, test, and implement solutions, which allows for the immediate sharing of knowledge and faster implementation of improvements. Training materials, including animations explaining AI, are also jointly purchased and shared.
From Experiment to Scaling Up
Implementing AI requires more than just technology. Martijn Franken, clinical physicist at Bravis Hospital: “Technology changes rapidly, but processes and people need more time to adapt to changes; after all, healthcare providers demand reliable solutions that are carefully implemented and managed.”
Questions that healthcare professionals and healthcare innovators are grappling with include scaling up, collaborating within the care chain, and the practical application of AI. Successful implementation requires care, patience, and continuous improvement. A number of prerequisites are essential:
-
t
- AI Literacy: Employees Need to Understand What AI Does and How to Use It
- Data and Infrastructure: Reliable Data and Effective Storage Are Crucial
- Governance and Ethics: Clear Agreements on Use and Responsibility
- Validation and trust: Healthcare providers must be able to rely on the results
t
t
t
Also read: Tomorrow's healthcare professional: collaborating with AI
Data-Driven and Future-Proof Healthcare
Through collaboration, a focus on the front lines, and attention to implementation, hospitals are taking steps toward data-driven and future-proof care.
Instead of waiting for the perfect solution, hospitals are choosing to get started, test, and improve. This fosters a sense of community and aligns with the power of knowledge sharing.
Relatable Lessons from Real-World Experience
The key lessons from real-world experience are clear and relatable:
-
t
- AI is not a magic bullet, but a tool. AI supports, but does not replace
- Pay attention to AI literacy within your organization: people and processes require reliability and care, and therefore more time than technological development does
- A poor process with AI is still a poor process: so adjust your process
- Collaboration accelerates development and, as a result, improvement (and collaboration is just plain fun, too)
t
t
t
And perhaps Franken’s most important message to the audience: “Be proud to learn from one another. Share successes and learn from each other. AI isn’t magic. We must work together, learn together, and continuously improve—from, for, and by the healthcare sector.”
