Mar 25, 2025
Lizzy Herzer
Dr. Manfred Criegee-Rieck from Klinikum Nürnberg and René Herzer from basebox are giving a joint presentation to IT decision-makers and physicians at the "Digital Hospital" event in Erfurt.
Dr. Manfred Criegee-Rieck from Klinikum Nürnberg during his presentation at the "Digital Hospital" event in Erfurt. He presents a pragmatic approach to AI implementation in the hospital environment.
How can artificial intelligence be safely implemented in hospitals? At the "Digital Hospital" conference in Erfurt, Dr. Manfred Criegee-Rieck from Nuremberg Hospital demonstrated a down-to-earth approach. The mathematician and IT director dispelled the notion that AI projects must inevitably be complex.
Four Pillars for the Digital Future
"Shaping the future of medicine together" – this is the motto behind Nuremberg Hospital's digitalization strategy. It rests on four pillars: digitally supported decisions, standardized processes, networked systems, and machine-readable data.
Cloud or Own Server? The Crucial Question
The choice of operating model shapes the entire project trajectory. While cloud solutions attract with low initial investments starting at 10,000 euros, in-house systems offer decisive advantages despite higher startup costs of at least 50,000 euros: data remains on-site, operating costs are predictable, and the hospital retains full control over updates.
From Simple to Complex
Criegee-Rieck advocates for a pragmatic approach: "Start with easily accessible data such as files and Office documents." Only later should more complex systems like the hospital information system be integrated. The integration effort increases from a few weeks to up to one and a half years.
The Maturity Test for AI Applications
Not every use case is suitable for getting started. The IT expert recommends prioritization based on AI maturity:
Knowledge management: Finding and structuring information
Administration: Translating, documenting, writing applications
Automation: Scheduling appointments, managing resources
Decision support: Image analysis, diagnostics
Regulatory requirements, integration effort, and risk potential should always be kept in mind.
Four Steps to Success
The Nuremberg roadmap for AI integration is compelling:
Define Hosting-Option
Create technical foundation
Develop use cases with teams
Gradually connect more data sources
It is crucial to involve data protection and security officers from the very beginning. A core team consisting of a project manager (half position), an IT administrator (quarter position), DPO/ISO (5 percent), and two to three pilot users with 10 percent of their working time each is sufficient.
basebox: AI Management System for Clinics
In the second part of the event, René Herzer presented basebox. basebox implements the pragmatic approach technically and offers flexible hosting options with GDPR-compliant data processing. Initial productive applications are possible within four to eight weeks.
You can watch the complete presentation by Dr. Manfred Criegee-Rieck and René Herzer as a video here.
Dr. Manfred Criegee Rieck, Daniel Zboril, Bavarian Association of Statutory Health Insurance Physicians, Lizzy Herzer, René Herzer, basebox
Stay Up to Date