We look forward to our next publications:
- Dataset for sepsis detection in emergency rooms #OpenData
- CDS models for interpreting procalcitonin, detecting AKI under vancomycin treatment, and identifying hypokalemia
Let’s continue the path to open-source #ClinicalAI together! For more information and contact with our team, visit https://www.ampel-cdss.com !
3/3
-New and exciting were our appearances in national media, such as MDR, WDR, Süddeutsche, Spektrum, and Deutsches Ärzteblatt
Our #OpenSource publications:
- Multicenter #Dataset of over 5 million blood samples for #Sepsis detection -> https://github.com/ampel-leipzig/sbcdata
- #CDS #AImodel for early sepsis detection based solely on a complete blood count -> https://github.com/ampel-leipzig/sbcmodel
- Prospectively studied CDS models for detecting #RefeedingSyndrome and hypo- and #Hypercalcemia -> https://www.uniklinikum-leipzig.de/einrichtungen/ampel/modelle/%C3%BCbersicht
2/3
When we started our initiative this year to launch an #OpenSource #ClinicalAI, we expected it to start slowly. Then events quickly accelerated. What a fantastic start!
Our highlights from the first 10 months of Open Source #AMPEL:
- HIMSS'24 in Rome: Together with pioneer Bart de Witte, we discuss open source in European healthcare
- @Mer__edith (CEO of Signal) promotes AMPEL during the award ceremony for the Future Prize of the Helmut Schmidt Foundation
1/3
The #SaxonHospitalDay motto was 'Shaping the Future: Digitalization and AI in #Healthcare'. The FAQ of the day:
• What is most important for operating a #CDSS (Clinical Decision Support System)?
Digitized and easily accessible data
• What differentiates individual CDSS the most?
The respective scientific proof of actual benefit, i.e., the existing #ClinicalEvidence
• What makes AMPEL-CDSS so special?
#AMPEL goes #OpenSource, maximizing transparency, innovation and collaboration
What happens when you throw Google, the innovators of the Sheba Medical Center and German lighthouse projects into the atmosphere of the Pittlerwerke in Leipzig?
A big compliment to the organizers, who succeeded in creating an open space for ideas! Many thanks to all the speakers, it was great fun. The future of medicine can come! 😉
Following 'From the clinic for the clinic,' we have created #AMPEL, a university hospital #AI that provides real added value.
We are excited to present our system tomorrow at "Shaping the Future of Medicine" #STFOM. On Thursday, we will continue at the #SaxonHospitalDay
As an #OpenSource initiative, we aim to make not only the software but also scientific evidence and best practices transparent.
Will you be in Leipzig in the coming days? Feel free to stop by for a chat with us!
So trivial and yet somehow quite cool: The first #Spektrum (#ScientificAmerican) article to report on its own work.
https://www.spektrum.de/news/sepsis-fataler-amoklauf-des-immunsystems/2225505
For all those who want to know (without a paywall) what our #Sepsis early detection AI is all about, I recommend our homepage: https://www.uniklinikum-leipzig.de/einrichtungen/ampel/en/models/sepsis
But one thing is for sure: the interest in transparent and jointly developed #CDSS structures is overwhelming. We have already infected 14 university hospitals with this idea and the number is growing! This gives us enormous momentum for the further development of our open source #AMPEL!
3/3
The CDSN intends to create scientific evidence, i.e. quickly enabling studies about which AI algorithms really do help with treatment and diagnosis in everyday clinical practice.
This is a big step. Only through multicenter collaboration can #CDS move from theory to practice. What we are building: innovative, interoperable and transparent structures in order to have the most innovative AI algorithms in clinical care.
Now it's time to keep our fingers crossed for our application!
2/3
How can we use AI in hospitals to better handle future pandemics and other challenges?
The answer is: in emergency situations, we fall back on existing structures. And those do not yet exist for #AI applications.
Together with a total of 14 university hospitals, we have therefore submitted an application for the first Clinical Decision Support Network (#CDSN) in Germany.
1/3
As the new spokesperson for the core dataset ‘laboratory report’ of the Medical Informatics Initiative #MII, Martin Federbusch now represents science towards industry and politics in the standardisation of digitised #LabData.
With the introduction of #AI in clinics such as #AMPEL, the scientific secondary data of today will become the primary data of tomorrow. It is therefore crucial that the conditions for both training and application of AI systems are right.
Today we had the AI expertise of 7 university hospitals with us. This scientific exchange in person is so important. The great discussions about how to bring #OpenSource #AI into clinical reality would hardly be possible otherwise.
With such a new topic without real examples to follow, one relies on expert knowledge. Our #AMPEL project benefits from this in multiple ways. We want to shape this open and transparent future of #MedicalAI!
We are already looking forward to the next time!
"Detecting critical situations in real time:
At Leipzig University Hospital, for example, various systems equipped with AI are already being used in different areas. The so-called "AMPEL" - a clinical decision support system - helps to detect the rare "refeeding syndrome", for example. [...] Accordingly, more than half of all cases detected by the "AMPEL" are overlooked without the system."
Translated from: https://www.sueddeutsche.de/wissen/wissenschaft-einsatz-von-ki-veraendert-arbeitsweisen-an-kliniken-dpa.urn-newsml-dpa-com-20090101-240426-99-815178
- real-time montoring of #medicalAI use will bring about more assurance (Rachel Dunscombe, #openEHR International)
- thresholds (causation and damage) as to when we have a defective AI system must be established along with the system (Yiannis Tolias, DG Sante European Commission)
- the AI Act is a framework act, it needs more to come to fruition. Medical and legal practitioners will have to fill it with life! (Petra Wilson, #HIMSS)
What do you think about these points?
From other sessions, this one surprised me the most: "EU #AIAct - How can it pave the way to better AI regulation globally?" Why? Because irrespective of the title, the speakers made excellent points:
- training medical personel is crucial, they have to have AI-literacy (Jessica Workum, chief data officer - artificial intelligence at Elosabeth-TweeStedan hospital)
- who is liable in adaptive AI? Well, all agreed that currently it's the medical personel, but responsibility must be shared!
Open Source Medical AI and Summary of #HIMSS24 in Rome: First of all, what a buzzing event!
The EU Medical Device Regulation was not made with AI and #OpenSource principles in mind. So, instead of following a trodden path, we must pave our own. We at #AMPEL-CDSS are developing a clinical decision support system that is AI-ready, evidence-based, interoperable and, most of all, Open Source. Public funds started us, so we are giving back to the community.
Our journey at https://ampel-cdss.com
Program tip for today's HIMSS in Rome: Starting at 11am, #OpenSource AI in healthcare will be discussed. As AMPEL (https://www.ampel-cdss.com), we are pleased to be able to present our point of view there 😎
"The Hitchhiker's Guide to AI...and Thanks for All the Fish" (11:00 AM, Room Bologna)
Meredith Whittaker, president of Signal and winner of the Bundeskanzler-Helmut-Schmidt Foundation's Future Prize, speaks from our souls:
"To open, accountable and shared core infrastructures that, instead of being abandoned and extracted, are supported and nurtured."
The #CDSS AMPEL is exactly that in the medical sector. Let's get together!
https://www.helmut-schmidt.de/aktuelles/detail/die-rede-der-zukunftspreistraegerin
Organised by the EU-funded SHIFT-HUB project, we will present our #AMPEL - Clinical Decision Support System here next week.
Free registrations: https://www.digitalsme.eu/events/shifthub-online-demoday-leipzig/
