The knowledge healthcare already uses.
We have been building clinical reference material for 16 years. What began as an ECG course is today one of the most used and most cited knowledge sources for physicians, nurses, biomedical scientists, ambulance personnel and students.
- The year we started
- 2010
- Published articles
- 523
- Clinical areas
- 35
- Calculators and risk scores
- 795
The same stubborn work since 2010
We started in 2010 with a single conviction: that clinical knowledge deserves better than scattered local protocols, outdated compendiums and PowerPoint files circulating between wards. Since then we have step by step built a single knowledge base that holds the same standard whatever area you look up.
We produce content for courses in hospital care, in primary care and at universities. The knowledge base is used as official course material at university hospitals and in medical education. The English-language edition has grown into the world's most used ECG course.
What makes it possible is not the volume but the discipline: the same editorial requirements for every article, the same requirements for source and level of evidence, and a version history that makes every change traceable. The site is maintained continuously — nothing is published and then left.
From the resuscitation room to the first term
Hospitals
Course material in clinics and in-house training, from the emergency department to the specialist ward.
Primary care
A reference work at the consultation desk and a basis for continuing education at health centres.
Universities
Official course content on the medical programme and on nursing and biomedical scientist programmes.
Vocational training
Ambulance personnel, biomedical scientists and specialist nurses in further training.
Four rules that are never departed from
They are simple to state and costly to follow. That is precisely why they are worth something.
A level of evidence on every recommendation
A recommendation without a strength is an opinion. Every recommendation carries its grade and its source, so that you can tell the difference between what is established and what is merely reasonable.
The source stays in the text
References appear where the statement appears, not in a list at the bottom that no one can trace back. If it cannot be substantiated, it is not there.
Written for the clinical decision
The text is ordered by how a work-up actually proceeds: what you see, what you rule out, what you do. Not by how a textbook is arranged.
Version history on everything
Every change to every article is saved. It is always possible to see what was there before, when it was changed and why.
The site is maintained and written by practising clinicians
Araz Rawshani is the founder and original principal author of the material, and the person responsible for keeping the site up to date. He is a senior consultant in cardiology at Sahlgrenska University Hospital and a research group leader at the Wallenberg Centre for Molecular and Translational Medicine at the University of Gothenburg.
He defended his doctoral thesis on socioeconomic aspects of diabetes and cardiovascular disease and has since published more than 125 scientific papers in cardiovascular medicine, diabetes and obesity. Between 2018 and 2023 he was registry holder for the Swedish Register of Cardiopulmonary Resuscitation. He supervises some forty doctoral students, postdoctoral researchers and master's students.
Alongside the clinical and scientific work he leads the AI strategy at the Centre for Digital Health at Sahlgrenska, focusing on decision support and prediction models in cardiology. That work is also the reason this site uses AI agents in its editorial flow — always as a draft, never as the final word: every AI-generated text is reviewed by a human before publication.
The site is operated by Rawshani Research AB, which also runs ecgwaves.com and diabetes.nu.
EBM AI
An agent-based co-author that does what no single human has time for: read everything.
EBM AI is not a language model asked to write an article. It is a pipeline of specialised clinical-reasoning agents that work on a single topic in turn. The search agents query PubMed, PubMed Central, Europe PMC and DOAJ and retrieve full texts, not just summaries — a conclusion resting on an abstract alone is a conclusion without a methods section.
The review agents then sort and appraise the material one by one: study design, size, population, conflicts of interest and how well the finding holds up against the guidelines. They also read one another's assessments and may object, which is the whole point — a lone appraiser who is wrong is wrong in silence. Only then does the writing agent distil what remains into a clinical text with a level of evidence and a source on every recommendation. Thousands of papers per topic become a few paragraphs you can actually act on.
The most important step is the last one, and it is human. Nothing EBM AI writes is published without a practising clinician having read it through, checked the sources and approved the text. The draft can be sent back any number of times. EBM AI is credited as a co-author on the articles it has helped produce, so that it is always apparent how a text came about, and every change is in the version history.
Does the whole department or course need access?
We supply accounts to departments, courses, universities and companies — with invoicing, and accounts activated within 48 hours.
Some icons on the site come from Icons8.