Why AI Will Not Replace Your Doctor and Lessons From a 2021 Clinical Experiment
- 21 hours ago
- 4 min read
Hilaryano Ferreira is a cardiologist and the founder of Aethera Health, an EU-licensed telehealth platform delivering physician-led online care across Europe. He works in cardiology and co-authored early peer-reviewed research on using AI to automate remote cardiac monitoring.
Years before chatbots made artificial intelligence a dinner-table topic, a small clinical team in Portugal taught software to read cardiac device reports. What they learned about the right division of labour between machines and physicians still defines how responsible telehealth is built today.
What did AI in healthcare look like before ChatGPT?
It is easy to forget how recently artificial intelligence became fashionable in medicine. In May 2021, about 18 months before ChatGPT was released and large language models entered public consciousness, AI in healthcare was mostly the quiet work of research teams solving narrow, unglamorous problems: reading scans, flagging arrhythmias, and sorting paperwork.
The paperwork mattered more than most people realise. Research published in the Annals of Internal Medicine found that, for every hour physicians spent with patients, they spent nearly two additional hours on electronic records and desk work. Medicine did not have an intelligence shortage. It had an administration surplus.
That was the problem a group of Portuguese clinicians set out to test in 2021, in a study whose lessons would later shape AETHERA Health.
What was the 2021 experiment?
Patients with pacemakers and other cardiac implantable electronic devices generate a steady stream of follow-up reports. Reviewing them is essential, yet, for the majority of routine cases, it is slow, repetitive work that consumes clinicians’ time without requiring clinical judgement.
In research presented to Europe’s electrophysiology community and published in EP Europace, a team led by Hilaryano Ferreira, then working in cardiology at Hospital Prof. Doutor Fernando Fonseca in Amadora, asked whether that review could be automated. Among his coauthors was Tchitchamene Nelumba of the paediatric cardiology department at Hospital de Santa Cruz in Lisbon. Today, they are the cofounders of AETHERA Health.
In collaboration with a Canada-based data science company, the team built a software application that combined optical character recognition with artificial intelligence and natural language processing to read device reports automatically and extract clinically relevant information. In testing, the system processed thirty device reports in under five minutes.
The conclusion was careful, as good clinical research should be: the algorithm showed that it was possible to facilitate and automate the remote follow-up of cardiac implantable electronic devices, with the potential to improve efficiency, reduce unnecessary in-person visits, and free clinicians for the work that actually needs them. The full abstract is available in EP Europace.
Why does a small study from 2021 still matter?
It was a modest experiment. Thirty reports, one workflow, no grand claims. Nobody involved would say it changed medicine.
But consider the timing. In May 2021, remote care was still widely treated as a pandemic stopgap rather than a permanent layer of the health system, and clinical AI was a specialist pursuit rather than a boardroom priority. Two working clinicians spending their evenings teaching software to read device reports were, in a quiet way, somewhat ahead of their time.
More important than the timing was the boundary the study drew. The software read, extracted, and organised. It never diagnosed, never decided, and never treated. Every clinical judgement remained with a physician. That boundary, tested with real cardiac data in a real hospital, became a design principle.
How should AI actually be used in patient care?
The lesson from that experiment translates directly to how modern telehealth should work, and it can be summarised in one sentence: automation should absorb the administrative weight of medicine, and physicians should retain every clinical decision.
In practice, at Aethera Health, that means artificial intelligence structures each patient’s health assessment, screens for contraindications, and prepares a clinical summary before a consultation. What it never does is diagnose, prescribe, or decide. A licensed physician reviews every case and approves treatment where appropriate. The machine reads and organises. The clinician judges.
This is not caution for its own sake. Cardiovascular disease alone is estimated by the European Heart Network to cost the EU economy around 282 billion euros a year, and no algorithm addresses that burden by replacing doctors. The gains come from giving clinicians their time back: less transcribing, more treating.
What should patients take from this?
Plenty of health platforms have added AI to their stories in recent years. A more useful question than “Does this platform use AI?” is “What is the AI allowed to do?” If the answer is that it drafts, sorts, and summarises while a licensed physician makes every clinical call, the technology is working for the patient. If the answer is unclear, that is worth pausing on.
Aethera’s founders were publishing on clinical AI before the platform existed, and that history shapes how the product behaves today: ambitious about automation, conservative about clinical authority. Somewhat pioneering, then. They were not the inventors of AI in medicine, and they would be the first to say so. But they were early, practical, and clinically grounded, which, in healthcare, counts for more.
Experience physician-led telehealth for yourself
If you would like to see what AI-supported, physician-led care feels like in practice, you can take a free health assessment that takes under five minutes with Aethera Health. Every assessment is reviewed by an EU-licensed physician, in keeping with the physician oversight principle described in this article.
Read more from Hilaryano Ferreira
Hilaryano Ferreira, Cardiologist. Founder of Aethere Health
Hilaryano Ferreira is a cardiologist and the founder of Aethera Health, an EU-licensed telehealth platform based in Portugal. The platform provides physician-led online care in weight management, cardiovascular health, men's and women's health, and longevity, supported by AI-assisted triage. He has over a decade of clinical practice. Besides founding Aethera, he works in cardiology and co-authored research published in EP Europace in 2021 on using artificial intelligence to automate remote follow-up of cardiac devices. That early work shaped the principle Aethera is built on: automation absorbs the administrative weight of medicine, while licensed physicians keep every clinical decision.
Reference:
Ferreira HDS, Ferrer F, Cabanelas N, Nelumba T, et al. “Artificial intelligence: an attempt to automate remote device follow-up.” EP Europace, Volume 23, Issue Supplement 3, May 2021, euab116.517.
