Visiomed, a company specialising in medical imaging, has just formed a strategic partnership with Medical Intelligence Service, the group of medical and technical experts behind MEDVIR, a revolutionary artificial intelligence system to provide medical decision support. Already used by SOS Médecins Paris and experimented on 400 patients from the emergency ward of Paris’s Lariboisière Hospital, MEDVIR has achieved an astonishing rate of 83% accurate diagnostics. We had a look at this system that’s so much more than just a symptom checker…
UNDERSTANDING HOW DOCTORS THINK
Have you ever wondered what’s going on inside a doctor’s head when you’re having a consultation? Dr Loïc Etienne, one of the people in charge of the MEDVIR project, attempted to analyse this on his blog.
It usually starts with the doctor listening to the list of symptoms, which can be extremely long and varied. The doctor’s aim at this stage is to determine whether anything requires urgent attention and to sort through the information to try and narrow down the number of possible diagnoses. To do this, they establish the order of importance of the symptoms and try to rule out as many possibilities as they can before carrying out a physical examination.
Then a decision must be made. When the diagnosis is sure or at least highly probable, it’s fairly easy. If, on the other hand, no obvious theory emerges, it’s a lot more complicated.
For Loïc Etienne and his employees, understanding and creating the way doctors reason during a consultation took nearly 25 years of research. The real challenge was working with fuzzy logic, i.e. an approach to computing based on “degrees of truth” rather than the usual “true or false”, and thus closer to the way human brains work.
AN expert systEM THAT ACTS AS A “virtual doctor”
MEDVIR – and e-docteur, the consumer version – can now ask questions that take into account previous answers. Tested and approved by doctors, it is based on 150 symptoms and can suspect over 500 illnesses. Its real genius lies in its ability to distinguish emergencies from conditions that can wait. And thanks to additional data provided by patients after the doctor has made a diagnosis, the system can constantly learn and evolve.
As far as Loïc Etienne is concerned, medicine 3.0 is already here. Whilst medicine 1.0 only included the doctor/patient relationship, medicine 2.0 also included social media, with more and more patients going on them before and after consulting a doctor (often to the despair of doctors).
With medicine 3.0, things have come one step further as a machine – an expert system like MEDVIR or a smart connected object – is added into the mix and helps build knowledge.
With artificial intelligence making faster and faster progress and connected health taking off, will doctors soon become obsolete? Hardly. Because whilst there are already a number of health 3.0 solutions on the market, they can’t replace a physician’s knowledge: they can just improve and modernise their everyday methods.