Before the birth of iCareX, we wondered how advanced medical practices applied today in hospitals around the world had been. We are first and foremost technology experts and, more often than not, we expect technology to travel hand in hand with science. What is the level of adoption of artificial intelligence by medical facilities? Are patients’ health data always accessible to physicians and is it correctly analyzed? Are the treatments administered to us personalized on the basis of our health profile?
These are just some of the questions we asked ourselves, the answers of which led us to a disconcerting truth: in order to protect everyone’s health, today’s medicine needs a large dose of innovation.
We understood that there is an important gap between current health protocols and the technological advancements of artificial intelligence. AI is widespread in every sector and has now reached a level of maturity that makes it ready to bring its incredible contribution to the healthcare sector as well. We believe that the initial concerns regarding the use of artificial intelligence algorithms – “black box” technology that, behind the processing of its dense neural networks, produces incredibly accurate but uninterpretable predictions – are nowadays outdated.
The algorithms that populate iTwinDiscover – iCareX’s end-to-end cloud platform for the advanced analysis of patient health data – have been made limpid through mechanisms that allow a clear interpretation of the models. In this way, we have chosen to give a coat of white paint to the “black box” of AI, making it fully transparent and understandable to all physicians.
In order for an artificial intelligence model to produce solutions that help treat the patient correctly, it must necessarily be fed with as much useful data as possible. However, we noticed that medical facilities are not equipped to provide physicians with all the patient’s health information. It took us very little to understand it, especially considering the fact that few or none of us succeed in the arduous task of delivering the complete ecosystem of his health data to his own doctor, which are scattered among a fairly large number of sources (mobile applications, smartwatches, wearable medical devices, analytical laboratories, alternative tests such as genetics or microbiome and hospitals different by the one where you are being treated).
So why do not integrate all these data into a single secure, configurable, dynamic and easily accessible database? Based on this idea, we decided to develop a privacy-first and GDPR-compliant channel – called iTwinSense – that would allow the patient to share their health data with trusted doctors, powering the iTwinDiscover platform and making sure it was always updated and constantly evolving. It seemed right to develop this mobile application – available in its first version from November 2020 – also to support patients, who in this way will have for the very first time the possibility of viewing in a usable way (punctual values, graphs, temporal variations, etc.) all their health data and have a clearer picture of their health: have you ever gone to the doctor and do not remember the result of a medical examination you have done or the name and dosage of a drug you have been treated with?
iTwinSense and iTwinDiscover are part of a common application – iTwin – that we created with the aim of bringing data and artificial intelligence to our best physicians. The names were certainly not chosen by chance: the term “twin” refers to a health concept that we consider extremely important, which is the “digital twin” that we all need and that we all would like to have. This is a digital representation of all our health data that allows the physician to have a complete overview of our health profile, helping him in the diagnosis of new diseases and in the development of more effective treatments. It is a revolutionary way of looking at medicine that uses two extremely important activities: the “Sense” or the perceptual collection of all health information that interest the patient in his daily life and the “Discover” that is the advanced exploration of data searching for correlations, trends and hidden patterns. We like to think that the patient’s health data is a cosmos of stars, within which we want to help the internaut physician to discover new planets and trace constellations never observed before.
Making data and artificial intelligence available to physicians has a very specific meaning for us, an important mission: to change the paradigm of standardized medicine by promoting a more complete personalization of the treatments administered to the patient. Today the same care protocol is applied to all patients, a bit like dressing everyone in one size suit even when we have a very different size from each other. Some will not be able to button their jackets, others will end up with pants that are too short or sleeves that are excessively long; but the point is we all need a tailored suit!
Medicine based on standard principles pushes physicians to deliver treatments that are too aggressive for some patients and too weak for others, and in this way ineffective for many: some people who die today could survive if personalized treatments were given to them. We are not the first to support it, as Hippocrates – the father of modern medicine – promulgated it 2500 years ago when he reminded everyone the need to: “give different ones [liquid medicines] to different patients, for the sweet ones do not benefit everyone, nor do the astringent ones, nor are all the patients able to drink the same things” .
All patients are characterized by extraordinary individual variability. Think for example of the nephrons or the functional units of the kidneys that work continuously to eliminate part of the waste present in the body. Generally, an individual who has a number of nephrons ranging between 200,000 and 2.5 million  is considered normal, a huge variability that can mean that one patient who has far fewer nephrons than another one can still be considered normal. Let’s imagine, therefore, that there are three people: the first with a number of nephrons equal to 200,000, the second with 1 million nephrons and the third with 2.5 million nephrons. Now suppose there is acute kidney damage (from undernutrition, hypoxia, or nephrotoxic drug) that neutralizes 100,000 nephrons. In the first individual, the damage corresponds to half of the nephronic patrimony, which means that he will have renal failure. In the second, however, the damage corresponds to a tenth of the total: there could presumably be damage to the kidneys. Finally, in the third, the damage is marginal, to the point of being negligible. It follows that, due to their basal variability, these three individuals will have to be treated with completely different approaches.
If Hippocrates had artificial intelligence available, today we would probably find ourselves faced with an extremely advanced medicine, able to personalize treatments for each individual and treat diseases with extreme ease without collateral damage for the patient. But it does not mean that this is not a feasible scenario, on the contrary we see a very near future taking exactly this direction.
Our AI algorithms allow us to direct the enormous wealth of knowledge of physicians towards the development of fully personalized treatments, the effects of which can be measured a priori through advanced clinical scenario analysis. Starting from current national and international guidelines, we have analyzed different types of diseases by identifying the parts of the process – mainly those that push the doctor and the patient to make a choice on the type of treatment to be administered – in which we can create value with our technology. In the next 5 years, we want to revolutionize the approach to treatment for over 100 diseases!
At iCareX we work every day to instill a new health culture, alongside the best physicians and for the wellness of all.
 Sykiotis, G. P., et al., Pharmacogenetic principles in the Hippocratic Writings. Journal of Clinical Pharmacology, 45: 1218-1220 (2005)
 John F Bertram, R. D., et al., Human nephron number: implications for health and disease, Pediatric Nephrology, 26, 1529-1533 (2011)