After its disappointing early stages in the 1990s, AI now seems to be on the right track, with its real-world applications multiplying. Naturally, the healthcare sector intends to benefit from it. An aging population, growing demand for care, increased life expectancy, an insufficient number of doctors, sustained demographic growth in Southern countries… the challenges are significant, and AI has a role to play. Insights.
Workload = high risk of medical error
France is often cited as an example for the quality of its healthcare. What is less often mentioned are the efforts made by healthcare professionals to meet their patients’ expectations, sometimes at the expense of their quality of life. Workload is a real issue in hospitals: long hours, emotional exhaustion, and workload linked to macro variables such as demographic growth and an aging population. To address this, the Ministry of Health launched a national strategy in late 2016 to improve the quality of working life for healthcare professionals, whether they work in hospitals or as independent practitioners.
While it is true that AI’s potential is extremely promising for medicine (as well as for other fields), the technology can cause some apprehension in a sector where errors can have dramatic consequences. Let’s remember that, to date, artificial intelligence retains a degree of mystery because the technology has not reached its full potential. In other words, “we don’t know” where the AI train will ultimately stop. Elon Musk, the renowned founder of Tesla and Space X, even released a documentary to warn about the “dangers” of AI. That being said, AI can already boast some real achievements in healthcare. Radiologists daily use AI technologies to process tens of thousands of images to further streamline diagnosis and decision-making.
AI and healthcare: restoring the quality of care
The flip side of spectacular technological advancement in healthcare is that current equipment produces 1,000 to 1,500 images per examination, 40 times more than thirty years ago. If we add to this all the macro-environmental factors causing the “demand” for healthcare to explode, we arrive at an astonishing average of 50,000 images passing before a radiologist each day. AI is therefore a decisive aid in reducing analysis times. But more is expected from this technology, because even if more advanced breakthroughs regularly make headlines, they often remain isolated and are frequently the preserve of upscale healthcare institutions. The role of startups in democratizing AI is crucial. Screenpoint, a young Dutch startup, for example, has developed an application capable of detecting breast cancers “very early.”
Other AI applications are more or less proven, but their availability remains anecdotal for now. These include assisted operations, remote patient monitoring, smart prostheses, data cross-referencing through big data, etc. However, there is an almost philosophical problem: the essence of AI lies in learning… and to learn, one must do, make mistakes, and correct them. The healthcare sector does not lend itself to this exercise. This also explains the “slowness” of AI solution deployment. Fortunately, the technology of the future places humans at the center of its concerns. Fascinating!
