Artificial Intelligence In Healthcare To Rescue The Health System
As our healthcare system reaches its limits, Artificial Intelligence In Healthcare is increasingly emerging as a tool that can help make it more effective by transforming the way we deliver care. At CHUM, she has already invested many specialties, but she remains simply an aid to the decision-making by the doctor, who will remain the main responsible for care for a long time.
 |
| Automated Machine Learning For Healthcare |
Currently, artificial intelligence (AI) allows you to analyze a photo of a mole on the skin taken with a smartphone, for example. Thanks to an image recognition algorithm, it will indicate whether this lesion has malignancy characteristics or not. The result obtained by the machine will help the doctor, general practitioner or dermatologist decide whether to speed up the care of a surgeon, in the event that the lesion is cancerous, or on the contrary, wait and see the patient a few months later if it is not cancerous.
Image recognition algorithms can also detect signs of diabetic retinopathy on a patient's eye fundus photographs, which will require consultation with the ophthalmologist. The comparison of several images taken at different times in time also allows to see if the lesions have evolved and therefore if the disease has worsened, therefore emphasizing the urgency of consulting an ophthalmologist at the earliest. On the other hand, if the background images do not reveal anything abnormal, in this case, the patient is reassured, and we will follow up by redoing a photo a few months later.
"AI thus reduces the number of visits to the specialist by detecting the most relevant cases. It reduces the concern of patients, or on the contrary guarantees them to be taken care of quickly since specialists are freed from the need to see all patients so that they focus only on suspicious cases, " explained Dr. Fabrice Brunet, president and CEO of CHUM, on the sidelines of the day of exchanges between Lyon and Montreal on the future of Health Systems, which took place on June 5 at the CHUM Research Center.
"For example, if you are worried about a mole, you will make an appointment with your dermatologist, but you will only be able to see it late because of the very long waiting list. The advantage of AI is that it directs the patient to the right resource based on the severity of the problem, which reduces waiting lists and optimizes existing resources, all of which will improve the health system. »
Artificial Intelligence In Healthcare
Accelerate Diagnosis
In general, tissues that are taken from a patient for analysis to determine if they are abnormal or even cancerous, are fixed on a glass slide. The pathologist then carefully examines them, counting the cells and identifying those that differ from others. "The AI can now perform a much faster and better analysis of these blades, and thus save diagnostic time. It allows us to use our pathologists more effectively, who can focus on more specific problems that require their expertise, " says Dr. Brunet.
 |
| Future Of Artificial Intelligence In Healthcare |
"AI increases our ability to analyze and make decisions, it improves the way we diagnose, detect, treat, predict a disease and accompany a patient," he points out while giving the example of the Radiological image of an organ that does not have known characteristics that testify to the presence of cancer.
On the other hand, "by correlating different sources of information about the patient, such as his genetic data, his lifestyle, his environment, as well as information that has been collected about him by health professionals, the neural networks [i.e. AI] will be able to make suspicious this image that, until then, seemed normal, and attract the attention of the radiologist, who will then make the decision to review the patient," adds Dr. Brunet.
Pristina, How To Use AI In Health
Artificial Intelligence has recently emerged as the ideal solution to transform the world of Health. There are different ways to use robots and AI to diagnose and treat patients.
A first real help of healthy AI is through chatbot, the exchange before the consultation; the use of RPA (Robotic Process Automation) in this case represents a real progress, since it increases the productivity of hospital agents by performing back office tasks. Voice dictation, now very effective, could also allow us to do without some 4000 medical assistants. AI also makes a more homogeneous supply of care, which helps to solve the problem of
Medical Deserts.
Another example of AI's success in health is the creation by General Electric of Pristina, a machine capable of detecting suspicious micro-calcification smaller than a grain of salt, and thus better detecting breast cancers.
How Does It Work?
In fact, the machine revolves around the patient to capture even blind spots, and thus outputs more reliable 3-dimensional photos (unlike the classic mammography that remains in 2 dimensions), which then allows better detection of cancer cells. Pristina then integrates AI by connecting the shots with a data bank, and comparing it with hundreds of thousands of others to assess the risk of cancer in the patient in question. If the machine detects a tumor, it will request accurate genetic, clinical and family data to identify the risk, avoid over treatment and enter the case into the database.
Pristina
For doctors, the challenge is to feed the machine with a maximum of information and pictures-controlled by a team of radiologists — in order to make the diagnosis more efficient than that which will be done by a general radiologist. The technique used is called “deep learning", that is to say that the AI machine uses computer algorithms for machine learning of recognition and comparison of forms.
This is a real customization of breast cancer screening that Pristina offers thanks to AI !
Ethical and Psychological Issues…
AI therefore offers tools to adapt behavior and treatment to patients ' lifestyles. But these tools are not an end in themselves, and to believe that the machines will cure us tomorrow would be inappropriate. Indeed, these tools will never replace professionals because they lack the psychological dimension, important in the treatment.
Finally, the use of Artificial Intelligence In Healthcare raises ethical issues, including the sharing of patients ' personal data. Establishing ethical rules of transparency is therefore a necessity to guarantee the confidentiality of care. Many data companies are already looking to buy back the radiological images produced by Pristina, so that they can develop new and even more efficient AI tools ... the question then arises: Should we refuse and defend our privacy or be open to help progress ?
No comments: