AI in medicine
MedTech
AI as doctor and forecasting model
- AI or Doctor: Who Patients Will Choose?
- “Depends on where AI is applied,” states Sanjay Aneja, a Yale MD, who interviewed 926 patients.
- “Depends on where AI is applied,” states Sanjay Aneja, a Yale MD, who interviewed 926 patients.
Sanjay Aneja
Assistant Professor within the Department of Therapeutic Radiology at Yale School of Medicine
1
AI or Doctor
Who Patients Will Choose?
Medical Visualisation
55% of patients are ready to allow AI to "read" a chest X-ray.
▪️ Of these, 12% would feel very comfortable doing so,
▪️ 43% would feel somewhat comfortable.
Diagnostics
Most patients believe that a doctor, not AI, should diagnose serious illnesses, such as cancer.
▪️ Only 6% would feel very comfortable,
▪️ and 25% would be somewhat comfortable if they were diagnosed with cancer by AI.
▪️ Of these, 12% would feel very comfortable doing so,
▪️ 43% would feel somewhat comfortable.
Diagnostics
Most patients believe that a doctor, not AI, should diagnose serious illnesses, such as cancer.
▪️ Only 6% would feel very comfortable,
▪️ and 25% would be somewhat comfortable if they were diagnosed with cancer by AI.
Procedures
There is no data on how patients would feel about AI participation in procedures.
On the one hand, AI helps. When a 63-year-old New Yorker came in for a colonoscopy, AI was able to “see” what the doctor missed - a polyp with precancerous cells.
On the other hand, there was a known case in the USA where AI regulated the supply of drugs to a patient in a coma but due to reprogramming the patient received the wrong drug and died.
Management
The area where AI is welcomed is management.
88% would approve AI making appointments. 86% would favor AI scheduling and workload management.
There is no data on how patients would feel about AI participation in procedures.
On the one hand, AI helps. When a 63-year-old New Yorker came in for a colonoscopy, AI was able to “see” what the doctor missed - a polyp with precancerous cells.
On the other hand, there was a known case in the USA where AI regulated the supply of drugs to a patient in a coma but due to reprogramming the patient received the wrong drug and died.
Management
The area where AI is welcomed is management.
88% would approve AI making appointments. 86% would favor AI scheduling and workload management.
2
Forecasting model
zoonotic viruses
Zoonotic viruses.
"Zoon" in ancient Greek means "animal," "nosos" means "disease."
Zoonotic diseases are those that are transmitted to humans from animals. Under the data of the United Nations Environment Program, they account for: 60% of all known and 75% of all "new" infectious diseases.
Zoonotic diseases are those that are transmitted to humans from animals. Under the data of the United Nations Environment Program, they account for: 60% of all known and 75% of all "new" infectious diseases.
Of the zoonotic viruses, scientists isolated aggressor viruses with similar properties. Their genome was structured in such a way as to "circumvent” immunity. Or "cheat" it by mimicking the human one.
Artificial intelligence detected 313 such virus species. Of these, 272 were classified as dangerous and 41 as very dangerous ones.
Then, the scientists identified the so-called circle of hosts, the animal vectors. They were non-human primates.
Although these conclusions must be confirmed by laboratory tests, we can already say that one of the most promising tools for pandemic prediction may have been found.
Artificial intelligence detected 313 such virus species. Of these, 272 were classified as dangerous and 41 as very dangerous ones.
Then, the scientists identified the so-called circle of hosts, the animal vectors. They were non-human primates.
Although these conclusions must be confirmed by laboratory tests, we can already say that one of the most promising tools for pandemic prediction may have been found.
3
Google AI
What if AI can be a researcher or even a doctor?
AI-powered dermatology
Actually, it happened. Now, AI-powered dermatology assist tools research the largest human organ - skin. It is based on machine learning technology and currently analyzed over 65 000 images.
Google's AI-powered dermatology assist tool was classified as a Class I medical device in the EU. According to Peggy Bui and Yuan Liu this web-based application will be available for general use in 2023.
Google's AI-powered dermatology assist tool was classified as a Class I medical device in the EU. According to Peggy Bui and Yuan Liu this web-based application will be available for general use in 2023.
One of the main problems of dermatology - explain your symptoms by describing what a person sees in his/her skin. AI-powered dermatology assist tool enables to search diagnosis by taking a photo of patient skin, hair, and nails from 3 different angles.
Then patients should answer some base questions such as skin type, and how long they have had symptoms. As a result, AI gives different possible diagnoses and answers to frequently asked questions.
Warning:
AI-powered dermatology assist tool cannot replace medical doctors or suggest medical treatment since further examinations such as biopsy or blood analyze needed to be sure about the patient's diagnosis.
AI-powered dermatology assist tool cannot replace medical doctors or suggest medical treatment since further examinations such as biopsy or blood analyze needed to be sure about the patient's diagnosis.
If you program or produce similiar technologies, you can apply for the Medtop: Tech Award
You can read other articles here:
We find, award and provide digital marketing tools for healthcare industry members in order to help them become known worldwide
Germany, Berlin, Gross-Berliner-Damm 153, 12489
email us: pr@medtop.org
call us: +49 15735982125
Germany, Berlin, Gross-Berliner-Damm 153, 12489
email us: pr@medtop.org
call us: +49 15735982125