Healthcare AI Can Save Millions of Lives by Preventing Medical Errors
Diagnostic artificial intelligence has been publicly tested for the first time in Russia! Skychain Global ran a contest to compare its AI systems’ error rate to that of flesh-and-blood doctors.
Sadly, people make mistakes. Even in our modern times, diagnostic errors are a common thing in healthcare. Medical errors are the third leading cause of death in the United States, behind heart disease and cancer.
For example, when diagnosticians analyze X-ray lung images, they miss early lung cancer in 70 percent of cases. But artificial intelligence can detect it much more reliably! There are thousands of examples like that in healthcare.
Skychain Global has been developing a blockchain platform that can dramatically increase the efficiency and accuracy of healthcare AI systems by bringing together the key parties involved in creating such systems. Skychain’s central idea is to use smart contracts to enable secure transactions between healthcare big data providers, AI developers, and consumers.
Skychain publicly tested its prototype system on February 20, measuring the false positive and false negative rates when detecting the following diseases:
· Melanoma (based on dermoscopy images)
· Breast cancer (based on biopsy sample images)
· Heart disease (based on EKG)
Click the link to watch a short video of the event:
It should be noted that the AI systems’ error rates were from 4% to 14%, and the doctors’ error rates were from 18% to 32%. It’s easy to see that Skychain can easily halve the number of medical errors! That’s especially important for early disease detection. If widely used, healthcare artificial intelligence can save millions of lives a year all over the world! It took Skychain AI systems only about 0.1 seconds to make each diagnosis, while the doctors spent 20–30 seconds per case.
Skychain Global has invited some highly experienced doctors to take part in the contest, so we can be assured that their diagnostic skills are substantially higher than those of the average doctor. Here are their opinions:
Prof. Valery Volgin, Doctor of Medicine; dermatologist: “Your research and development is very important. It will be most useful for mass screening.”
Maria Ovsyannikova, first category doctor; anatomic pathologist, cancer pathologist, medical examiner: “It’s a state-of-the-art project that doesn’t have any rivals in healthcare. The project provides artificial intelligence for preliminary assessment of different medical conditions. It can be used both by patients and by doctors. An increasing use of such technologies in modern healthcare is necessary and inevitable.”
Maksim Golikov, top category doctor; anesthesiologist, emergency physician, functional diagnostician (EKG): “At the very least, such research and development, especially involving artificial intelligence, can help medical doctors in their work. Say, the computer produces a result that can draw the doctor’s attention, for example, to a specific EKG or other medical examination data. Then the doctor can take a closer look at that data and, possibly, arrive at a different conclusion.”
Maria Sadovskaya, top category doctor, dermatologist: “The doctors-vs-artificial-intelligence contest was kind of a new thing for us. The task wasn’t difficult for me, but it was very useful. The dermatologists have achieved the best results compared to other doctors, and I’m happy about that. I was only 2 points behind artificial intelligence at the end of the contest!”
Some other doctors attended the event, too:
· Natalia Solntseva, top category doctor; cardiologist, anesthesiologist, emergency physician, functional diagnostician (EKG)
· Vasily Krivopuskov, anatomic pathologist
Currently we are only talking about medical diagnoses made by AI on par with the best doctors in their field of medicine. But accumulating massive amounts of healthcare big data in the Skychain machine-learning ecosystem promises to bring unprecedented opportunities in the future. Among other things, AI can potentially evaluate the effectiveness of a medicine or a treatment regimen based on a billion case histories, discover new diseases, or develop the most effective treatment regimen for a specific patient. It might even be able to develop new medicines and medical treatment approaches! But if we want any of that to become possible, we need to accumulate vast amounts of healthcare big data in a single ecosystem and make all of that data easily available for neural network training. That’s the problem Skychain is going to tackle!
Skychain Global is starting to sell its tokens on February 26, 2018. Until the soft cap is reached, they will be offered at a 5% discount. Eventually, the tokens will be converted into Skychain’s internal currency, Skychain Global Coin (SGC).