Browse
Health Pages
Categories
Could a really smart computer that won a game show become your doctor in the future? IBM seems to think so. And why not? Watson possesses immense amount of information and can analyze them really quickly.

At least one in 20 American adults are misdiagnosed by their doctors. This was the shocking finding of a recent study. The researchers' conclusion was simple — policymakers, healthcare organizations and researchers should try to do something about that. What is the real reason behind the frequency with which patients aren't diagnosed correctly though?

Isn't part of the reason that doctors' knowledge is limited by definition? No matter how much experience a doctor has, or how much expertise, no doctor can ever have heard of all obscure disorders. This certainly becomes clear whenever you talk to someone with a rare medical disorder. Initial misdiagnosis almost seems inevitable for anyone who has a very unusual medical condition. 

Enter Watson, IBM's supercomputer. If you've heard of Watson, it's because the supercomputer won a game show. "Watson, the supercomputer that is now the world Jeopardy champion, basically went to med school after it won Jeopardy," Andrew McAfee, the author or The Second Machine Age, recently said in an interview.

He added: "I’m convinced that if it’s not already the world’s best diagnostician, it will be soon."
Dr Computer isn't just a futuristic idea: Watson has already been familiarized with dozens of medical textbooks, databases of medical journals, and thousands of medical journals. It's still learning. The MD Anderson Cancer Center already uses Watson to help recommend leukemia treatments, and the supercomputer was also found to be better at diagnosing lung cancer than humans.

Dr Computer — Would You Be OK With That?

Employing computers, especially really intelligent ones like Watson, in the diagnostic process makes sense. Watson isn't merely a data processor. Watson's artificial intelligence allows it to learn, to come up with hypotheses, and to weigh the symptoms of one condition against those of another — something it can do because it is basically able to store information about all medical conditions science currently knows about. 

Watson won't exclude certain conditions just because they are statistically really rare, and it won't come to conclusions without looking into all possibilities. Unlike a human, it will be able to look at many different diagnoses at the same time. It would never skip or forget anything. McAfee also adds that Watson doesn't suffer from human ailments like sleep deprivation, divorce, and stress.

Watson does not, in order words, not suffer from any human biases. It's much more likely to make an accurate diagnosis much more quickly. 

Making computers like Watson costs a lot, both in terms of construction and entering data. McAfee says that individual diagnostic tests will cost almost nothing once "Dr Watson" is already running, however. What's more, one single supercomputer could be used all over the world. You don't need to be in its presence to make use of its diagnostic abilities, after all. A smart phone and access to the internet would do it.

How? IBM describes the diagnostic process as following:

"First, the physician might describe symptoms and other related factors to the system. Watson can then identify the key pieces of information and mine the patient’s data to find relevant facts about family history, current medications and other existing conditions. It combines this information with current findings from tests, and then forms and tests hypotheses by examining a variety of data sources—treatment guidelines, electronic medical record data and doctors’ and nurses’ notes, as well as peer-reviewed research and clinical studies. From here, Watson can provide potential treatment options and its confidence rating for each suggestion."

Would human doctors become completely obsolete if a supercomputer came into use for diagnostic purposes? Of course not! Having Watson's diagnostic help available would give doctors more time to actually treat patients, rather than attempting to diagnose them.

Sources & Links

Post a comment