AI medical diagnosis is interesting. Having observed a big city emergency room, do diagnostic medical robots assume people are answering truthfully?
A private company with a string of health service contracts is to launch a national scheme which allows patients to receive a full diagnosis by smartphone – without ever having to see a GP.
Babylon Health has just begun a pilot scheme which means patients in five boroughs of London are encouraged to consult a chatbot instead of a human being, when they contact the 111 non-emergency line.
Under the system, patients key in their symptoms, with artificial intelligence used to assess the urgency of each case, and determine whether users should be told to go to A&E, a pharmacy or tuck up at home.
Now the company’s chief executive has revealed it is to launch a more sophisticated model which will allow any individual to receive a diagnosis by smartphone.
Dr Ali Parsa, the company’s founder said the system would allow doctors to work in tandem with artificial intelligence – so that medics could focus on treating rather than diagnosing diseases.
The entrepreneur said: “There are 300 million pieces of knowledge that we have collected.
“No human brain can do that. This is the largest amount of primary care clinical semantic knowledge in the would that is held by any computer, as far as we know.”
The model remains in development, but tests so far have shown it is faster and more accurate than the doctors in risk assessing cases, Dr Parsa said.
In the coming months, research will test the thesis that it can also outperform medics in making a full diagnosis. So far, trials have found it can do so in all abdominal diseases, the company said.
“I think we will soon be able to diagnose more accurately and faster than a doctor in most cases. That leaves the doctor to focus on the management of the diseases,” Dr Parsa said. …
Its app, which triages cases – making a risk assessment of urgency – is already available free to consumers in any part of the country.
Consumers can pay £25 if they need a webcam consultation with a private doctor, or subscribe to the service for £5 a month for unlimited access.
More than 190,000 consumers have signed up for the scheme, with 120 corporate clients – including the companies Boots, Bupa and Sky – providing free subscriptions to staff. …
“Everybody has a mobile phone,” Dr Parsa added. “This way you get access to a doctor 14 hours a day, seven days a week in your pocket.”
… tests comparing speed, accuracy and safety of the artificial intelligence system showed the computer consistently outperforming the human.
Tests comparing accuracy of triage forund that nurses results were accurate in 73.5 per cent of cases, while doctors achieved accuracy levels of 77.5 per cent, while the computer reached rates of 90.2 per cent. …
While doctors took an average of 3 min 12 seconds to make a diagnosis and a nurse took 2 minutes 27 seconds, the computer took 1 minute 7 seconds.
The findings came from a panel of senior doctors who judged the accuracy and speed of diagnosis after the event. …
“If you think of the game of chess – no person can beat the machine – but the best games come when chess players are assisted by machine,” he said.
It was impossible for any medic to retain the levels of knowledge required to perform at the highest levels, he said.
“There were 11,000 papers published in dermatology last year – doctors need to be able to harness all that information; it’s about making humans focus on what they do best.”
There’s also Watson:
“Watson, the supercomputer that is now the world Jeopardy champion, basically went to med school after it won Jeopardy,” MIT’s Andrew McAfee, coauthor of The Second Machine Age, said recently in an interview with Smart Planet. “I’m convinced that if it’s not already the world’s best diagnostician, it will be soon.”
Watson is already capable of storing far more medical information than doctors, and unlike humans, its decisions are all evidence-based and free of cognitive biases and overconfidence. It’s also capable of understanding natural language, generating hypotheses, evaluating the strength of those hypotheses, and learning – not just storing data, but finding meaning in it.
As IBM scientists continue to train Watson to apply its vast stores of knowledge to actual medical decision-making, it’s likely just a matter of time before its diagnostic performance surpasses that of even the sharpest doctors.