Healthcare everywhere
Watson’s oncology expertise is now being used in a software-as-a-service (SaaS) format in Thailand, India, Korea and China. In Singapore, Watson has also been tested in intensive care units, where it uses real-time data from vital sign equipment to help nurses predict which patients are prone to sepsis and organ failure, said Nakhooda.
But because it is device-agnostic, Watson can also be used for numerous healthcare applications outside of a clinical setting.
“To me that’s where it gets very exciting and powerful—you could use any device in the home as a mode of communication, with Watson behind the scenes analyzing the information,” said Nakhooda.
IBM’s researchers in Japan, for example, have developed a personal assistant device designed to chat with elderly people in their homes. The goal: to use the conversation to detect early signs of dementia, and thus allow for early interventions, Nakhooda explained.
When specialists are scarce
Watson’s oncology applications have started to gain traction in the medical community, especially now that studies are being done to validate the system’s ability to recommend treatments, said Nakhooda.
For example, a study conducted by IBM and collaborators at India’s Manipal Comprehensive Cancer Center found a high concordance between Watson’s recommendations and those of a multidisciplinary tumor board of experts. Man and machine agreed 96 percent of the time on treatments for lung cancer, 81 percent for colon cancer and 93 percent for rectal cancer.
Doctors see a role for Watson in countries where specialists are hard to come by, says Nakhooda. India, for example, has one oncologist for every 1,600 patients. “The ability to give this tool to a more junior oncologist or to another doctor at least gives the hospital the ability to do something to treat the patient.”
Watson is now a domain expert on cancer. But the Holy Grail, thinks Nakhooda, lies in the ability to recommend treatments for all manner of ailments. To appreciate how complex this problem is, consider that while a cough could be due to something as innocuous as the common cold, it might also herald the start of cancer.
“My ultimate dream would be for Watson to be used in developing countries, where there’s very little clinical support. A nurse could put in some very basic information and get something back on what could be good for the patient,” said Nakhooda.
For something like this to happen, Watson will have to learn and apply our entire corpus of medical knowledge—no mean feat even for a cognitive computing system.
This article was first published in the print version of Supercomputing Asia, July 2017. Click here to subscribe to Supercomputing Asia in print.
———
Copyright: Asian Scientist Magazine.
Disclaimer: This article does not necessarily reflect the views of AsianScientist or its staff.
