As IBM Moves Forward, Study Finds Promise in Watson Health Decision-Making

In a recent study, IBM Watson Health found artificial intelligence (AI) improves medical decision-making for cancer treatment. Findings suggested that physicians changed their treatment decisions in 13.6% of cases based on information provided by the AI system Watson for Oncology.

The supercomputer support system was designed to provide physicians with the most up-to-date treatment options for patients. IBM hoped it would become an innovative tool to transform cancer treatment, even as news reports of shortcomings in IBM’s AI for healthcare and resultant skepticism surfaced.

The company partnered with the Memorial Sloan Kettering Cancer Center and other clinics to train the system to offer top-notch professional judgement by processing over 20 million pieces of literature in seconds.

“In oncology, there is so much data and so much literature, especially in the last few years,” said Mehool Patel, M.D., associate chief medical officer for Watson Health Oncology and Genomics. “Each and every week, there’s so much coming out, in really a breathless fashion, that it’s going to be very hard to ingest all that information for one single person. So part of the ethos of IBM Watson is to really see how we can improve the clinical flow of a clinician who’s treating a patient who really needs cutting-edge care.”

The findings came from a blinded evaluation of 1,000 breast, lung and colorectal cancer patients in India at Manipal Hospitals. Physicians who changed their treatment decisions saw newer treatment options appear in 55% of cases and more personalized alternatives in 30% of cases. In 15% of cases, Watson for Oncology provided new insights from genotypic and phenotypic data and evolving clinical experiences.

Patel said the results of the study further highlight the benefits and potential of Watson for Oncology.

“That’s really why I think this study is very critical because even in centers that are considered expert places in their nation, they are making a significant change. You know, 1% you can say is maybe not much, but about 14% is a large number of patients to have a change in recommendation,” Patel added.

Researchers presented their findings at the annual American Society for Clinical Oncology late last month. IBM Watson Health and its partners also presented 21 other studies featuring their progress in AI technology. The other studies include one finding that suggests Watson for Oncology improves cancer patient confidence and another that found Watson for Genomics surfaces new insights for oncologist treating patients with hematological malignancies.

While the potential benefits of this technology are promising, there continues to be skepticism on the true impact of Watson for Oncology. Previously, Stat News revealed internal IBM Watson documents that showed the technology has offered incorrect cancer treatment advice. Generally, this has left some questioning whether IBM Watson Health has lived up to its promise. Particularly, the larger question of whether AI could ever computerize medical intuition remains to be seen.

Navigate the digital transformation with confidence. Register for our newsletter.

Related

What IBM Watson Health Hopes to Accomplish Through $50M in Deals with Brigham & Women’s, Vanderbilt

IBM Watson Health Releases Diabetes App, Touts Power of Data

Podcast: The Artificial Intelligence Question