Photo: Getty

Internal company documents from IBM show that medical experts working with the company’s Watson supercomputer found “multiple examples of unsafe and incorrect treatment recommendations” when using the software, according to a report from Stat News.

Stat reviewed documents that were included in two presentations given in June and July 2017 by IBM Watson’s former deputy health chief Andrew Norden. The documents were reportedly shared with IBM Watson Health management.

Advertisement

According to Stat, those documents provided strong criticism of the Watson for Oncology system, and stated that the “often inaccurate” suggestions made by the product bring up “serious questions about the process for building content and the underlying technology.”

One example in the documents is the case of a 65-year-old man diagnosed with lung cancer, who also seemed to have severe bleeding. Watson reportedly suggested the man be administered both chemotherapy and the drug “Bevacizumab.” But the drug can lead to “severe or fatal hemorrhage,” according to a warning on the medication, and therefore shouldn’t be given to people with severe bleeding, as Stat points out.

A Memorial Sloan Kettering (MSK) Cancer Center spokesperson told Stat that they believed this recommendation was not given to a real patient, and was just a part of system testing.

Advertisement

According to the report, the documents blame the training provided by IBM engineers and on doctors at MSK, which partnered with IBM in 2012 to train Watson to “think” more like a doctor. The documents state that—instead of feeding real patient data into the software—the doctors were reportedly feeding Watson hypothetical patients data, or “synthetic” case data. This would mean it’s possible that when other hospitals used the MSK-trained Watson for Oncology, doctors were receiving treatment recommendations guided by MSK doctors’ treatment preferences, instead of an AI interpretation of actual patient data.

And the results seem to be less than desirable for some doctors.

“This product is a piece of shit,” a doctor at Florida’s Jupiter Hospital said to IBM, according to the documents reviewed by Stat. “We bought it for marketing and with hopes that you would achieve the vision. We can’t use it for most cases.”

Advertisement

That doctor was reportedly one of many whose complaints were included in the internal documents.

Within days of when Norden gave one of these presentation, Gizmodo spoke with an oncologist at Jupiter Hospital for a report on the overzealous hype and shortcomings of Waston. During the interview, which was arranged by IBM Watson Health, Shah said Watson sometimes served as an extra opinion when Jupiter doctors could not agree on treatment. Shah did not provide a ringing endorsement or the sort of harsh criticism that the unnamed Jupiter doctor apparently candidly shared with IBM executives, as shown in the internal documents.

Responding to Stat’s report, a IBM spokesperson told Gizmodo that Watson for Oncology is trained to help oncologists treat 13 cancers and is being used by 230 hospitals around the world, and has “supported care for more than 84,000 patients.”

Advertisement

“At the same time, we have learned and improved Watson Health based on continuous feedback from clients, new scientific evidence, and new cancers and treatment alternatives,” the spokesperson said. “This includes 11 software releases for even better functionality during the past year, including national guidelines for cancers ranging from colon to liver cancer.”

Norden told Stat he could not comment since he is no longer working for IBM. He left weeks after the presentations to work for Cota, a heath care data-analytics company that had partnered with IBM.

Memorial Sloan Kettering spokesperson Caitlin Hool told Stat that the criticisms in the internal documents are a reflection of “the robust nature of the process” of developing software like Watson.

Advertisement

“While Watson for Oncology provides safe treatment options, treatment decisions ultimately require the involvement and clinical judgement of the treating physician,” Hool told Stat. “No technology can replace a doctor and his or her knowledge about their individual patient.”

[Stat News]