International Business Machines Corp. IBM -1.44% is exploring a potential sale of its IBM Watson Health business, according to people familiar with the matter, as the technology giant’s new chief executive moves to streamline the company and become more competitive in cloud computing.
IBM is studying alternatives for the unit that could include a sale to a private-equity firm or industry player or a merger with a blank-check company, the people said. The unit, which employs artificial intelligence to help hospitals, insurers and drugmakers manage their data, has roughly $1 billion in annual revenue and isn’t currently profitable, the people said….Watson was one of IBM’s highest-profile initiatives in recent years and a big bet on the growing healthcare sector, though results disappointed in part because physicians were hesitant to adopt artificial intelligence. While the effort made strides in areas including oncology and genomics, it never became the cohesive business IBM envisioned and has lost key executives in recent years.
I owe it to IBM Watson. Long time back, I got a brief on exploring its potential in healthcare. I was naturally curious to find out how it would disrupt the healthcare markets. Perhaps it was a trigger to this blog too.
I followed it up with a video call (and the executive on the other end was a radiation oncologist who had quit clinics and was giving me a demo). It was perplexing for her choice; I didn’t argue. I was more keen to observe the “holy grail” of the AI that was right in front of me- drop-down menus! She had to manually enter the required clinical information- I remained clueless about how IBM Watson served the purpose in the first case? The “recommendations” were no better than the NCCN PDF’s that are littered everywhere on the workspace of other colleagues.
IBM Watson only gave information for the breast cancer (it was the easiest to game) while the “recommendations” for other tumour types were missing. I didn’t get to see the genomics module (which was albeit more interesting) as there was no provision to get it locally. During the discussion, she also shared an example of a “leading hospital chain” that had invested in IBM Watson and was offering it as a stand-alone service to generate “recommendations on treatment course from experts and AI”. There were some papers shared about the “concurrence of opinion” between the AI and the “experts”.
Why it failed or why it was doomed to failure? I have written extensively about the hype that precedes the actual product. Interoperability between the EMR’s and open API’s are the key to success. IBM Watson had none. Besides, I have a distinct feeling that they had hyped the products to gain legitimacy for “reimbursements”- had they succeeded, it would have been a surefire recipe for policy makers but IBM would have, of course, made promises to “improve the product”. We know how promises from these corporations go.
It reminds me of another company called as NantHealth. They made me sign non-disclosure agreements that prohibit me from speaking about the product too. It was again doomed for failure, although the promise of “top-down integration” makes perfect business sense. However, established hospital chains won’t shift their EMR’s overnight to conform to an upstart- even though the owner was a fancy billionaire. These two failures remain a blot on what otherwise was a promising idea around integrating AI and “onco-mics” (oncolog+genomics). I have been looking for an integrated solution for a long time.
The central question is- would this solution actually help the patients? The answer depends on whom you are asking. For the hospital administrators, it is an excellent investment to show that they are “ahead of the technology curve” and the marketing brochures can scream about AI “serving the communities”. For the doctors, it is a tedious investment of time to navigate through another software with horrible UI and UX. For “scientists” in the domain, they would not wait with bated breath on how it is just another expensive proposition for a routine “Statistical inference”.
For the patients, it is a bearer of false hopes as they remain bereft of both preventive and palliative care.
As clinicians, we owe more than that as a moral and ethical responsibility to communicate in clear uncertain terms on what works (in black and white) instead of speaking generically in statistical uncertainties.