August 2, 2018 | Written by: Deborah DiSanzo
Categorized: AI | Blog Post | Oncology & Genomics
Share this post:
I would like to provide an update on Watson Health and address a few misperceptions. If you only read certain media reports recently, you would have an incomplete and inaccurate perspective, especially as it relates to our Watson for Oncology product.
IBM launched Watson Health three years ago with a clear goal: find the best ways to bring AI and other technologies to bear to help medical and health professionals tackle the world’s biggest healthcare challenges. Across the various areas of Watson Health our efforts are having an impact, for example:
- helping people with diabetes better manage glucose levels,
- informing clinical decision-making related to cancer with a suite of oncology products, and
- supporting important new findings that could accelerate new treatments for diseases like ALS and Parkinson’s
Specifically in Watson for Oncology, we set out to achieve a very ambitious objective: leveraging the power of AI and data to inform the daily decisions doctors make in treating cancer. Today, 230 hospitals and health organizations around the world use our oncology and genomics offerings, supporting care for 84,000 patients as of the end of June 2018 – almost double the number of patients at the end of 2017.
We work closely with our training partner Memorial Sloan Kettering, which ranks treatment options for Watson for Oncology based on their leading expertise. Watson is currently trained on a total of 13 cancers, chosen purposely to address the broadest number of patients who are suffering with those diseases. That training includes MSK oncologists creating representative cases based on real patient scenarios that allow for a range of treatment choices. This approach leverages their real-world expertise, based on evidence, and includes industry changes on ongoing basis. For example, Watson for Oncology is being trained on cases based on practice-changing breast cancer research coming out of the American Society of Clinical Oncology meeting in June. Historical patient cases wouldn’t reflect this advancement.
We are 100% focused on patient safety. It’s important to remember this tool does not diagnose. It provides treatment options to physicians, supported by evidence. Ultimately the treatment decision is always up to the doctor and patient. While a recent media report created the misimpression a potentially harmful treatment was recommended to a doctor in the field, the option cited was actually part of internal testing being done to improve Watson for Oncology, not a real patient. We also have an extensive quality management protocol through which we assess and address feedback and items that have been flagged by client, partners, and our own teams and systems.
IBM has issued 11 software releases in the past year to enhance Watson for Oncology’s functionality. We have increased our science and research, which now includes more than 50 peer-reviewed papers and abstracts on our AI offerings. We are working closely with medical professionals to ensure Watson continues to evolve with their needs and technical processes, and we are localizing our products to meet international needs where treatment options may differ from the U.S.
IBM knew at the outset the healthcare industry is unique, complex, local and personal for all involved. We also believed three years ago – and still do today – that AI can make possible what was previously unthinkable. And we know IBM, working closely with partners and clients, is uniquely positioned to deliver on that promise.