Watson Health: The Promise to Unlock Healthcare Data
“Technology is worth nothing if it doesn’t solve an important problem or improve lives.” – Dr. Robbie Pearl (CEO of the Permanente Medical Group)
Healthcare and technology have a strong joint history with seemingly limitless applications at their intersection.[1] Yet despite the rise of the first electronic medical record forty years ago, the healthcare world has yet to fully benefit from its digitization.
Policy as a Catalyst for Digitization
In 2009, Congress passed the Health Information Technology for Economic and Clinical Health Act (HITECH) to incentivize clinicians to adopt electronic health records (EHRs) and penalize non-users.[2] In the 7 years since this ruling, clinicians digitized how they practice medicine and invited the looming presence of computers into their intimate patient interactions, which came at a high cost to clinician satisfaction and productivity.
The rise of EHRs has forced clinicians to transform the way they practice medicine, yet many have yet to reap benefits of the data they have painstakingly captured. Today, clinicians are inundated with data from many sources (i.e., updated clinical guidelines, real-time data from wearables or remote monitoring devices, and EHR data), yet struggle to draw insights from this deluge of data, especially when making decisions in real-time. As many EHRs do not have robust analytics capabilities (especially when aggregating multiple datasets), technologists are flocking to healthcare to unlock the next frontier of big data.
Hello Healthcare, Says IBM Watson
IBM is one of those companies aiming to make use of all of this data, and Watson Health is the frontrunner in applying artificial intelligence to healthcare data. Watson Health delivers its cognitive computing abilities to “translate information into knowledge that can help drive more informed decision-making.”[3] Watson Health’s business model addresses three major gaps in the existing healthcare analytics market:
- Horsepower: Watson can read 200 million pages of text in 3 seconds, allowing clinicians to gather information on complex situations instantly.[4]
- Potential to unlock the elusive unstructured data: Watson can digest the free-text notes, images, and scanned documents that are inconsumable by average analytics software.[5]
- Ingestion of multiple datasets: Watson can aggregate data from various sources, which is helpful because it would take clinicians 29 hours each workday to keep up with new professional insights.[6]
To address the unique needs across the spectrum of care, Watson Health has tailored its technology to six specific care solutions: Watson for Oncology; Watson for Genomics; Population Health Management Solutions; Care Management; Imaging Solutions; and Enterprise Imaging Solutions. These existing use cases are just the tip of the iceberg in terms of value artificial intelligence is expected to bring to healthcare in the next 5 years.[7] This post focuses on how Watson for Oncology is already delivering unprecedented value in healthcare and offers more promise in the years to come.
Watson for Oncology
Watson is particularly beneficial to the field of oncology because of two characteristics: (1) new clinical trials, evidence, and therapeutics emerge daily; and (2) treatment selection is extremely time-sensitive. Watson Health addresses these by matching patients to clinical trials and recommending potential treatment options.
Matching Patients to Clinical Trials:
Clinical trials give patients access to potentially life-saving treatment and contribute to our growing body of cancer research. The current process for matching patients to applicable clinical trials is manual and slow. Dr. Steven Alberts (Chair of Medical Oncology at the Mayo Clinic) explains that Watson Health is able to accelerate this process and quickly identify relevant opportunities.[8]
Sharing Potential Treatments:
Watson Health also provides clinical decision support in the form of treatment plans. Alberts further describes: “In an area like cancer — where time is of the essence — the speed and accuracy that Watson offers will allow us to develop an individualized treatment plan more efficiently, so we can deliver exactly the care that the patient needs.”[9]
What Next?
IBM is placing a big bet on Watson Health being the first to apply artificial intelligence to an industry full with data but hungry for insight. In the next few years, I recommend Watson Health focus on two major areas:
- Watson in the Workflow: As mentioned previously, the care team is overburdened with technology. To ensure adoption, Watson Health should learn from missteps in EHR adoption and aim to better integrate its technology into existing clinical workflows, or even integrate more seamlessly with EHRs rather than have a distinct platform.
- Watson as a Training Tool: Watson Health has the ability to be a powerful training tool for new clinicians – showing them how care plans are determined based on historical clinician data and academic data. Watson Health has the ability to disrupt the current medical education and training programs in the US.
Wordcount: 794
[1] http://www.nytimes.com/interactive/2012/10/05/health/digital-doctor.html?_r=0#/#time15_361
[2] U.S. Department of Health and Human Services Centers for Medicare & Medicaid Services 42 CFR Parts 412, 413, 422 et al. Medicare and Medicaid Programs; Electronic Health Record Incentive Program; Final Rule
[3] https://www.ibm.com/watson/health/
[4] Ibid.
[5] Ibid.
[6] Ibid.
[7] http://www.forbes.com/sites/reenitadas/2016/03/30/top-5-technologies-disrupting-healthcare-by-2020/#52f4f2ef6252
[8] http://www.mayoclinic.org/giving-to-mayo-clinic/your-impact/features-stories/artificial-intelligence-gets-real
[9] Ibid.
Great read MB, I enjoyed learning more about how IBM Watson is effecting change.
From your suggestion about how to best implement Watson, what do you think of trying piloting implementation plans in underdeveloped countries, so as to ‘get it right’ while still making a meaningful impact? For instance, India has such a scarcity of physicians (their shortage in doctors is about 16 times greater than in the US) and prevalence of new cancer cases (>1 million per year). This would provide an excellent opportunity to understand how best to leverage the technology in an environment that is behind from an EHR and labor standpoint. And socially, you could expect it to be more helpful in India than say the USA because of the lower availability of resources and tech currently on the market.
Watson undoubtedly has huge potential to create meaningful value with health data; thanks for sharing your insights. The biggest issue right now is that Watson is before its time, specifically in the healthcare space. All of its ability is reliant on a wealth of data. Large hospitals may already be able to use Watson for a number of things, but the real value will be when it has access to much larger datasets.
Also, I do question how relevant Watson will be to health data whenever we have a fully function system of interoperability based on HL7 standards, in which case, a lot of Watson’s capability to tie together disparate pieces of data becomes redundant. That said, Watson’s capabilities extend far beyond this; I am intrigued to see how it will work in practice. My old company had a Watson machine at our annual conference a few years back, the excitement it is generating in the industry is incredible.
Using Watson in healthcare is very exciting. It looks to be useful in situations that return high probabilistic results. Much of healthcare cancer and disease treatment research is still unknown. I see Watson’s foray into this area as a big blank space that Watson can eventually fill in, so more of an opportunity than an unsolvable problem. How much trust will we bestow upon Watson in the future? How will future physician’s be trained? Healthcare doesn’t seem like a field we want to soley rely on Watson for all the answers.
http://www.cnbc.com/2016/02/19/can-ibm-fix-the-health-care-problem.html