DMC Bytes - Virtual Care, IBM Watson Health

Ensuring Quality in the Era of Virtual Care
By Kurt Herzer, Peter Pronovost
JAMA Network Viewpoint (Free Access) 

In this article, the authors use the 2001 Institute of Medicine framework for high-quality care (safe, effective, efficient, timely, patient centered, equitable) to assess how virtual care stacks up. 

Under Safe and Effective, the priority is to avoid harm and deliver evidence-based care. They point out that current virtual care for low-level complexity is generally accepted. Over time, however, virtual care could become more complex and help facilitate care coordination and effective treatment. The authors point out that there is limited high-quality evidence to ensure that virtual care does no harm and can achieve the same or better outcomes than traditional care. A statistic references that in the 2^nd quarter of 2020, almost 70% of primary care office-based visits had a blood pressure recorded but only 9.6% of telemedicine. (Note: I am not surprised by the telemedicine rate but why only 70% in primary care?). 

For Efficient and Timely the authors acknowledge that visits that do not require in-person evaluations could be quicker, avoid the costs of travel and limit time away from work (or other productive pursuits). Of course, they then identify the potential for unnecessary visits, more tests due absence of a physical exams and concern for medical liability leading to an increase the total cost of care. 

The Patient Centered and Equitable section raises the question as to whether virtual care will facilitate the appropriate level of engagement between patients and clinicians, especially as care expands to new diagnoses, difficult treatment decisions and sensitive topics. The authors also describe both positive and negative effects that virtual care could have on health care disparities. 

The article closes with three potential guiding principles. Virtual care should...
1. …achieve comparable safety and effectiveness as traditional care.
2. …a net increase in efficiency and not add to the total cost of care.
3. …be respectful of patient preferences and values and not exacerbate health care disparities. 

I encourage DMC members to review the details behind these principles. The authors reference development of clinical practice guidelines, new payment models, quality reporting, the potential adverse consequences to continuity, use of wearables…and so on. 

My take: I like the application of the IOM framework to virtual care and appreciate the challenges and opportunities outlined. As virtual care evolves, we need to be careful not to let comparisons to “traditional care” limit its potential. We need to optimize care for people based on their unique needs, preferences and availability of services using all available methods. That includes in-person care (which can also be very innovative), virtual care – or a combination of both. And all care should be safe, effective, efficient, timely, patient-centered and equitable. 

IBM’s Retreat from Watson
By Daniela Hernandez, Asa Fitch
IBM Explores Sale of IBM Watson Health
By Laura Cooper, Cara Lombardo
Wall Street Journal (Subscription required) 

 Two interesting articles about what was once thought to be a great example of how artificial intelligence (or augmented intelligence as a colleague from IBM used to say) would help solve clinical problems, reduce cost and unwarranted variation in healthcare services. This WSJ article focuses, as you might expect, on the business angle of this news: “…billed as a ‘bet the ranch’ move by Big Blue;” “IBM spent several billion dollars on acquisitions to build up Watson.” We all probably remember the Jeopardy IBM challenge and television advertisements with Bob Dylan.  

But did it improve health care? According to some, the hype exceeded the reality. This article from STAT identifies concerns with Watson for Oncology (2017) and MD Anderson’s Cancer Center canceled Watson after spending $60M in 2017. This Forbes article criticizes IBM for deciding not to facilitate or build APIs for third party applications. Google returned 3.4M results with a search on “IBM Watson failure.” And it wasn’t just in healthcare. The Motley Fool article (2018) summarizes Watson’s difficulty picking exchange-traded funds. Surprised?

And if you want the American Medical Association’s take from 2019, this AMA Journal of Ethics article entitled, “Should Watson Be Consulted for a Second Opinion” states: 

“In the future, it may very well be considered unethical (and create liability) not to consult Watson or intelligent systems like it for a second opinion…” 

But wait – there’s more: 

“… assuming that such systems prove effective in what they purport to do.” 

So, it seems the answer – at least for now – is that there is no ethical need to ask Dr. Watson. 

Any DMC Forum members have experience with Watson? Will this news deflate interest in AI applications for healthcare? 

Disclosure: I was a member of the IBM Watson Healthcare Advisory Committee (~2012 – 2015).