Confusing Watson, Confusing Me


Leonard Zwelling

            I was just
sent the video of the conference held last Friday, October 18, that
demonstrated the Oncology Expert Advisor, the Madison Avenue name for the MD
Anderson/IBM Watson collaboration. It’s definitely worth a look.

going into the details of the presentation, I came away very confused. Was this
“OEA” (Dr. Kantarjian’s term for it):

An electronic medical record (EMR) on steroids?

Comparative effectiveness research (CER) on
steroids? OR

An attempt to market MD Anderson care to a
world-wide audience (what was repeatedly called the “democratization of our
knowledge”)? OR

A new research tool? OR

All of the above?

I think it is 5), all of the above.
So let’s examine what was shown and compare it with these descriptions of the
collaboration’s putative purposes.

I couldn’t believe that MD Anderson
is struggling to install an EMR and a clinical research data base once I saw
this demo. It looked great and exactly what I would have wanted to manage my
oncology patients when I cared for them many years ago. A click generated
beautiful flow sheets in many colors. You could click on icons and they would open
to reveal all kinds of additional information, Wow! This was a cool tool to
manage a patient. Why not let everyone in the place use it?

What a tool for CER as well! It
gathers data in any form, repackages it so that a human can understand and
manipulate it and spot trends. Isn’t that what CER is supposed to do? Can you
imagine if instead of leukemia as the first disease selected, primary prostate
cancer had been analyzed and we had a better handle on what really works for
these patients? As one of the only places that has treated primary prostate
cancer patients with all known modalities from proton therapy to watchful
waiting, MD Anderson is sitting on a treasure trove of primary prostate cancer

There was also a high quality, heavily produced Hollywoodized promotional film in the middle of the hour that suggested that
this new tool could be the vehicle by which MD Anderson could export its
knowledge to the world. This struck me as strange given that every patient that
was discussed in the demo had come not from Asia, but from the US. Why? Because
the local leukemia care in America where the patient was diagnosed was
inadequate. This suggested that one need not travel to another a third world time
zone to find poor quality cancer care. I was not convinced that knowing that
tumor lysis syndrome was right around the corner would equip a local physician in
Africa or even Kansas to treat it in as skillful a manner as it would be
managed here at 1515. I think an MD Anderson physician might be the best key
missing piece for someone with a serious disease like leukemia and even a 3D
printer can’t send one of those to Malaysia on the internet. I am not sure that
MD Anderson quality care can be franchised like Starbucks with the assistance
of Watson even if Dr. Kantarjian was on the other end of the world wide web (if
he was not asleep when his insight was needed in Thailand because of the
difference in time zones).

I saw nothing presented that
suggested there were any new research findings that emerged from the first year
of the collaboration, but perhaps it is too early. It is however important to
realize that a whole lot of laboratory based basic molecular and biochemical investigations
into the very nature of leukemia will not be done so that this project can be
paid for with the best possible outcome for the IBM/MD Anderson collaboration
being the “democratization” of currently available treatments only. That was never
what MD Anderson was about before. Sure, Anderson investigators were more than
happy to share their findings with the world through consultations, meeting
reports and journal articles, but by the time their data were publicly available,
our faculty were on to other, newer things.

It is noteworthy that the major
donor for this project had a grandfather with—surprise—leukemia and that Dr.
Kantarjian helped manage that patient—on the phone. Do you really think a
computer can substitute for Hagop? I don’t either.

So I would assess the news
conference as providing little news but a lot of technology, little in new
ideas, but a lot of dissemination of old ones, and little in novelty but a lot
of cost.

One thought I did have was that
perhaps the team should offer Watson to the Department of Health and Human
Services to run the federal health insurance exchanges. These IBM guys seemed
to have it all over the folks in DC when it came to IS. 

On October 22, I finally was able
to register on after 4 previous attempts had failed due to
system overload. Unfortunately, the system still could not verify my data so I
could not have purchased insurance.

The Watson presentation started
with Dr. Chin asking us to IMAGINE a world without cancer.

First, I am not at all sure that is
very likely until we understand the origins of the collection of maladies known
as cancer, which we currently do not. It is not at all a consensus that the
aberrancies unearthed by genomics (with or without Watson) are the cause or the
effect of the carcinogenic process.

Second, I saw nothing in this
presentation, more a triumph of marketing than one of science, to suggest this
project will generate any new insights using old data. I always thought CER
generated the best next question and rarely the absolute answer.

Third, I find it unseemly that we
address a critical problem in clinical research with the type of hype used to
sell corn flakes.

I guess I really expected a rock
star to finish the hour, not a wannabee. Actually, I really expected William
Shatner. I just wasn’t sure if he would be Captain Kirk or the Priceline
Negotiator when he appeared.

Beam me up, Scotty!

Leonard Zwelling