Friday, March 20, 2015

A $6.6 Million CEO Dreams of a "Doctor-Less" Future

The CEO of giant hospital system UPMC, Mr Jeffrey Romoff, has been one of the best compensated CEOs of ostensibly non-profit hospital systems.  As we noted here, his 2013-14 compensation was $6.6 million.  UPMC has become so big and its top managers so rich that a former Mayor of the city of Pittsburgh sued the organization claiming it was not really not-for-profit (look here and here).  The leadership of UPMC has previously supplied us with some interesting examples of conflicts of interest (look here and here). 

The announcement of a new alliance of Pittsburgh organizations provided an interesting insight into the thinking for which such a CEO is paid the big bucks.  Leaders of three big organizations, UPMC, the University of Pittsburgh (with which UPMC is affiliated), and Carnegie-Mellon University announced an alliance to use "big data" in health care (see this article in the Pittsburgh Post-Gazette).

UPMC, the University of Pittsburgh and Carnegie Mellon University on Monday announced the formation of the Pittsburgh Health Data Alliance to 'revolutionize health care and wellness' by using data to detect potential outbreaks as well as create health care innovations that will spawn spinoff companies.

The clinical goal, the leaders of the three institutions said, is to remake health care so that it is at once more computerized, yet more personalized, using millions of gigabytes of accumulated health records to predict and treat patients’ health issues in a manner far more specific than is possible today.

Big data now seems to be the latest rage in business schools and among the high-tech crowd, never mind the failures of fancy statistical modeling based on big data that helped lead to the global financial collapse of 2008.  Similarly, despite at least 30 years of research, multivariate prediction and diagnostic modeling in medicine has never lived up to its expectations.  Few models have been demonstrated to be better than mediocre predictors when tested in real-life clinical settings.  Finally, there are numerous concerns about privacy and data security when patients' data is being avidly traded back and forth.

The most striking talk in this meeting, however, was by UPMC CEO Jeffrey Romoff.  The Pittsburgh Post-Gazette noted,

Mr. Romoff said he envisioned 'doctor-less health care,' which is not to say there will be no doctors in the future, but they will be greatly aided by computerized diagnoses, by biometric data gathered on smartphones and transmitted in real time, and by a patient’s own genome. It could result in a new form of  'artificial intelligence,' he said.

The reporter, however, seemed to have edited Mr Romoff to take the edge off what he said.  A video of that part of the conference can be found, for the moment, here.  I transcribed, I believe accurately, Mr Romoff's three most relevant sentences.

The majority of healthcare that everybody receives will be accessible on their handheld device.

We will be thinking about 'doctor-less' healthcare.

We will in fact create an artificial intelligence better than the superb level of intelligence we now have among our physicians and our healthcare professionals.

So, in my humble opinion, it did not sound like Mr Romoff was just envisioning that physicians someday may actually have access to diagnostic or predictive models that are highly accurate for real patients.  He was envisioning replacing physicians with machines, with artificial intelligence. 

Again, never mind that despite years of work and billions of dollars, artificial intelligence so far has proved remarkably dumb.

So furthermore, in my humble opinion, this provided a glimpse into how health care managers now think.  Mr Romoff appears to be a generic manager.  He is not a health care professional, and has no apparent experience taking care of patients (see his official bio, listing his most advanced degree as a Masters in Philosophy).  Generic managers now often seem to think of themselves as some sort of new aristocracy, far removed from the peasants who work for them.  Would not it be easier for such aristocracy to avoid working with such peasants at all?  Machines would be so much neater and cleaner, would not ask for raises or think of unionizing or rebelling (at least outside of the world of Terminator movies).

Leaving aside such fantasies for the moment, the most concerning problem with Mr Romoff's dream of robotic doctors is that anyone who has ever had any direct involvement in health care knows that doctors need to do much more than crunch data and make predictions and diagnoses.  Doctors and other health care professionals have sworn to put patients' interests first.  That implies that doctors must talk to, endeavor to understand, and be empathetic towards their patients.  Many times we doctors may not do this anywhere near perfectly.  But we are human, so can at least try.  Artificial intelligence may be getting closer to making better health care predictions and diagnoses, but does anyone seriously think we are close to making an understanding, empathetic machine?

I believe that Mr Romoff has unwittingly made another argument why he and his fellow generic managers should not be leading health care.  Health care should be lead by people who understand the actual care of patients, uphold health care professionals' values, and are willing to be accountable for putting patients' and the public's health first. 


Anonymous said...

This man is scary,

Darren Gorensk said...

I'd also note that the collaboration will also create a Center for Commercial Applications of Healthcare Data, which sounds like it will push forward the technocratic view of healthcare through commercialization of research...

Anonymous said...

Did you see him on 60 Minutes? When L. Stahl asked him how much money he made, he mumbled and stumbled, and finally spit it out.

Joseph j7uy5 said...

It is safe to say that this is a dumb idea: "The majority of healthcare that everybody receives will be accessible on their handheld device. We will be thinking about 'doctor-less' healthcare. We will in fact create an artificial intelligence better than the superb level of intelligence we now have among our physicians and our healthcare professionals."
I should add, I have seen this coming. Over the past twenty years or so, there has been a trend toward deprofessionalization of health care. That is, the tech and orderlies, rather than being seen as valuable parts of the team, have been marginalized to the status of interchangeable parts. Then this happened to nurses and social workers. Now it is happening to doctors. In hospital chains, even the CEO/COO/CFOs are basically floor managers. All of this is in the service of "shareholder value."

Note that the originator of the concept of shareholder value, Jack Welch, has been quoted as saying it is “the dumbest idea in the world”.

The Dumbest Idea In The World: Maximizing Shareholder Value

James Gaulte said...

Outstanding commentary.frightening that so much power is in the hand of folks so ignorant of medical care and the limits of statistics and risk factor epidemiology.

Judy B said...

Considering that far too many doctors practice "guideline medicine," we may have already strayed into this territory. When will doctors make an effort to take back their profession from the "managers?"

InformaticsMD said...
This comment has been removed by the author.
InformaticsMD said...

When I see claims that something will "revolutionize" medicine, which describe as exhibits of bellicose grandiosity (or grandiose bellocosity), I am reminded of my 2009 paper of a few yeara ago:

"The Syndrome of Inappropriate Overconfidence in Computing: An Invasion of Medicine by the Information Technology Industry"

Per Dr. Poses. on Naked Capitalism Yves Smith called this type of Big Data mania an example of "lunatic triumphalism."

The problem is, this lunacy really is a very destructive fanaticism.

-- SS