An Outsourcing Urban Myth

April 19th, 2006

Posted by: Roger Pielke, Jr.

In today’s New York Times David Leonhardt has a column which debunks the supposed exodus of radiology work to India, finding such claims to be vastly overstated:

A few years ago, stories about a scary new kind of outsourcing began making the rounds. Apparently, hospitals were starting to send their radiology work to India, where doctors who make far less than American radiologists do were reading X-rays, M.R.I.’s and CT scans.

It quickly became a signature example of how globalization was moving up the food chain, threatening not just factory and call center workers but the so-called knowledge workers who were supposed to be immune. If radiologists and their $350,000 average salaries weren’t safe from the jobs exodus, who was?

On ABC, George Will said the outsourcing of radiology could make health care affordable again, to which Senator Charles E. Schumer of New York retorted that thousands of American radiologists would lose their jobs. On NPR, an economist said the pay of radiologists was already suffering. At the White House, an adviser to President Bush suggested that fewer medical students would enter the field in the future.

“We’re losing radiologists,” Representative Sherrod Brown, an Ohio Democrat, said on CNN while Lou Dobbs listened approvingly. “We’re losing all kinds of white-collar jobs, all kinds of jobs in addition to manufacturing jobs, which we’re losing by the droves in my state.”

But up in Boston, Frank Levy, an economist at the Massachusetts Institute of Technology, realized that he still had not heard or read much about actual Indian radiologists. Like the once elusive Snuffleupagus of Sesame Street, they were much discussed but rarely seen. So Mr. Levy began looking. He teamed up with two other M.I.T. researchers, Ari Goelman and Kyoung-Hee Yu, and they dug into the global radiology business.

In the end, they were able to find exactly one company in India that was reading images from American patients. It employs three radiologists. There may be other such radiologists scattered around India, but Mr. Levy says, “I think 20 is an overestimate.”
Some exodus.

Nonetheless, Leonhardt suggests that issues related to outsourcing remain:

For now, the practical effect on radiology is small. At its highest levels, the United States health care system may be the best the world has ever known. India doesn’t even have many radiologists today, let alone a large number who measure up to American standards.

But that’s going to change. Eventually, Indian doctors will be able to do the preliminary diagnoses that are a big part of radiology. Something similar will happen in accounting, architecture, education, engineering and the law, as Mr. Levy and his colleagues suggest in the coming Milken Institute Review.

These fields tend to be regulated already, giving them noble excuses — like certification, client privacy and legal accountability — to put up trade barriers. But the real reason will usually be a simple desire to protect jobs and salaries.

When factory workers have asked for that kind of protection, the country has told them no. So why does the answer change when the request comes from a wealthier, more influential group of workers?

3 Responses to “An Outsourcing Urban Myth”

    1
  1. Mark Bahner Says:

    Hi,

    “Eventually, Indian doctors will be able to do the preliminary diagnoses that are a big part of radiology.”

    You ain’t seen nothin’ yet! (To quote BTO. ;-) )

    I think it was an issue or two ago in MIT’s Technology Review that talked about putting 50,000 cases of breast cancer x-rays into a massive database. Eventually, they hope to get *millions* of cases into the database. It’s called a “medical grid” (as in electrical grid).

    They will also tie those x-rays to patient therapies used (e.g., surgeries, drugs) and patient outcomes.

    With the database completed, computer programs can be written to relate x-rays to therapies used, and patient outcomes.

    In other words, radiologists themselves can be replaced by computer programs that have “looked” at literally tens of thousands–if not millions–of x-rays, and can relate the results not only to appropriate therapies, but likely outcomes.

    And THEN…they will start putting in the *DNA* of each patient, to relate that to X-rays, treatments, and outcomes.

    The power of the whole thing is simply mind-boggling. Rather than having to rely exclusively on the patient’s oncologist and radiologist, the patient–probably through the oncologist–will have access to a set of computers that eventually have perfect memory of literally millions of cases of breast cancer. (And in 20 years, it could be hundreds of millions of cases of breast cancer.)

    The only way this can happen is because of the mind boggling decrease in the cost of computer power. Say each x-ray, scanned in very fine detail, contains 20 megabytes of information. (I’m just guessing; anyone knowledgeable feel free to correct me.) Let’s say 10 x-rays per person (I have no idea). That’s 200 meg per patient. Now, 50,000 patient: that’s 10,000 gigabytes of data. Well, even 20 years ago, that would be a hundred-million-dollar proposition. But today, at 50 cents a gig, it’s only $5,000!

    Amazing. And this power will be available *anywhere in the world*…from the U.S. to India, to Botswana and Mongolia. The X-rays simply get uploaded into the grid (the database), and out comes information on the ~100 or ~1000 most similar cases, with treatments and outcomes.

    That’s part of why I’m very confident that world economic growth will continue to accelerate as the 21st century progresses:

    http://markbahner.typepad.com/random_thoughts/2005/11/why_economic_gr.html

    P.S. I guessed about 200 meg per patient. That appears to be an excellent swag. Through the miracle of Google:

    http://www.upenn.edu/almanac/v46/n34/digital-mam.html

    …the answer apparently is approximately 160 meg per patient.

    P.P.S. I think that Internet article may be on the exact same project as described in the MIT Technology Review article.

  2. 2
  3. Rabett Says:

    They will never outsource call centers. It’s impossible.

  4. 3
  5. Markk Says:

    The issue of outsourcing in this area crosses with the issue of HIPPA. There are tough legal issues related to who has control of patient information that I think will lead to fat lawsuits in the future.

    One of the main problems, I think, with outsourcing from the US is that what it is really doing is changing where the research and development money will be spent. All the corporate research institutes going up in India and China, WITH the factories there for real developmental interaction, means that the focus of new developments will be there, not in the US and the U.S. is becoming a hollow society in terms of technical expertise combined with real world issues. Of course this is not 100% or even 50%, but even a 10% swing means that the U.S. is much less resiliant. I think this is inevitable.