Gawande’s Kitchen

Atul Gawande is a fine writer and probably a good doctor. He was a key health care advisor to the Obama administration, largely based on writing very long articles mostly for the New Yorker about the problems he sees in the current health care system.

His most notable contribution to the health care debate was a major article he wrote in June 2009, which claimed to show that physicians in McAllen Texas were systematically over treating patients to enrich themselves. This article “became required reading in the White House and Congress during the health care debate and turned McAllen into shorthand for America’s medical spending problem,” according to NPR (National Public Radio).

Gawande argued that costs in McAllen are far higher than in El Paso, a town with similar characteristics. His only explanation: for some reason physicians in McAllen are greedier than physicians elsewhere.

As it turns out, he was wrong.

This notion suited the White House reformers to a tee. Most of them figured all the problems in health care are the fault of greedy doctors and the solution is to control the docs. Hence, the inclusion of Accountable Care Organizations, pay-for-performance, electronic medical records, and comparative effectiveness research — all designed to get doctors under the control of bureaucrats (and we all know what models of efficiency those bureaucrats are).

Unfortunately, Dr. Gawande failed to consider a number of other factors that would complicate his simple explanation. One correspondent who is very familiar with the medical market in Texas wrote to me saying, “It seems pretty clear that Gawande drew the wrong conclusions. The most notable difference between McAllen and ANYWHERE else, according to the published Dartmouth data, is that McAllen has far fewer physicians. The result of this is that there is very little ambulatory care, far more inpatient days and a larger proportion of the care is delivered in the emergency room.  It’s the shortage of physicians in McAllen that is causing costs to be so high.  I don’t know how Gawande missed it.” She added that there is also a physician shortage in El Paso, but people there tend to cross the border into Juarez, Mexico, for physician services, while there is no similar opportunity across the border from McAllen.

And the NPR story linked above was about a later study published in Health Affairs that looked at Blue Cross Blue Shield data, rather than just Medicare data, and found that Blue Cross actually pays less per person in McAllen than it does in El Paso. So the rush to find simple answers — greedy doctors — resulted in flawed policy prescriptions. I wonder if Dr. Gawande would practice medicine the same way he practices public policy.

All of this is just by way of introducing his latest offering. His most recent article in the New Yorker is called “Big Med: Restaurant chains have managed to combine quality control, cost control, and innovation. Can health care?” (See John Goodman’s previous post on this.)

Gawande looks closely at The Cheesecake Factory and is astonished at how well-organized and efficient the kitchens are in these restaurants. Costs are low, quality is superb, and customer satisfaction is high. Everything is standardized and every worker’s job is well defined and limited to just a few functions. As I said, he is a very fine writer and he made me hungry describing how food is prepared in these places.

He wonders why medical care can’t be delivered the same way, and finds a couple of examples of where similar practices are being introduced in hospital settings. He believes this is the future of medicine — centralized control by national chains that prescribe in detail what must be done by each worker and oversee how long they spend on each task.

One of his examples is the joint replacement department at his own Brigham and Women’s Hospital in Boston. Gawande’s mother needed a knee replacement, so he was able to track her progress carefully. They have standardized what they do — the parts they order, the anesthesia they use, the post-op therapy. The hospital got some resistance from the surgeons who were used to doing things their own way, so it allowed for some deviation from the “default option,” but only if they could show there was good reason for doing so. Gawande says “the start-to-finish standardization has led to vastly better outcomes.”

A couple of observations about all this:

  1. This example reminds me of Regina Herzlinger’s concept of “focused factories,” in which a facility learns to do one or two things very efficiently. That may work well for routine and commonplace services like joint replacement, or Herzlinger’s favorite example, hernia repair, but it is hard to see it applying to the extremely wide range of maladies most hospitals deal with.
  2. The methods this department settled on for its standards were based on a great deal of experience with a wide range of alternatives. If there had not been alternatives, they would have had no way of knowing which was optimal. This hospital had been using nine different knee prostheses. They found that all worked equally well. So they were confident in using the three least costly ones. Similarly with post-op therapy: Most of the surgeons had been prescribing passive-motion machines to exercise the knee, but closer examination found that these machines didn’t do much good. So they stopped using them. They were able to do this because no one treatment had been standardized and required for all patients. If there hadn’t been a variety of treatments used, no one would have known there was a better method available.

But there is a much more important question being ignored by Gawande — How well does The Cheesecake Factory analogy really apply to health care? We can see how similar the kitchen is to an operating room — lots of busy people rushing about in a sterile environment, each concentrated on a task. But what about the rest of the “system?”

At The Cheesecake Factory, the customer is the diner. That’s who orders the service, pays the bill, and comes back again if he is happy. That is who all of the efficient, standardized food preparation is designed to please.

In Gawande’s ideal health care model, however, the customer isn’t the patient, but the third-party payer, be it an insurer or government. Let’s call that entity the TPP. The TPP never enters the kitchen. The TTP has no idea what happens in there, and doesn’t really care as long as the steak is cooked to his satisfaction and the tab is affordable.

In this model, the patient is actually the steak. It is the steak who is processed in the kitchen. It is the steak that is cut and cooked and placed on a platter. The steak doesn’t get a vote. Nobody cares if the steak is happy. The steak doesn’t pay the bill. The steak isn’t coming back again.

So here we are in Dr. Gawande’s kitchen, where you and I are slabs of meat and Chef Gawande will cook us to the specifications of his TPP customers — satisfaction guaranteed.

Comments (14)

Trackback URL | Comments RSS Feed

  1. Otis says:

    This is eyeopening.

  2. Jane Orient says:

    Exactly, Greg.

    Also remember that at the Cheesecake Factory they throw out a lot of meals. And they probably also throw out any ingredients that aren’t perfect.

    I think that “managing the case mix” is the term for it.

  3. Uwe Reinhardt says:


    There are a number of physicians among the disciples following this blog. I wonder if they truly treat patients like slabs of meat in their practice and just do not care what happens to patients.

    I hope they will enlighten us on this point. It is not my personal experience.


  4. Dayana Osuna says:

    Excellent post Mr. Scandlen! You and Dr. Goodman nailed it with this and the previous post on this issue. Really enjoyed reading this!

  5. Corey says:

    Doctors will draw the time somewhere, but the incentives of the third party payer (especially if over standardized) will push them there.

  6. Devon Herrick says:

    I loved Gawande’s article. It was interesting both from the standpoint of how complex meals are prepared to order; and how disintegrated health care is. One of the biggest differences between health care and meal preparation is that meals are made to order, whereas health care is fixing something that is wrong.

    Hindsight is 20/20. It’s easier to learn the mistakes of past experiences than start from scratch. But, each patient is often just that: starting from scratch. Nonetheless, I still believe Gawande has a good point. A better point is made by John Goodman. That is, until we get the incentives right, we cannot expect doctors and hospitals (and industrial engineers and venture capitalists) to solve the problems in health care like they seem adept at solving the problems in a host of other industries. You don’t have to actually like the Cheesecake Factory to benefit from the innovation it provides. Restaurants are a competitive industry — you can find anything from poor immigrants making Chinese food to gourmet chefs operating 4-star restaurants. They may not all be as efficient or have as varied a menu as the Cheesecake Factory, but if the quality falls too far, patrons will leave and a new competitor will attract their business. In health care, Medicare pays the same to inefficient hospitals as it pays to efficient ones. It’s hard to say exactly what a Focused Factory in health care would look like. But we know it would look very different than what we see today.

  7. Ashley says:

    Freer markets are nearly always better!

  8. Linda Gorman says:

    The claim that the health care system is “disintegrated” always interests me as it always seems to be deployed by those who favor less choice for patients and more centralized control in general. This in spite of the fact that there is evidence suggesting that in the real world integrated systems increase expenditures and, maybe, costs.

    I’m told that once upon a time physicians had referral groups. They referred their patients to physicians that they trusted and admitted patients to hospitals that they worked well with. I’ve even experienced it once or twice and was happy with it as it seemed to operate along the lines of the good recognize other good people model.

    That tacit policy changed with the arrival of the HMO and with laws responding to repeated claims (that haven’t, as far as I know been well supported) that physicians were only after the money and didn’t care about patients. These live on in the unprecedented denigration of fee for service care and the recent claims that the pricing system encourages physicians to concentrate on volume so we need to move everyone into the kind of capitated care that has been shown to harm the seriously ill.

    The evidence showing that FFS leads physicians to carve up patients for the money isn’t as far as I can tell, very convincing. I figure I’m missing something as even supposedly market oriented health policy people participate in FFS bashing.

    Therefore, it is also interesting to see the comments on this thread saying heck no, physicians are professional and care about their patients.

  9. Dale Rasmussen says:

    The key concept in your post is your first statement: “Atul Gawande is a fine writer and probably a good doctor.” I am a fan of his ability to conceptualize and communicate his observations about our health care culture on a level that is understandable and entertaining. Dr. Gawande’s research and presentations are thought provoking and inspiring, and they are important in the discussion to evolve and improve processes in our health care system. What I’m not a fan of is the way his stories and opinions are used by politicians as the singular scientific and authoritative voice of credibility in support of bad health care policy.

  10. Chuck says:

    Interesting analysis.

  11. Dave says:

    I’m an orthopedic surgeon and agree to some extent with the restaurant analogy for my specialty. I think most orthopedic surgeons would tell you that they enjoy the mechanics and engineering involved in their specialty. Treatments can be engineered by applying epidemiolocic principles to research. The ACA plans to attempt this with “comparative effectiveness research”, to be funded by the PCORI.
    This is all well and good for the treatment end of medicine. However, the diagnosis end of medicine is much more gray and complex than treatment. It is not possible to compare one thought process for diagnosis to another. This becomes more difficult as patients become older and/or sicker. The assembly line approach fails when we realize that we only want the best approach to our medical problems supervised by a physician we trust, and not a pre-approved “guideline” certified by an annonymous academic board.

  12. Michelle says:

    Dr. Gawande is an excellent writer and an excellent doctor committed to preserving lives, not special interests.

  13. Pave Low John says:

    As analogies go, the health care/restaurant is quite problematic. The immediate flaws that jump out are:

    1) Restaurants face a vast amount of competition. In fact, the restaurant game is so cut-throat that about 30% of restaurants fail in the first year of operation (there is a myth that the failure rate is much higher, but this study at UCSB addresses that misconception. see Hospitals and medical professionals are in a much different environment with respects to competition and business model.

    2) Most people are going to consume roughly the same amount of food, say 30 dollars worth in Cheesecake factory for the sake of arguement. Heathcare, on the other hand, sees a wildly skewed expenditure pattern for each patient, based on a whole host of factors.

    3) No one, to my knowledge, uses insurance or direct government assistance to buy food at a restaurant. Even EBT cards (food stamps) can’t be used at a restaurant. It’s pretty much cash, check or card.

    4) The paperwork associated with eating a meal at Cheesecake and heathcare aren’t even in the same universe, much less the same ballpark. Bureaucratic red tape is a huge drag on healthcare, a topic I would actually like to see someone like Dr. Gawande tackle.

    5) The Cheesecake Factory, as far as I know, does not routinely deal with life-and-death issues (although their desserts can give you an out-of-body experience, especially anything with chocolate…) That part of healthcare is probably the biggest difference – The stakes are much higher. If people were dying or being permanently crippled because a restaurant served a lousy meal, it would completely change the restaurant dynamic.

    I could probably think of more, but those are the ones that came to mind the fastest. I applaud Dr. Gawande for attempting to shed light on an important issue, but trying to find the right analogy can be a tricky thing.

  14. Bob Geist says:

    Greg, a good piece.
    Gawande missed the fact that patients seek health care when in pain and/or worried about being sick–it’s always both. They are medicine’s customers and may have disease, but they seek care when they have symptoms.
    Gawande and “The Managed Care Industry” see care the other way around—diseases in patients, not symptoms are the problem. In this view the factory-like efficiency of food franchising businesses can control disease. Control can be extrapolated into the fantasy of “efficiency” curbing costs for employer and government agency “buyers” of care.
    Too bad that factory-like efficiency, desirable as it may seem, is irrelevant to demand inflation from political tax subsidies for insurance that have created a “free” care medical market. There is no efficiency cure for a politically perverted market gone awry, a market where “customers” are the few “buyers” of insurance for a clientele of “consumer cost centers”.
    Thanks, Bob