Code Red: Two Economists Examine the U.S. Healthcare System

June 2, 2011

The Summer of Sequels

Filed under: Uncategorized — David Dranove and Craig Garthwaite (from Oct 11, 2013) @ 9:28 am

I have seen this film before. Folks get all excited about the potential for vertical integration to save our healthcare system, and then the facts emerge.

The results of the first major ACO demonstration project are in and unless there is some hidden meaning behind all the data, it looks like ACOs may not be the magic bullet that the Obama administration had hoped. The demonstration began under President Bush and the specific payment structure and quality incentive differ somewhat from the ACO rules under the Affordable Care Act, but the main features are the same – give an integrated provider organization a share of the savings if it can hold down Medicare spending while also offering some quality bonuses. Despite the fact that the participants included ten of the nation’s best known physician-led integrated organizations, less than half were able to lower Medicare costs by the final year of the project and only two demonstrated consistent cost savings. And the methods used to achieve savings – nurse call centers and telephone health checkups – are the sorts of thing that don’t exactly require vertical integration.

There are going to be excuses – the ACOs need to be run by hospitals, they need more time to develop their information technologies, the performance incentives need to be strengthened. But that is the kind of ex post rationalizing one hears any time an experiment fails to support a theory. Maybe the theory (that vertical integration is the panacea for our ailing system) is wrong.

There remains a deep divide in both academia and amongst practitioners about the merits of vertical integration. Supporters of the ACA will continue to pin their hopes for Medicare savings on ACOs, the new evidence be damned. I have blogged that we need less top down direction about how to organize care delivery and I am sorry in a way to see the data bear me out. (Sorry because I am fresh out of magic bullets.) If the government is to play a role in the future of the health system, then it should either go all in on regulation (i.e., single payer) or fix the problems that are limiting the effectiveness of the free market (e.g., subsidize and standardize integrated health information systems; double down on antitrust enforcement.) Will these ideas work? I don’t know. But at least there isn’t a strong theoretical case to be made that they will fail.

Let’s stop this love affair with vertical integration. How many times do we have to keep seeing this bad movie?

P.S. I am currently reading The Quantum Story which is about the evolution of quantum physics. I barely understand much of it. But this much I do understand. When physicists perform experiments and the results do not confirm their theory, they reject the theory.


  1. Isn’t this a case of glass half full vs. empty? 7 out of 10 institutions earned the bonus at least once and 2 earned it all five years.

    I am very curious about other proxies for success rate, e.g.,
    – Of the 50 Institution-Years (5 institutions x 5 years), how many earned the bonus
    – What was the average % reduction in costs. Did Marshfield Clinic and other overachievers overshoot the 2% target?

    Comment by Manu — June 2, 2011 @ 9:56 am

  2. PS Typo correction in prior comment

    50 Institution years are based on 5 years * 10 Institutions

    Comment by Manu — June 2, 2011 @ 9:56 am

    • Considering that they only needed to show 2% savings, the fact that less than half did so is entirely unimpressive. And remember, these organizations volunteered for this, so they must have expected to succeed. What does this say for others who are impressed into duty?

      Comment by dranove — June 2, 2011 @ 10:26 am

RSS feed for comments on this post.

Blog at

%d bloggers like this: