Rational economic creatures?

I heard a wonderful talk by Abhijit Banerjee, economics professor at MIT, about his and Esther Duflo's new book entitled, Poor Economics.  Here is a short summary, accompanying this video interview of the authors.

Why do the poor remain poor despite a million different strategies to counter poverty? Well, perhaps because policies that deal with poverty alleviation are often based on cultural and literary stereotypes of how the poor are "lazy or enterprising, noble or thievish, angry or passive, helpless or self-sufficient." And therefore we often rely on over simplistic policies with readymade formulae - "Free markets for the poor," "Make human rights substantial," "Give more money to the poorest." A new book, Poor Economics, tries to make one key point - let’s stop staring at data and theories, and understand instead the coherent story of how really poor people live their lives.

The authors present several examples of policy and programmatic interventions that have failed because policy-makers do not take the time to understand how things work on the ground in these poor communities.

I am struck by the similarity to many proposed interventions in health care.  In the last several days, I have discussed this with regard to penalties for failure to meet certain metrics regarding patient readmissions to hospitals.  But it is a broader issue.  For example, a move to capitated rates of pay is viewed by some as the sine qua non of health care policy.  I have noted that there is little empirical support for this approach, even if it might have a sound economic rationale.

But does it have a sound economic rationale?

Still feeling the after-effects of a morning at MIT, where I first learned the term over four decades ago, I propose we conduct a gedanken experiment.  That is, let's consider a hypothesis for the purpose of thinking through its consequences.

I put forth the following thought experiment.  Advocates of capitated, or global, payments argue that the current system of fee-for-service medicine leads to overuse, in that doctors and hospitals have a financial incentive to conducts tests and procedures to generate revenue.  The economic underpinning of a global payment system is that hospitals and doctors are rational economic creatures.  Setting a per-patient budget, it is argued, will cause the hospitals and doctors to work within that revenue envelope to deliver care more efficiently.  They are at risk for any over-spending and they get to keep the surplus if they beat the budget.

But, answer me this.  Let's say, we have a system where, say, 25% of the patients are on a global budget and the remainder are on a fee-for-service payment plan.  

If the economic theory is correct, that the hospital and doctors are rational economic creatures, shouldn't we notice a difference within the same provider network in how the global patients are treated from how the FFS patients are treated?

Let's turn away from the thought experiment briefly to review real data.  I pose a question for my readers:  Has such a difference been documented in those systems that have this mixed payment regime?  I think not.  But if you have counter examples, please provide cites to support your answer.

But now, pretend you are running that hospital and physicians network.  As suggested above, you believe that professional ethics should not allow your system to treat people differently based on the kind of insurance plan that covers them.  So, you instruct everyone to think about all patients as though they are covered by the global fee.  You do this even though you suffer revenue losses from the FFS patients, who, by the way, remain the majority of your patients.

If we do this, haven't we just disproven the hypothesis that doctors and hospitals are rational economic creatures?

So, which is it?  Are they rational economic creatures, willing to treat identical patients differently based on pricing?  Or, are they not rational economic creatures -- treating all patients alike -- in which case the theoretical basis for global payments appears to be problematic?