Abstract
This paper presents evidence that doctors behave very differently in making treatment recommendations depending on the region where they work, creating large variations in the quantities of care delivered to seemingly standardized populations. This evidence on “variations” (and the failure of normal explanations of the variations) leaves almost by default the idea that incomplete diffusion of information about the efficacy of medical information must be largely responsible. The paper then discusses reasons why this problem might occur: difficulties in collecting information about the success of medical procedures; difficulties in establishing property rights to such information, even if it were to be collected; and liability considerations that hinder adoption of any information that is collected. It concludes with some suggestions for addressing these problems.

This publication has 8 references indexed in Scilit: