Funding

The purpose of this study is to test whether physicians change their use of non-recommended tests, procedures, or medications more in response to evidence based-guidelines, price information, or an individual patient's story.

We will perform a randomized controlled trial (RCT) of information presented to physicians to test the hypothesis that an identifiable victim affects physician practice behavior more than a statistical victim.

Specifically, we will answer the following research questions: 1) do physicians order fewer non-recommended tests, procedures, or medications if they are told about a patient or a physician who had a bad outcome from that test, procedure, or medication than if they are simply told the guideline or the cost of the test, procedure or medication, 2) does the effect of learning about the identifiable victim last longer than the effect of learning about the guideline or the cost of a test, procedure, or medication, and 3) does the identifiable victim effect differ if the victim is a patient or a physician? We hypothesize that because of the propensity to respond more to the identifiable victim rather than the statistical victim that physicians will order fewer unnecessary tests when they are told about an individual patient case than if they are simply told about the guideline, that the effect of the identifiable victim will last longer than the effect of the statistical victim, and that a patient as the identifiable victim will have more effect than a physician as the identifiable victim.

The identifiable victim effect refers to the tendency to offer more aid to a specific, identifiable victim rather than a vaguely defined group of people with the same need. In the this study, the identifiable victim is a fictional patient who experience a negative consequence as a result of an unnecessary test. The identifiable victim effect is described and studied in the following articles:

Small D. Sympathy and callousness: The impact of deliberative thought on donations to identifiable and statistical victims. Organizational Behavior and Human Decision Processes 2007;102:143-53.

George Loewenstein, Deborah A. Small, and Jeff Strand. "Statistical, identifiable, and iconic victims" in Edward J. McCaffery, Joel Slemrod (2006). Behavioral public finance. Russell Sage Foundation; pp. 32-35.

Inclusion Criteria:

• Primary care physicians in the Weill Cornell Physicians Organization

Exclusion Criteria:

• None