Rationality versus reality: the challenges of evidence-based decision making for health policy makers
BACKGROUND In previous studies, the authors found that most people, given a fixed budget, would rather offer a less effective screening test to 100% of a Medicaid population than a more effective test to 50% of the population. In a subsequent study, the authors found that the number of people preferring the less effective screening test was dramatically reduced when the percentage of Medicaid enrollees receiving it was less than 100. In this article, 2 new studies are reported that explore whether people's preferences for equity versus efficiency are susceptible to a framing effect. METHODS In 2 studies, the authors presented subjects with multiple scenarios involving screening tests that vary in the proportion of people who could receive the tests within a budget constraint and the number of people whose lives each test would save. Across scenarios, the proportion of Medicaid enrollees who could receive each test was varied, as was the question of whether scenarios involved Medicaid enrollees from the same or a different state. In addition, the authors varied the order in which subjects received the scenarios. RESULTS In the 1st study, people's preferences for equity over efficiency varied significantly depending on the way situations were framed. Preference for equity was stronger when the more widely distributed choice covered the entire population than when it covered only half the population (P < 0.001). In addition, people's preferences were susceptible to order effects, with preference for equity being significantly stronger when the 1st scenario received by subjects involved 1 screening test that could be offered to the entire population (P < 0.001). In the 2nd study, preferences for equity over efficiency diminished even when the different framings were descriptions of identical circumstances--preference for equity was significantly reduced when the population to be screened was framed broadly, in terms of the percentage of patients across 2 states who could receive testing rather than narrowly, in terms of the percentages of patients in 1 state who could receive testing (P = 0.04). CONCLUSION Policy planners should be careful about accepting public preferences for equity over efficiency at face value, because such preferences can be dramatically influenced by framing effects and order effects.