Are preferences for equity over efficiency in health care allocation "all or nothing"?

BACKGROUND. In a previous study we showed that within a budget constraint, most people would rather offer a less effective screening test to 100% of a Medicaid population, thereby saving 1,000 lives, than a more effective tes t to 50% of the population, thereby saving 1,100 lives. We present here a s tudy exploring whether this preference for equity over efficiency persists when neither test can be offered to the entire population. METHODS. Members of Physicians' Online and prospective jurors at the Philad elphia County Courthouse randomly received I of 3 questionnaires (41, 42, o r 43) describing a limited budget to screen Medicaid enrollees for colon ca ncer, In all questionnaires, test I was said to save 1,000 lives, and test 2, a more effective and more expensive test, was said to save 1;100. In 41, test I was offered to 100% and test:! to 50% of the population In 42, the 2 tests could be offered to 50% and 25%, respectively; in 43,to 90% and 40%, respectively. Respondents indicated which test they recommended and provid ed justification. RESULTS. The majority of physicians (59%) and the general public (56%) receiving Q1 favored the less effective screening test Howeve r, of those receiving Q2,only 26% of physicians and 27% of the general publ ic recommended the less effective screening test And of: those receiving Q3 , only 38% of physicians and 28% of the general public recommended the less effective test. Justifications for these recommendations were based largel y on concerns for equality of: treatment among those who chose the less eff ective test and concerns for saving the most lives among those who preferre d the more effective test. CONCLUSIONS, Although most: respondents show a preference for equity over e fficiency when equity means that 100% of a population can receive a service , many fewer;respondents maintain this preference when equity is no longer absolute. This result suggests that the preference for equity is sometimes "all or none."