Lisa A Prosser1, Eve Wittenberg. 1. Department of Ambulatory Care and Prevention, Harvard Medical School and Harvard Pilgrim Health Care, Boston, Massachusetts 02215, USA. lprosser@hms.harvard.edu
Abstract
OBJECTIVE: To evaluate differences in risk attitude across the domains of health and money for 2 types of respondents, patients and community members. METHODS: Two groups of respondents, patients with multiple sclerosis (n = 56) and members of the general community (n = 57), completed a survey that collected information on risk attitudes and socioeconomic and clinical variables (e.g., disability level). Risk attitude was measured using 2 standard-gamble questions on money and 1 standard-gamble question on health outcomes. Multivariate regression was used to evaluate the relationship between risk attitude and respondent type (patient v. community), adjusting for covariates that could affect risk attitude. RESULTS: The median certainty equivalents for money gambles were significantly different from and less than the expected value of the gamble for both types of respondents. Median certainty equivalents for the health gamble were not significantly different from the expected value for either group of respondents. For all 3 gambles, there was no difference in median certainty equivalents between the 2 types of respondents in both unadjusted and adjusted analyses. CONCLUSIONS: Risk attitude varied across domains but not by respondent type. Patients and community members were predominantly risk neutral with respect to health outcomes and risk averse with respect to money. Research on risk preferences on money outcomes may not be an appropriate proxy for risk preferences regarding health outcomes. Risk preferences may depend more on characteristics of the choice than on respondent type.
OBJECTIVE: To evaluate differences in risk attitude across the domains of health and money for 2 types of respondents, patients and community members. METHODS: Two groups of respondents, patients with multiple sclerosis (n = 56) and members of the general community (n = 57), completed a survey that collected information on risk attitudes and socioeconomic and clinical variables (e.g., disability level). Risk attitude was measured using 2 standard-gamble questions on money and 1 standard-gamble question on health outcomes. Multivariate regression was used to evaluate the relationship between risk attitude and respondent type (patient v. community), adjusting for covariates that could affect risk attitude. RESULTS: The median certainty equivalents for money gambles were significantly different from and less than the expected value of the gamble for both types of respondents. Median certainty equivalents for the health gamble were not significantly different from the expected value for either group of respondents. For all 3 gambles, there was no difference in median certainty equivalents between the 2 types of respondents in both unadjusted and adjusted analyses. CONCLUSIONS: Risk attitude varied across domains but not by respondent type. Patients and community members were predominantly risk neutral with respect to health outcomes and risk averse with respect to money. Research on risk preferences on money outcomes may not be an appropriate proxy for risk preferences regarding health outcomes. Risk preferences may depend more on characteristics of the choice than on respondent type.