BACKGROUND: Evidence that long-term hormone therapy (HT) may increase the risk of serious adverse events led to a sharp reduction in all HT use, including short-term use for vasomotor symptom relief. We estimated women's willingness to accept adverse event risks in exchange for vasomotor symptom relief when risks are communicated in absolute vs. relative terms. METHODS: We developed a conjoint survey to elicit preferences across benefit and risk attributes of HT. The survey was administered via the Internet to 523 U.S. women aged 46-60 years. Participants evaluated pairs of hypothetical treatments and indicated preferences using a Likert-based scale. Risks were presented in absolute and relative terms. Satisfaction scores for HT risks and benefits were estimated using random-effects, ordered-probit regression. Maximum acceptable risk (MAR) was calculated as the increase in risk that reduces the satisfaction score for a given level of HT benefit to 0. RESULTS: For both risk versions, the least important attribute (smallest difference in satisfaction scores) was night sweat frequency; heart attack risk was the most important (largest difference in scores). Participants were more willing to accept risks in return for symptom relief when shown absolute vs. relative risks, although differences in MAR were statistically significant only for breast cancer risk. MARs for breast cancer and heart attack exceeded reported rates in most cases. CONCLUSIONS: Many women may be willing to accept risks to control vasomotor symptoms. However, describing risks in different, but technically equivalent, ways affects women's willingness to trade risks for benefits.
BACKGROUND: Evidence that long-term hormone therapy (HT) may increase the risk of serious adverse events led to a sharp reduction in all HT use, including short-term use for vasomotor symptom relief. We estimated women's willingness to accept adverse event risks in exchange for vasomotor symptom relief when risks are communicated in absolute vs. relative terms. METHODS: We developed a conjoint survey to elicit preferences across benefit and risk attributes of HT. The survey was administered via the Internet to 523 U.S. women aged 46-60 years. Participants evaluated pairs of hypothetical treatments and indicated preferences using a Likert-based scale. Risks were presented in absolute and relative terms. Satisfaction scores for HT risks and benefits were estimated using random-effects, ordered-probit regression. Maximum acceptable risk (MAR) was calculated as the increase in risk that reduces the satisfaction score for a given level of HT benefit to 0. RESULTS: For both risk versions, the least important attribute (smallest difference in satisfaction scores) was night sweat frequency; heart attack risk was the most important (largest difference in scores). Participants were more willing to accept risks in return for symptom relief when shown absolute vs. relative risks, although differences in MAR were statistically significant only for breast cancer risk. MARs for breast cancer and heart attack exceeded reported rates in most cases. CONCLUSIONS: Many women may be willing to accept risks to control vasomotor symptoms. However, describing risks in different, but technically equivalent, ways affects women's willingness to trade risks for benefits.
Authors: F Reed Johnson; George Van Houtven; Semra Ozdemir; Steve Hass; Jeff White; Gordon Francis; David W Miller; J Theodore Phillips Journal: J Neurol Date: 2009-04-27 Impact factor: 4.849