Hareth Al-Janabi1, Terry N Flynn2, Joanna Coast1. 1. Health Economics Unit, University of Birmingham, Birmingham, United Kingdom (HA-J, JC) 2. Centre for the Study of Choice (CenSoC), University of Technology Sydney, Sydney, Australia (TNF)
Abstract
BACKGROUND: There is growing interest in incorporating the effects of interventions on unpaid carers in economic evaluation. Current methods focus on using health measures (which neglect important aspects of care-related quality of life) or using carer-specific sum score measures (which are not preference based). OBJECTIVE: To estimate preference-based index values for a profile measure of the caring experience (the Carer Experience Scale). METHODS: Eighteen profiles from the Carer Experience Scale were included in a best-worst scaling experiment. In each profile, respondents were asked to pick the best and worst attribute level from the profiles. The choice task was completed in a postal questionnaire by 162 unpaid carers of older people from 5 geographical locations in the United Kingdom. Logistic regression was used to estimate utility weights for the attribute levels of the Carer Experience Scale. Alternative modeling assumptions were employed to determine the stability of the parameter estimates. These parameter estimates were rescaled so that the profile index values for the Carer Experience Scale lay on a 0-to-100 scale. RESULTS: The results indicate that low levels of "activities" and "getting on" result in larger decrements to utility than other attributes of the caring experience. In general, greater value is placed on differences between the bottom and middle levels of attributes than between the middle and top levels. Alternative modeling approaches had a negligible effect on the index values. CONCLUSION: The index values reported in this study offer a new preference-based approach to incorporating the effects on carers in economic evaluation, focusing on care (rather than health)-related quality of life.
BACKGROUND: There is growing interest in incorporating the effects of interventions on unpaid carers in economic evaluation. Current methods focus on using health measures (which neglect important aspects of care-related quality of life) or using carer-specific sum score measures (which are not preference based). OBJECTIVE: To estimate preference-based index values for a profile measure of the caring experience (the Carer Experience Scale). METHODS: Eighteen profiles from the Carer Experience Scale were included in a best-worst scaling experiment. In each profile, respondents were asked to pick the best and worst attribute level from the profiles. The choice task was completed in a postal questionnaire by 162 unpaid carers of older people from 5 geographical locations in the United Kingdom. Logistic regression was used to estimate utility weights for the attribute levels of the Carer Experience Scale. Alternative modeling assumptions were employed to determine the stability of the parameter estimates. These parameter estimates were rescaled so that the profile index values for the Carer Experience Scale lay on a 0-to-100 scale. RESULTS: The results indicate that low levels of "activities" and "getting on" result in larger decrements to utility than other attributes of the caring experience. In general, greater value is placed on differences between the bottom and middle levels of attributes than between the middle and top levels. Alternative modeling approaches had a negligible effect on the index values. CONCLUSION: The index values reported in this study offer a new preference-based approach to incorporating the effects on carers in economic evaluation, focusing on care (rather than health)-related quality of life.