TY - JOUR
T1 - Exploring the importance of controlling heteroskedasticity and heterogeneity in health valuation
T2 - a case study on Dutch EQ-5D-5L
AU - Karim, Suzana
AU - Craig, Benjamin M.
AU - Groothuis-Oudshoorn, Catharina G.M.
N1 - Funding Information:
The authors would like to thank the EuroQol Group for funding this study (Grant EQ Project 2016220, EQ Project 138-2020RA. The authors would also like to thank Dr. Terry Flynn and Dr. Sander Arons for their assistance with the project's proposal, experimental design, and primary data collection.
Funding Information:
EuroQol Group funded this study Grant EQ Project 2016220, EQ Project 138-2020RA.
Funding Information:
Catharina G.M. Groothuis-Oudshoorn and Suzana Karim received grants from the EuroQol Research Foundation. Dr. Benjamin M Craig is a member of the EuroQol Research Foundation.
Publisher Copyright:
© 2022, The Author(s).
PY - 2022/12
Y1 - 2022/12
N2 - Background: Respondents in a health valuation study may have different sources of error (i.e., heteroskedasticity), tastes (differences in the relative effects of each attribute level), and scales (differences in the absolute effects of all attributes). Although prior studies have compared values by preference-elicitation tasks (e.g., paired comparison [PC] and best–worst scaling case 2 [BWS]), no study has yet controlled for heteroskedasticity and heterogeneity (taste and scale) simultaneously in health valuation.Methods: Preferences on EQ-5D-5L profiles were elicited from a random sample of 380 adults from the general population of the Netherlands, using 24 PC and 25 BWS case 2 tasks. To control for heteroskedasticity and heterogeneity (taste and scale) simultaneously, we estimated Dutch EQ-5D-5L values using conditional, heteroskedastic, and scale-adjusted latent class (SALC) logit models by maximum likelihood.Results: After controlling for heteroskedasticity, the PC and BWS values were highly correlated (Pearson's correlation: 0.9167, CI: 0.9109–0.9222) and largely agreed (Lin's concordance: 0.7658, CI: 0.7542–0.7769) on a pits scale. In terms of preference heterogeneity, some respondents (mostly young men) failed to account for any of the EQ-5D-5L attributes (i.e., garbage class), and others had a lower scale (59%; p-value: 0.123). Overall, the SALC model produced a consistent Dutch EQ-5D-5L value set on a pits scale, like the original study (Pearson's correlation:0.7295; Lin's concordance: 0.6904).Conclusions: This paper shows the merits of simultaneously controlling for heteroskedasticity and heterogeneity in health valuation. In this case, the SALC model dispensed with a garbage class automatically and adjusted the scale for those who failed the PC dominant task. Future analysis may include more behavioral variables to better control heteroskedasticity and heterogeneity in health valuation. Highlights: The Dutch EQ-5D-5L values based on paired comparison [PC] and best-worst scaling [BWS] responses were highly correlated and largely agreed after controlling for heteroskedasticity.Controlling for taste and scale heterogeneity simultaneously enhanced the Dutch EQ-5D-5Lvalues by automatically dispensing with a garbage class and adjusting the scale for those who failed the dominant task.After controlling for heteroskedasticity and heterogeneity, this study produced Dutch EQ-5D-5L values on a pits scale moderately concordant with the original values.
AB - Background: Respondents in a health valuation study may have different sources of error (i.e., heteroskedasticity), tastes (differences in the relative effects of each attribute level), and scales (differences in the absolute effects of all attributes). Although prior studies have compared values by preference-elicitation tasks (e.g., paired comparison [PC] and best–worst scaling case 2 [BWS]), no study has yet controlled for heteroskedasticity and heterogeneity (taste and scale) simultaneously in health valuation.Methods: Preferences on EQ-5D-5L profiles were elicited from a random sample of 380 adults from the general population of the Netherlands, using 24 PC and 25 BWS case 2 tasks. To control for heteroskedasticity and heterogeneity (taste and scale) simultaneously, we estimated Dutch EQ-5D-5L values using conditional, heteroskedastic, and scale-adjusted latent class (SALC) logit models by maximum likelihood.Results: After controlling for heteroskedasticity, the PC and BWS values were highly correlated (Pearson's correlation: 0.9167, CI: 0.9109–0.9222) and largely agreed (Lin's concordance: 0.7658, CI: 0.7542–0.7769) on a pits scale. In terms of preference heterogeneity, some respondents (mostly young men) failed to account for any of the EQ-5D-5L attributes (i.e., garbage class), and others had a lower scale (59%; p-value: 0.123). Overall, the SALC model produced a consistent Dutch EQ-5D-5L value set on a pits scale, like the original study (Pearson's correlation:0.7295; Lin's concordance: 0.6904).Conclusions: This paper shows the merits of simultaneously controlling for heteroskedasticity and heterogeneity in health valuation. In this case, the SALC model dispensed with a garbage class automatically and adjusted the scale for those who failed the PC dominant task. Future analysis may include more behavioral variables to better control heteroskedasticity and heterogeneity in health valuation. Highlights: The Dutch EQ-5D-5L values based on paired comparison [PC] and best-worst scaling [BWS] responses were highly correlated and largely agreed after controlling for heteroskedasticity.Controlling for taste and scale heterogeneity simultaneously enhanced the Dutch EQ-5D-5Lvalues by automatically dispensing with a garbage class and adjusting the scale for those who failed the dominant task.After controlling for heteroskedasticity and heterogeneity, this study produced Dutch EQ-5D-5L values on a pits scale moderately concordant with the original values.
KW - Best–worst scaling
KW - EQ-5D
KW - Health valuation
KW - Heteroskedasticity
KW - Scale heterogeneity
UR - http://www.scopus.com/inward/record.url?scp=85130728609&partnerID=8YFLogxK
U2 - 10.1186/s12955-022-01989-9
DO - 10.1186/s12955-022-01989-9
M3 - Article
C2 - 35614472
AN - SCOPUS:85130728609
SN - 1477-7525
VL - 20
JO - Health and quality of life outcomes
JF - Health and quality of life outcomes
IS - 1
M1 - 85
ER -