Predicting Italian EQ-5D-3L Health State Utilities From UK Values: Validation of a Nonlinear Utility Adjustment Model Using Real-World Evidence
Author(s)
Abdul Jabbar Omar Alsaleh, MA, MS, PharmD1, Elisheva LEW, MS2, Jérôme Msihid, MSc3, Lorenzo Pradelli, MD4, Massimiliano Povero, PhD4, Charlie Nicholls, PhD5.
1Sanofi, Milan, Italy, 2Sanofi, Paris, France, 3Sanofi, Gentilly, France, 4AdRes HE&OR, Torino, Italy, 5Sanofi, Reading, United Kingdom.
1Sanofi, Milan, Italy, 2Sanofi, Paris, France, 3Sanofi, Gentilly, France, 4AdRes HE&OR, Torino, Italy, 5Sanofi, Reading, United Kingdom.
OBJECTIVES: The direct transfer of unadjusted European quality of life five-dimension 3-level (EQ-5D-3L) utility values between countries may misrepresent local health preferences, potentially biasing health technology assessments. This study aimed to validate an established method that translates published mean EQ-5D-3L utilities based on the United Kingdom (UK) value set to reflect Italian societal preferences using real-world data.
METHODS: An existing nonlinear model utilizing 243 EQ-5D-3L health state utility weights (HSUWs) from UK [Dolan (1997)] and Italian [Scalone (2013)] value sets was employed. The Italian HSUWs were regressed against the UK HSUWs, yielding the final model: HSUWjIT=0.3388+1.026×HSUWjUK−0.42×(HSUWjUK)2, (j=1−243 health states; IT=Italy). The model was extended to mean utilities: UTILITYIT=0.3388+1.026×UTILITYUK−0.42×[(UTILITYUK)2+VarUTILITY-UK]. A targeted literature review was conducted to identify studies reporting both UK and Italian EQ-5D-3L utility values. The model was applied to the published UK mean utilities, and the predicted Italian utilities were compared with the published Italian mean utilities.
RESULTS: Two studies were included: Mozzi (2016), a survey-based study on the health-related quality of life in patients with Crohn’s disease (n=500), and Joelson (2021), a large observational study (n=27,328) that reported pre- and post-operative (OP) utilities in patients with spinal stenosis (SS) or disc herniation (HNP). Both studies indicated that Italian utilities were higher than UK utilities, emphasizing the need for country-specific adjustments. The model demonstrated strong predictive accuracy (R²=0.8723), exhibiting close alignment between predicted and published Italian utilities: in Mozzi (2016) (0.74 vs. 0.76); Joelson (2021) pre-OP SS (0.67 vs. 0.68), post-OP SS (0.81 vs. 0.80), pre-OP HNP (0.59 vs. 0.59), and post-OP HNP (0.86 vs. 0.86).
CONCLUSIONS: The validated model provides a simple yet effective method for adapting UK mean EQ-5D-3L utilities to Italian settings, with high predictive accuracy. The model offers a practical solution to economic assessments by transposing utilities between different countries.
METHODS: An existing nonlinear model utilizing 243 EQ-5D-3L health state utility weights (HSUWs) from UK [Dolan (1997)] and Italian [Scalone (2013)] value sets was employed. The Italian HSUWs were regressed against the UK HSUWs, yielding the final model: HSUWjIT=0.3388+1.026×HSUWjUK−0.42×(HSUWjUK)2, (j=1−243 health states; IT=Italy). The model was extended to mean utilities: UTILITYIT=0.3388+1.026×UTILITYUK−0.42×[(UTILITYUK)2+VarUTILITY-UK]. A targeted literature review was conducted to identify studies reporting both UK and Italian EQ-5D-3L utility values. The model was applied to the published UK mean utilities, and the predicted Italian utilities were compared with the published Italian mean utilities.
RESULTS: Two studies were included: Mozzi (2016), a survey-based study on the health-related quality of life in patients with Crohn’s disease (n=500), and Joelson (2021), a large observational study (n=27,328) that reported pre- and post-operative (OP) utilities in patients with spinal stenosis (SS) or disc herniation (HNP). Both studies indicated that Italian utilities were higher than UK utilities, emphasizing the need for country-specific adjustments. The model demonstrated strong predictive accuracy (R²=0.8723), exhibiting close alignment between predicted and published Italian utilities: in Mozzi (2016) (0.74 vs. 0.76); Joelson (2021) pre-OP SS (0.67 vs. 0.68), post-OP SS (0.81 vs. 0.80), pre-OP HNP (0.59 vs. 0.59), and post-OP HNP (0.86 vs. 0.86).
CONCLUSIONS: The validated model provides a simple yet effective method for adapting UK mean EQ-5D-3L utilities to Italian settings, with high predictive accuracy. The model offers a practical solution to economic assessments by transposing utilities between different countries.
Conference/Value in Health Info
2025-11, ISPOR Europe 2025, Glasgow, Scotland
Value in Health, Volume 28, Issue S2
Code
PCR189
Topic
Patient-Centered Research
Topic Subcategory
Health State Utilities
Disease
No Additional Disease & Conditions/Specialized Treatment Areas