Article
Direct and indirect mapping algorithms of the EORTC QLQ-C30 to the EQ-5D-5L utility index in patients with metastatic breast cancer
Search Medline for
Authors
-
Pimrapat Gebert
- Charité – Universitätsmedizin Berlin, corporate member of Freie Universität Berlin and Humboldt-Universität zu Berlin, Berlin, Germany
-
Felix Fischer
- Charité – Universitätsmedizin Berlin, corporate member of Freie Universität Berlin and Humboldt-Universität zu Berlin, Department of Psychosomatic Medicine, Center for Internal Medicine and Dermatology, Berlin, Germany
-
Christoph Paul Klapproth
- Charité – Universitätsmedizin Berlin, corporate member of Freie Universität Berlin and Humboldt-Universität zu Berlin, Department of Psychosomatic Medicine, Center for Internal Medicine and Dermatology, Berlin, Germany
-
Anna Maria Hage
- Charité – Universitätsmedizin Berlin, Corporate Member of Freie Universität Berlin and Humboldt Universität zu Berlin, Department of Gynecology with Breast Center, Berlin, Germany
-
Adam David Dordevic
- Charité – Universitätsmedizin Berlin, Corporate Member of Freie Universität Berlin and Humboldt Universität zu Berlin, Department of Gynecology with Breast Center, Berlin, Germany
-
Ulrike Grittner
- Charité – Universitätsmedizin Berlin, corporate member of Freie Universität Berlin and Humboldt-Universität zu Berlin, Berlin, Germany
-
Maria Margarete Karsten
- Charité – Universitätsmedizin Berlin, Corporate Member of Freie Universität Berlin and Humboldt Universität zu Berlin, Department of Gynecology with Breast Center, Berlin, Germany
| Published: | September 6, 2024 |
|---|
Outline
Text
Introduction: Cost-effectiveness analyses in clinical studies require health utility data. Unfortunately, the collection of utility information with e.g. the instrument EQ-5D is not consistently prioritized in oncological trials. Instead, non-preference-based measures like the EORTC QLQ-C30 are commonly used but are lacking utility information. This study developed and compared the algorithm performances of a direct and indirect mapping algorithm for estimating the EQ-5D-5L utility index using the German value set [1] from EORTC QLQ-C30.
Methods: We included 1,900 observations from 879 patients with metastatic breast cancer from the PRO B study [2] after removing missing data of both questionnaires. Using computer-generated random numbers, we split the observations into two sets. The data of 70% (1,328 observations) were allocated to the estimation set for developing the mapping model, while the remaining 30% (572 observations) formed the validation set for testing the model's performance. Direct mapping involved Adjusted Limited Dependent Variable Mixture Models (ALDVMM) and Tobit regression, while indirect mapping employed Generalized Ordered Logit model. Model performance was assessed using mean absolute error (MAE), root mean squared error (RMSE), and a comparison of mean predicted and mean observed values.
Results: Direct mapping consistently outperformed indirect mapping. Among the direct algorithms, ALDVMM demonstrated superior performance in both estimation and validation sets. The mean differences between predicted and observed utilities for the estimation and validation sets were 0.011 and 0.013, while the RMSE values were 0.126 for both, and the MAE values were 0.081 and 0.083, respectively.
Discussion: We developed a mapping algorithm for estimating the EQ-5D-5L utility index and compared direct and indirect mapping methods. Our results imply that ALDVMM is a reliable method for mapping EQ-5D-5L health states to utility values because of its flexibility, potential to capture complex relationships, and robustness to model misspecification. This has also been reported in other mapping studies [3], [4].
Conclusion: This study developed a robust algorithm for estimating EQ-5D-5L utility German tariff in patients with metastatic breast cancer, derived from the EORTC QLQ-C30. The direct mapping approach using ALDVMM demonstrated superior performance compared to Tobit or indirect mapping.
The authors declare that they have no competing interests.
The authors declare that a positive ethics committee vote has been obtained.
References
- 1.
- Ludwig K, Graf von der Schulenburg JM, Greiner W. German Value Set for the EQ-5D-5L. Pharmacoeconomics. 2018;36:663-674. DOI: 10.1007/s40273-018-0615-8
- 2.
- Karsten MM, Kuhn F, Pross T, Blohmer JU, Hage AM, Fischer F, et al. PRO B: evaluating the effect of an alarm-based patient-reported outcome monitoring compared with usual care in metastatic breast cancer patients-study protocol for a randomised controlled trial. Trials. 2021;22:666. DOI: 10.1186/s13063-021-05642-6
- 3.
- Aghdaee M, Gu Y, Sinha K, Parkinson B, Sharma R, Cutler H. Mapping the Patient-Reported Outcomes Measurement. Information System (PROMIS-29) to EQ-5D-5L. Pharmacoeconomics. 2023;41:187-198. DOI: 10.1007/s40273-022-01157-3
- 4.
- Gray LA, Hernandez Alava M, Wailoo AJ. Mapping the EORTC QLQ-C30 to EQ-5D-3L in patients with breast cancer. BMC Cancer. 2021;21:1237. DOI: 10.1186/s12885-021-08964-5
