Mapping the disease-specific LupusQoL to the SF-6D
Rachel Meacock1, Mark Harrison, Kathleen McElhone
1The University of Manchester, 4.311 Jean McFarlane Building, Oxford Road, Manchester, M15 4QE, UK, rachel.meacock@manchester.ac.uk.
Summary
A new algorithm predicts SF-6D utility scores using LupusQoL data for systemic lupus erythematosus (SLE) patients. This method offers a reliable way to estimate health utility values from non-preference-based questionnaires.
Area of Science:
- Rheumatology
- Health Economics
- Psychometrics
Background:
- Systemic lupus erythematosus (SLE) impacts patient quality of life.
- Accurate health utility scores are crucial for economic evaluations.
- The LupusQoL is a non-preference-based measure of quality of life in SLE patients.
Purpose of the Study:
- To develop and validate a mapping algorithm predicting SF-6D utility scores from LupusQoL responses.
- To assess the generalizability of the developed algorithm.
Main Methods:
- Data from 320 SLE patients (UK) were used to build the prediction model.
- Ordinary least squares (OLS) regression was employed to map LupusQoL domains to SF-6D scores.
- External validation was performed on an independent dataset of 113 SLE patients.
Main Results:
- The final model incorporated four LupusQoL domains: physical health, pain, emotional health, and fatigue.
- The algorithm demonstrated good predictive performance with R-squared values of 0.7219 (estimation) and 0.7431 (validation).
- Low prediction errors (MAE: 0.0557/0.0528, RMSE: 0.0706/0.0663) were observed in both datasets.
Conclusions:
- A generalizable method for estimating SF-6D utility values from LupusQoL data was successfully derived.
- OLS regression proved suitable for developing this mapping algorithm, supported by the normality of SF-6D data.
- This approach facilitates health utility estimation in SLE patients using non-preference-based instruments.
