Machine learning algorithms for predicting EQ-5D full health state in systemic lupus erythematosus

Abstract

Objectives: To determine factors associated with EQ-5D full health state (FHS) in systemic lupus erythematosus (SLE) before and after a trial intervention, resorting to machine learning algorithms. Methods: We conducted a post-hoc analysis of data from two phase III clinical trials of belimumab (BLISS-52, BLISS-76). Demographic, laboratory, and clinical features were retrieved, and the Monte Carlo Feature Selection algorithm was employed, further refined upon consideration of collinearity and clinical relevance/expertise. The models used were support vector machine with radial basis function kernel (SVMRadial), least absolute shrinkage and selection operator (LASSO), neural network (NNet), and logistic regression (LR). Results: In a cohort of 1642 SLE patients, 12.9% reported FHS at baseline and 23.1% at week 52. Selected features were age, sex, Asian ancestry, baseline clinical Systemic Lupus Erythematosus Disease Activity Index-2000, Safety of Estrogens in Lupus National Assessment-SLEDAI Physician Global Assessment, and urine protein/creatinine ratio (UPCR), and baseline EQ-5D utility index score (week-52 models only). The models predicting FHS demonstrated comparable performance at baseline and week 52, where for baseline a maximum area under the curve of 0.73 was seen for the LASSO and LR models, versus a 0.77 maximum for the week-52 LASSO and NNet models. Particularly high for all models was the negative predictive value (0.88–0.94). Calibration showed marginal improvement in week-52 models. Conclusion: Machine learning identified older age, female sex, non-Asian ancestry, high disease activity, and low UPCR to be associated with a lack of FHS experience in SLE patients at baseline and week 52. Baseline EQ-5D utility index constituted the most informative feature for predicting FHS experience at week 52.

Date of Award	30 Jan 2024
Original language	English
Awarding Institution	Universidade Católica Portuguesa
Supervisor	Ioannis Parodis (Supervisor)