Publication
Machine learning algorithms for predicting EQ-5D full health state in systemic lupus erythematosus
| datacite.subject.fos | Ciências Médicas::Biotecnologia Médica | pt_PT |
| dc.contributor.advisor | Parodis, Ioannis | |
| dc.contributor.author | Botto, João Ricardo da Costa Monteiro | |
| dc.date.accessioned | 2024-03-14T12:27:22Z | |
| dc.date.available | 2025-03-01T01:31:00Z | |
| dc.date.issued | 2024-01-30 | |
| dc.date.submitted | 2023-12 | |
| dc.description.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. | pt_PT |
| dc.description.abstract | Objetivos: Determinar fatores associados ao estado de saúde perfeita (ESP) do EQ-5D no lúpus eritematoso sistémico (LES) antes e depois da intervenção num ensaio clínico, recorrendo a algoritmos de machine learning. Métodos: Realizámos uma análise post-hoc de dados de dois ensaios clínicos de fase III do belimumab (BLISS-52, BLISS-76). Foram recolhidas variáveis demográficas, laboratoriais e clínicas, tendo sido utilizado o algoritmo Monte Carlo Feature Selection, posteriormente refinado considerando colinearidade e relevância/perícia clínica. Os seguintes modelos foram usados: support vector machine with radial basis function kernel (SVMRadial), least absolute shrinkage and selection operator (LASSO), neural network (NNet), e logistic regression (LR). Resultados: Numa coorte de 1642 doentes com LES, 12.9% reportaram FHS na baseline e 23.1% na semana 52. As variáveis selecionadas foram idade, sexo, ascendência Asiática, e clinical Systemic Lupus Erythematosus Disease Activity Index-2000, Safety of Estrogens in Lupus National Assessment-SLEDAI Physician Global Assessment, e rácio proteína/creatinina urinário (RPCU) na baseline. Exclusivamente para os modelos da semana 52, o índice do EQ 5D na baseline também foi selecionado. Os modelos preditores do ESP demonstraram um desempenho comparável na baseline e na semana 52. Na baseline, a máxima área sob a curva foi vista nos modelos LASSO e LR (0.73), enquanto na semana 52 foi vista nos modelos LASSO e NNet (0.77). O valor preditivo negativo foi particularmente alto para todos os modelos (0.88–0.94). A calibração mostrou uma ligeira melhoria nos modelos da semana 52. Conclusão: Idade avançada, sexo feminino, ascendência não-Asiática, alta atividade de doença, e baixo RPCU foram associados, através de machine learning, à não-experiência de um ESP em doentes com LES na baseline e na semana 52. O índice do EQ-5D na baseline constituiu a variável mais informativa para prever a experiência de um ESP na semana 52. | pt_PT |
| dc.identifier.tid | 203550897 | pt_PT |
| dc.identifier.uri | http://hdl.handle.net/10400.14/44287 | |
| dc.language.iso | eng | pt_PT |
| dc.subject | Systemic lupus erythematosus | pt_PT |
| dc.subject | Quality of life | pt_PT |
| dc.subject | EQ-5D | pt_PT |
| dc.subject | Machine learning | pt_PT |
| dc.subject | Lúpus eritematoso sistémico | pt_PT |
| dc.subject | Qualidade de vida | pt_PT |
| dc.title | Machine learning algorithms for predicting EQ-5D full health state in systemic lupus erythematosus | pt_PT |
| dc.type | master thesis | |
| dspace.entity.type | Publication | |
| rcaap.rights | openAccess | pt_PT |
| rcaap.type | masterThesis | pt_PT |
| thesis.degree.name | Mestrado em Engenharia Biomédica | pt_PT |