M. Escorihuela Sahún, L. M. Esteban Escaño, Á. Borque-Fernando, G. Sanz Sáiz
This paper presents a library in R to validate and evaluate predictive models analytically and graphically. It includes functions such as CUC_plot, CUC_table, Efficacy, Efficacy_curve and Efficacy_test, which allow the construction of the clinical utility curve, a table of values, the efficacy of biomarkers, their efficacy curve and a comparative test between biomarkers. The objective is to define a biomarker that predicts disease events. The CUC_plot function displays the clinical utility curve, analyzing the benefit of a biomarker with a cut-off point, representing false negatives and avoided treatments. CUC_table provides the corresponding numerical values. Efficacy, defined as the difference between avoided treatments and false negatives, is evaluated with the functions Efficacy (numerical value) and Efficacy_curve (graph). Efficacy_test compares the efficacy of two biomarkers using McNemar's test or proportions test depending on whether the data are paired or not.
Palabras clave: Biomarkers, clinical utility , efficacy, predective models
Programado
Software II
10 de junio de 2025 15:30
Sala VIP Jaume Morera i Galícia