Optimal cutoff points for classification in diagnostic studies: new contributions and software development

• Autores: Mónica López Ratón
• Directores de la Tesis: Carmen María Cadarso Suárez (dir. tes.), Elisa M. Molanes (codir. tes.)
• Lectura: En la Universidade de Santiago de Compostela ( España ) en 2016
• Idioma: inglés
• Tribunal Calificador de la Tesis: Antonio Martín Andrés (presid.), Pablo García Tahoces (secret.), Francisco Gude Sampedro (voc.), María José Rodríguez Álvarez (voc.), Pablo Martínez Camblor (voc.)
• Materias:
  • Matemáticas
    • Estadística
      • Análisis de datos
      • Técnicas de inferencia estadística
  • Ciencias de la vida
    • Biomatemáticas
      • Bioestadística
• Enlaces
  • Tesis en acceso abierto en: MINERVA
• Resumen

  • Diagnostic tests are often used for discriminating between healthy and diseased populations. In continuous diagnostic tests (take values in a continuous range), it is useful to select a cutpoint or discrimination value c that defines the positive (patient is classified as diseased) and negative (patient is classified as healthy) tests results, such in general, individuals with a diagnostic test value of c or higher are classified as diseased. The objective consists in to select the better optimal cutpoint c, the “optimal” cutpoint. Several strategies have been proposed in the literature for selecting optimal cutpoints in diagnostic tests, depending on the underlying reason for this choice.

    The main objective of this doctoral thesis is to study and review the different criteria for selecting optimal cutpoints, mainly based on their application in clinical field, development of new estimation and inference techniques of the optimal cutpoint and implement user-friendly software in R that includes all these techniques.