ALTERNATIVE METHODOLOGICAL PROPOSALS BAYESIAN, SEMIPARAMETRICS AND EMPIRICAL PLAUSIBILITY IN LONGITUDINAL DATA, ANALYSIS OF SURVIVAL AND MODELLING OF HRQOL (Q3153291)

From EU Knowledge Graph
Revision as of 16:04, 12 October 2021 by DG Regio (talk | contribs) (‎Created claim: summary (P836): THE OBJECTIVES OF THIS PROJECT FOCUS ON THREE KEY METHODOLOGICAL ASPECTS: BAYESIAN PROPOSALS, SEMIPARAMETRICS AND EMPIRICAL PLAUSIBILITY, ALL IN CONTEXTS OF LONGITUDINAL DATA, ANALYSIS OF SURVIVAL AND MODELLING OF HEALTH-RELATED QUALITY OF LIFE (CVRS). FIRST, IN THE AREA OF LONGITUDINAL DATA, WE WILL DELVE INTO THE ECONOMETRIC MODELS IN PSEUDO-PANELS OR SYNTHETIC PANELS FOR THE CASE OF PSEUDO PANELS THAT, BY CONSTRUCTION, PRESENT A TEMPORAL DEPE...)
Jump to navigation Jump to search
Project Q3153291 in Spain
Language Label Description Also known as
English
ALTERNATIVE METHODOLOGICAL PROPOSALS BAYESIAN, SEMIPARAMETRICS AND EMPIRICAL PLAUSIBILITY IN LONGITUDINAL DATA, ANALYSIS OF SURVIVAL AND MODELLING OF HRQOL
Project Q3153291 in Spain

    Statements

    0 references
    0 references
    25,712.5 Euro
    0 references
    51,425.0 Euro
    0 references
    50.0 percent
    0 references
    30 December 2016
    0 references
    29 September 2021
    0 references
    UNIVERSIDAD DEL PAIS VASCO/EUSKAL HERRIKO UNIBERTSITATEA
    0 references
    0 references

    43°15'46.80"N, 2°56'6.00"W
    0 references
    48020
    0 references
    LOS OBJETIVOS DE ESTE PROYECTO SE CENTRAN EN TRES ASPECTOS METODOLOGICOS FUNDAMENTALES: PROPUESTAS BAYESIANAS, SEMIPARAMETRICAS Y DE VEROSIMILITUD EMPIRICA, TODOS ELLOS EN CONTEXTOS DE DATOS LONGITUDINALES, ANALISIS DE SUPERVIVENCIA Y MODELIZACION DE LA CALIDAD DE VIDA RELACIONADA CON LA SALUD (CVRS). EN PRIMER LUGAR, EN EL AREA DE DATOS LONGITUDINALES, PROFUNDIZAREMOS SOBRE LOS MODELOS ECONOMETRICOS EN PSEUDO-PANELES O PANELES SINTETICOS PARA EL CASO DE PSEUDO PANELES QUE, POR CONSTRUCCION, PRESENTAN UNA DEPENDENCIA TEMPORAL. ANALIZAREMOS EL MODELO GENERAL PARA PSEUDO-PANELES INDEPENDIENTES Y LO ADAPTAREMOS DE FORMA QUE PERMITA DEPENDENCIA TEMPORAL EN CADA PSEUDO-PANEL, MODELIZANDO ESTA DEPENDENCIA DE ACUERDO CON CRITERIOS DE BONDAD DE AJUSTE ESPECIFICOS Y CON CRITERIOS DICTADOS POR LAS CARACTERISTICAS DE DEPENDENCIA ESPECIFICAS DE LOS DATOS UTILIZADOS. DESARROLLAREMOS LOS PROGRAMAS ADECUADOS PARA LLEVAR A CABO LA ESTIMACION EN ESTOS MODELOS Y APLICAREMOS LAS TECNICAS DESARROLLADAS A DATOS REALES, ESPECIALMENTE EN BIOLOGIA Y DATOS DEL MERCADO DE TRABAJO VASCO. EN SEGUNDO LUGAR, EN EL AREA DE ANALISIS DE SUPERVIVENCIA, EXTENDEREMOS LAS PROPUESTAS METODOLOGICAS DE VEROSIMILITUD EMPIRICA A PROBLEMAS DE INFERENCIA CON RESTRICCIONES DE ORDEN, ESTUDIANDO LAS DISTRIBUCIONES ASINTOTICAS DE LOS ESTADISTICOS DE CONTRASTES Y EL COMPORTAMIENTO ASINTOTICO DE LOS ESTIMADORES DE VEROSIMILITUD EMPIRICA OBTENIDOS BAJO RESTRICCIONES DE ORDEN, PONIENDO ESPECIAL ENFASIS EN LOS ASPECTOS COMPUTACIONALES NO TRIVIALES DE LOS METODOS DE VEROSIMILITUD EMPIRICA. COMPARAREMOS FUNCIONES DE DISTRIBUCION Y RIESGOS COMPETITIVOS UTILIZANDO ESTA METODOLOGIA Y APLICAREMOS LOS RESULTADOS EN EJEMPLOS REALES EN FIABILIDAD, SALUD PUBLICA, ECONOMIA Y FINANZAS. FINALMENTE, ESTUDIAREMOS PROPUESTAS ALTERNATIVAS EN LA MODELIZACION ESTADISTICA DE LA CVRS, ESPECIALMENTE EN LO QUE SE RELACIONA CON RESULTADOS PERCIBIDOS POR EL PACIENTE, APLICANDO LAS MISMAS A INDICES DE SALUD Y CALIDAD DE VIDA RELACIONADA CON LA SALUD. ESTUDIAREMOS SU DISTRIBUCION Y MODELIZACION ESTADISTICA, PROPONIENDO MEJORAS EN LA CONSTRUCCION DE INDICES PARA MEDIR ESTOS RESULTADOS, Y COMPARAREMOS LAS DISTINTAS ALTERNATIVAS METODOLOGICAS PARA SU ANALISIS. PROPONDREMOS MODELOS AVANZADOS DE TECNICAS DE SUAVIZADO (MAG, P-SPLINES) PARA MODELOS APLICADOS AL ANALISIS DE RESULTADOS PERCIBIDOS POR EL PACIENTE. FINALMENTE, CONTEXTUALIZAREMOS LOS RESULTADOS OBTENIDOS, DAREMOS PAUTAS PARA LA SELECCION DE ALTERNATIVAS ADECUADAS, Y RECOMENDACIONES Y CONCLUSIONES BASADAS EN LOS ASPECTOS MAS RELEVANTES DESDE EL PUNTO DE VISTA ESTADISTICO Y CLINICO DE LA APLICACION. TAMBIEN DESARROLLAREMOS, IMPLEMENTAREMOS Y VALIDAREMOS MODELOS PREDICTIVOS UTILES EN LA PRACTICA CLINICA HOSPITALARIA, IMPLEMENTANDO HERRAMIENTAS TECNOLOGICAS A PARTIR DE LOS MISMOS. PARA ELLO, DESARROLLAREMOS MODELOS PREDICTIVOS EN EVOLUCION CLINICA, MORTALIDAD, SUPERVIVENCIA O CAMBIOS EN LA CVRS PARA PACIENTES CON ENFERMEDADES CRONICAS, PROPONIENDO TRANSFORMACIONES Y CATEGORIZACIONES OPTIMAS DE LAS VARIABLES PREDICTIVAS QUE PERMITAN SU CORRECTA INTRODUCCION EN EL MODELO PREDICTIVO. LOS DISTINTOS MODELOS PREDICTIVOS SERAN EVALUADOS Y COMPARADOS MEDIANTE DIFERENTES TECNICAS DE ANALISIS (REGRESION LINEAL, REGRESION LOGISTICA BINARIA, REGRESION LOGISTICA ORDINAL, REGRESION DE POISSON, ARBOLES DE REGRESION Y CLASIFICACION, REDES NEURONALES, ENTRE OTRAS), Y DIFERENTES CRITERIOS (BONDAD DE AJUSTE, CALIBRACION, CAPACIDAD PREDICTIVA, ENTRE OTROS). (Spanish)
    0 references
    THE OBJECTIVES OF THIS PROJECT FOCUS ON THREE KEY METHODOLOGICAL ASPECTS: BAYESIAN PROPOSALS, SEMIPARAMETRICS AND EMPIRICAL PLAUSIBILITY, ALL IN CONTEXTS OF LONGITUDINAL DATA, ANALYSIS OF SURVIVAL AND MODELLING OF HEALTH-RELATED QUALITY OF LIFE (CVRS). FIRST, IN THE AREA OF LONGITUDINAL DATA, WE WILL DELVE INTO THE ECONOMETRIC MODELS IN PSEUDO-PANELS OR SYNTHETIC PANELS FOR THE CASE OF PSEUDO PANELS THAT, BY CONSTRUCTION, PRESENT A TEMPORAL DEPENDENCE. WE WILL ANALYSE THE GENERAL MODEL FOR INDEPENDENT PSEUDO-PANELS AND ADAPT IT IN A WAY THAT ALLOWS TEMPORAL DEPENDENCE IN EACH PSEUDO-PANEL, MODELING THIS DEPENDENCE ACCORDING TO SPECIFIC ADJUSTMENT GOODNESS CRITERIA AND CRITERIA DICTATED BY THE SPECIFIC DEPENDENCY CHARACTERISTICS OF THE DATA USED. WE WILL DEVELOP THE APPROPRIATE PROGRAMS TO CARRY OUT ESTIMATION IN THESE MODELS AND APPLY THE TECHNIQUES DEVELOPED TO REAL DATA, ESPECIALLY IN BIOLOGY AND BASQUE LABOUR MARKET DATA. SECONDLY, IN THE AREA OF SURVIVAL ANALYSIS, WE WILL EXTEND THE METHODOLOGICAL PROPOSALS OF EMPIRICAL PLAUSIBILITY TO PROBLEMS OF INFERENCE WITH RESTRICTIONS OF ORDER, STUDYING THE ASINTOTIC DISTRIBUTIONS OF CONTRAST STATISTICS AND THE ASYMPTOTIC BEHAVIOR OF EMPIRICAL PLAUSIBILITY ESTIMATORS OBTAINED UNDER RESTRICTIONS OF ORDER, PLACING SPECIAL EMPHASIS ON THE NON-TRIVIAL COMPUTATIONAL ASPECTS OF EMPIRICAL PLAUSIBILITY METHODS. WE WILL COMPARE DISTRIBUTION FUNCTIONS AND COMPETITIVE RISKS USING THIS METHODOLOGY AND APPLY THE RESULTS IN REAL EXAMPLES IN RELIABILITY, PUBLIC HEALTH, ECONOMY AND FINANCE. FINALLY, WE WILL STUDY ALTERNATIVE PROPOSALS IN THE STATISTICAL MODELLING OF HRQOL, ESPECIALLY IN RELATION TO RESULTS PERCEIVED BY THE PATIENT, APPLYING THEM TO HEALTH AND HEALTH-RELATED QUALITY OF LIFE INDEXES. WE WILL STUDY THEIR DISTRIBUTION AND STATISTICAL MODELLING, PROPOSING IMPROVEMENTS IN THE CONSTRUCTION OF INDEXES TO MEASURE THESE RESULTS, AND WE WILL COMPARE THE DIFFERENT METHODOLOGICAL ALTERNATIVES FOR THEIR ANALYSIS. WE WILL PROPOSE ADVANCED MODELS OF SOFTENING TECHNIQUES (MAG, P-SPLINES) FOR MODELS APPLIED TO THE ANALYSIS OF RESULTS PERCEIVED BY THE PATIENT. FINALLY, WE WILL CONTEXTUALISE THE RESULTS OBTAINED, GIVE GUIDELINES FOR THE SELECTION OF SUITABLE ALTERNATIVES, AND RECOMMENDATIONS AND CONCLUSIONS BASED ON THE MOST RELEVANT ASPECTS FROM THE STATISTICAL AND CLINICAL POINT OF VIEW OF THE APPLICATION. WE WILL ALSO DEVELOP, IMPLEMENT AND VALIDATE USEFUL PREDICTIVE MODELS IN HOSPITAL CLINIC PRACTICE, IMPLEMENTING TECHNOLOGICAL TOOLS FROM THEM. TO DO THIS, WE WILL DEVELOP PREDICTIVE MODELS IN CLINICAL EVOLUTION, MORTALITY, SURVIVAL OR CHANGES IN HRQOL FOR PATIENTS WITH CHRONIC DISEASES, PROPOSING OPTIMAL TRANSFORMATIONS AND CATEGORISATIONS OF PREDICTIVE VARIABLES THAT ALLOW THEIR CORRECT INTRODUCTION INTO THE PREDICTIVE MODEL. THE DIFFERENT PREDICTIVE MODELS WILL BE EVALUATED AND COMPARED USING DIFFERENT ANALYSIS TECHNIQUES (LINEAR REGRESSION, BINARY LOGISTIC REGRESSION, ORDINAL LOGISTIC REGRESSION, POISSON REGRESSION, REGRESSION AND CLASSIFICATION TREES, NEURAL NETWORKS, AMONG OTHERS), AND DIFFERENT CRITERIA (GOODNESS OF ADJUSTMENT, CALIBRATION, PREDICTIVE CAPACITY, AMONG OTHERS). (English)
    12 October 2021
    0 references
    Bilbao
    0 references

    Identifiers

    MTM2016-74931-P
    0 references