Challenges in Cancer Incidence: Multivariate Models for Completing Registries and Forecasting

G. Retegui Goñi, J. Etxeberria Andueza, M. D. Ugarte Martínez

Cancer incidence and mortality data are crucial for understanding cancer burden, setting control targets, and evaluating policies. Most countries record cancer mortality through Statistical Offices, providing data at various levels (national, provincial, municipal or even census tracts). Cancer incidence, however, are recorded by national or regional Population-Based Cancer Registries (PBCRs), but these figures are usually available with a delay. Furthermore, in large countries, regional PBCRs are often established in different years, resulting in incomplete and non-harmonized data series within a country, with lack of information mainly at the beginning of the data series. This work focuses on deriving short‐term cancer incidence predictions in the presence of missing data using flexible shared spatio-temporal models. The new models, implemented in INLA, will be assessed using cancer data from England.

Palabras clave: Bayesian inference, disease mapping, predictions, shared component models.

Programado

Métodos bayesianos aplicados a la medicina
10 de junio de 2025  17:10
Sala de prensa (MR 13)

Evaluación de SEM y Redes Bayesianas en la Inferencia Causal de Modelos de Cuidado Centrado en la Persona

E. Benitez, M. Olano Lizarraga, J. Martín Martín

Selección Bayesiana de Variables en Riesgos Competitivos.

A. de la Cruz Herrán, A. Forte Deltell, J. A. Carbonell Asins