Banca de QUALIFICAÇÃO: MARCOS AURÉLIO MAIA DA SILVA

Uma banca de QUALIFICAÇÃO de MESTRADO foi cadastrada pelo programa.
STUDENT : MARCOS AURÉLIO MAIA DA SILVA
DATE: 29/05/2026
TIME: 14:00
LOCAL: Ambiente virtual
TITLE:

Spatial model for age-specific mortality and life expectancy in small areas with
defective vital records

KEY WORDS:

Mortality; Bayesian inference; hierarchical model

PAGES: 50
BIG AREA: Ciências Exatas e da Terra
AREA: Probabilidade e Estatística
SUMMARY:

This work presents a methodological extension of the model proposed in
Bayesian Estimation of Age-Specific Mortality and Life Expectancy for Small Areas
With Defective Vital Records, incorporating a spatial structure aimed at improving the
estimation of age-specific mortality rates and life expectancy at birth in small areas. The
proposal is based on the development of a hierarchical Bayesian model capable of
simultaneously estimating mortality risk and correcting for underreporting of deaths in
contexts characterized by deficiencies in vital records.
It is assumed that the number of deaths follows a Poisson distribution, whose
mean is defined by the exposed population, the mortality risk, and the death coverage
rate. The modeling adopts distinct prior distributions for these components, with the
mortality pattern structured based on the TOPALS model, allowing greater flexibility in
smoothing age-specific curves.
Spatial dependence is introduced in the log-mortality risks, modeled through
latent factors with spatial effects, assuming a multivariate normal distribution with a
covariance structure dependent on the geographic proximity between areas and age
groups. Additionally, underreporting of deaths is addressed through a monotonic
transformation of the proportion of observed records.
The hierarchical specification of the model allows capturing different intensities
of spatial association across regions, making it sensitive to contextual heterogeneities,
such as differences between urban and rural areas. Thus, the proposed approach
contributes to obtaining more accurate estimates of mortality levels and patterns,
providing a relevant tool to support evidence-based public policy formulation,
especially in contexts with poor-quality vital data.

COMMITTEE MEMBERS:
Interno - 3010614 - ELIARDO GUIMARAES DA COSTA
Presidente - 1781198 - FIDEL ERNESTO CASTRO MORALES
Externo ao Programa - 1346605 - FLAVIO HENRIQUE MIRANDA DE ARAUJO FREIRE - nullExterno ao Programa - 2002253 - MARCOS ROBERTO GONZAGA - nullInterno - 1342138 - RODRIGO MATHEUS ROCHA DE MEDEIROS