Bias correction in clustered underreported data
Autor(a) principal: | |
---|---|
Data de Publicação: | 2022 |
Outros Autores: | , , , , |
Tipo de documento: | Artigo |
Idioma: | eng |
Título da fonte: | Repositório Institucional da UFMG |
Texto Completo: | https://doi.org/10.1214/20-BA1244 http://hdl.handle.net/1843/56438 https://orcid.org/0000-0003-3220-6356 https://orcid.org/0000-0001-6554-9799 https://orcid.org/0000-0002-7655-6254 https://orcid.org/0000-0002-6724-9367 |
Resumo: | Data quality from poor and socially deprived regions have given rise to many statistical challenges. One of them is the underreporting of vital events leading to biased estimates for the associated risks. To deal with underreported count data, models based on compound Poisson distributions have been commonly assumed. To be identifiable, such models usually require extra and strong information about the probability of reporting the event in all areas of interest, which is not always available. We introduce a novel approach for the compound Poisson model assuming that the areas are clustered according to their data quality. We leverage these clusters to create a hierarchical structure in which the reporting probabilities decrease as we move from the best group to the worst ones. We obtain constraints for model identifiability and prove that only prior information about the reporting probability in areas experiencing the best data quality is required. Several approaches to model the uncertainty about the reporting probabilities are presented, including reference priors. Different features regarding the proposed methodology are studied through simulation. We apply our model to map the early neonatal mortality risks in Minas Gerais, a Brazilian state that presents heterogeneous characteristics and a relevant socio-economical inequality. |
id |
UFMG_110f57374ed777b7ca1bbb9ef3adee2d |
---|---|
oai_identifier_str |
oai:repositorio.ufmg.br:1843/56438 |
network_acronym_str |
UFMG |
network_name_str |
Repositório Institucional da UFMG |
repository_id_str |
|
spelling |
2023-07-17T18:51:07Z2023-07-17T18:51:07Z2022-0317195126https://doi.org/10.1214/20-BA12441931-6690http://hdl.handle.net/1843/56438https://orcid.org/0000-0003-3220-6356https://orcid.org/0000-0001-6554-9799https://orcid.org/0000-0002-7655-6254https://orcid.org/0000-0002-6724-9367Data quality from poor and socially deprived regions have given rise to many statistical challenges. One of them is the underreporting of vital events leading to biased estimates for the associated risks. To deal with underreported count data, models based on compound Poisson distributions have been commonly assumed. To be identifiable, such models usually require extra and strong information about the probability of reporting the event in all areas of interest, which is not always available. We introduce a novel approach for the compound Poisson model assuming that the areas are clustered according to their data quality. We leverage these clusters to create a hierarchical structure in which the reporting probabilities decrease as we move from the best group to the worst ones. We obtain constraints for model identifiability and prove that only prior information about the reporting probability in areas experiencing the best data quality is required. Several approaches to model the uncertainty about the reporting probabilities are presented, including reference priors. Different features regarding the proposed methodology are studied through simulation. We apply our model to map the early neonatal mortality risks in Minas Gerais, a Brazilian state that presents heterogeneous characteristics and a relevant socio-economical inequality.A qualidade dos dados de regiões pobres e socialmente carentes deu origem a muitos desafios estatísticos. Uma delas é a subnotificação de eventos vitais levando a estimativas enviesadas dos riscos associados. Para lidar com dados de contagem subnotificados, modelos baseados em distribuições compostas de Poisson têm sido comumente assumidos. Para serem identificáveis, tais modelos geralmente requerem informações extras e fortes sobre a probabilidade de relatar o evento em todas as áreas de interesse, o que nem sempre está disponível. Introduzimos uma nova abordagem para o modelo composto de Poisson assumindo que as áreas são agrupadas de acordo com a qualidade de seus dados. Aproveitamos esses clusters para criar uma estrutura hierárquica na qual as probabilidades de relatórios diminuem à medida que passamos do melhor grupo para o pior. Obtemos restrições para a identificabilidade do modelo e provamos que apenas informações prévias sobre a probabilidade de relatórios em áreas com a melhor qualidade de dados são necessárias. Várias abordagens para modelar a incerteza sobre as probabilidades de relatórios são apresentadas, incluindo prioris de referência. Diferentes características da metodologia proposta são estudadas através de simulação. Aplicamos nosso modelo para mapear os riscos de mortalidade neonatal precoce em Minas Gerais, um estado brasileiro que apresenta características heterogêneas e uma desigualdade socioeconômica relevante.CNPq - Conselho Nacional de Desenvolvimento Científico e TecnológicoFAPEMIG - Fundação de Amparo à Pesquisa do Estado de Minas GeraisCAPES - Coordenação de Aperfeiçoamento de Pessoal de Nível SuperiorOutra AgênciaengUniversidade Federal de Minas GeraisUFMGBrasilICX - DEPARTAMENTO DE CIÊNCIA DA COMPUTAÇÃOICX - DEPARTAMENTO DE ESTATÍSTICABayesian AnalysisEstatísticaDistribuição de PoissonDistribuição (Probabilidades)Mortalidade infantilCompound Poisson modelGeneralized beta distributionJeffreys priorModel identifiabilityNeonatal mortalityUnderreportingBias correction in clustered underreported dataCorreção de viés em dados subnotificados agrupadosinfo:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/articlehttps://projecteuclid.org/journals/bayesian-analysis/volume-17/issue-1/Bias-Correction-in-Clustered-Underreported-Data/10.1214/20-BA1244.fullGuilherme Lopes de OliveiraRaffaele ArgientoRosangela Helena LoschiRenato Martins AssunçãoFabrizio RuggeriMárcia D’Elia Brancoapplication/pdfinfo:eu-repo/semantics/openAccessreponame:Repositório Institucional da UFMGinstname:Universidade Federal de Minas Gerais (UFMG)instacron:UFMGLICENSELicense.txtLicense.txttext/plain; charset=utf-82042https://repositorio.ufmg.br/bitstream/1843/56438/1/License.txtfa505098d172de0bc8864fc1287ffe22MD51ORIGINALBias correction in clustered underreported data.pdfBias correction in clustered underreported data.pdfapplication/pdf4528433https://repositorio.ufmg.br/bitstream/1843/56438/2/Bias%20correction%20in%20clustered%20underreported%20data.pdf060f432411f6707e264e711ba742a888MD521843/564382023-07-17 15:51:07.661oai:repositorio.ufmg.br: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Repositório de PublicaçõesPUBhttps://repositorio.ufmg.br/oaiopendoar:2023-07-17T18:51:07Repositório Institucional da UFMG - Universidade Federal de Minas Gerais (UFMG)false |
dc.title.pt_BR.fl_str_mv |
Bias correction in clustered underreported data |
dc.title.alternative.pt_BR.fl_str_mv |
Correção de viés em dados subnotificados agrupados |
title |
Bias correction in clustered underreported data |
spellingShingle |
Bias correction in clustered underreported data Guilherme Lopes de Oliveira Compound Poisson model Generalized beta distribution Jeffreys prior Model identifiability Neonatal mortality Underreporting Estatística Distribuição de Poisson Distribuição (Probabilidades) Mortalidade infantil |
title_short |
Bias correction in clustered underreported data |
title_full |
Bias correction in clustered underreported data |
title_fullStr |
Bias correction in clustered underreported data |
title_full_unstemmed |
Bias correction in clustered underreported data |
title_sort |
Bias correction in clustered underreported data |
author |
Guilherme Lopes de Oliveira |
author_facet |
Guilherme Lopes de Oliveira Raffaele Argiento Rosangela Helena Loschi Renato Martins Assunção Fabrizio Ruggeri Márcia D’Elia Branco |
author_role |
author |
author2 |
Raffaele Argiento Rosangela Helena Loschi Renato Martins Assunção Fabrizio Ruggeri Márcia D’Elia Branco |
author2_role |
author author author author author |
dc.contributor.author.fl_str_mv |
Guilherme Lopes de Oliveira Raffaele Argiento Rosangela Helena Loschi Renato Martins Assunção Fabrizio Ruggeri Márcia D’Elia Branco |
dc.subject.por.fl_str_mv |
Compound Poisson model Generalized beta distribution Jeffreys prior Model identifiability Neonatal mortality Underreporting |
topic |
Compound Poisson model Generalized beta distribution Jeffreys prior Model identifiability Neonatal mortality Underreporting Estatística Distribuição de Poisson Distribuição (Probabilidades) Mortalidade infantil |
dc.subject.other.pt_BR.fl_str_mv |
Estatística Distribuição de Poisson Distribuição (Probabilidades) Mortalidade infantil |
description |
Data quality from poor and socially deprived regions have given rise to many statistical challenges. One of them is the underreporting of vital events leading to biased estimates for the associated risks. To deal with underreported count data, models based on compound Poisson distributions have been commonly assumed. To be identifiable, such models usually require extra and strong information about the probability of reporting the event in all areas of interest, which is not always available. We introduce a novel approach for the compound Poisson model assuming that the areas are clustered according to their data quality. We leverage these clusters to create a hierarchical structure in which the reporting probabilities decrease as we move from the best group to the worst ones. We obtain constraints for model identifiability and prove that only prior information about the reporting probability in areas experiencing the best data quality is required. Several approaches to model the uncertainty about the reporting probabilities are presented, including reference priors. Different features regarding the proposed methodology are studied through simulation. We apply our model to map the early neonatal mortality risks in Minas Gerais, a Brazilian state that presents heterogeneous characteristics and a relevant socio-economical inequality. |
publishDate |
2022 |
dc.date.issued.fl_str_mv |
2022-03 |
dc.date.accessioned.fl_str_mv |
2023-07-17T18:51:07Z |
dc.date.available.fl_str_mv |
2023-07-17T18:51:07Z |
dc.type.status.fl_str_mv |
info:eu-repo/semantics/publishedVersion |
dc.type.driver.fl_str_mv |
info:eu-repo/semantics/article |
format |
article |
status_str |
publishedVersion |
dc.identifier.uri.fl_str_mv |
http://hdl.handle.net/1843/56438 |
dc.identifier.doi.pt_BR.fl_str_mv |
https://doi.org/10.1214/20-BA1244 |
dc.identifier.issn.pt_BR.fl_str_mv |
1931-6690 |
dc.identifier.orcid.pt_BR.fl_str_mv |
https://orcid.org/0000-0003-3220-6356 https://orcid.org/0000-0001-6554-9799 https://orcid.org/0000-0002-7655-6254 https://orcid.org/0000-0002-6724-9367 |
url |
https://doi.org/10.1214/20-BA1244 http://hdl.handle.net/1843/56438 https://orcid.org/0000-0003-3220-6356 https://orcid.org/0000-0001-6554-9799 https://orcid.org/0000-0002-7655-6254 https://orcid.org/0000-0002-6724-9367 |
identifier_str_mv |
1931-6690 |
dc.language.iso.fl_str_mv |
eng |
language |
eng |
dc.relation.ispartof.pt_BR.fl_str_mv |
Bayesian Analysis |
dc.rights.driver.fl_str_mv |
info:eu-repo/semantics/openAccess |
eu_rights_str_mv |
openAccess |
dc.format.none.fl_str_mv |
application/pdf |
dc.publisher.none.fl_str_mv |
Universidade Federal de Minas Gerais |
dc.publisher.initials.fl_str_mv |
UFMG |
dc.publisher.country.fl_str_mv |
Brasil |
dc.publisher.department.fl_str_mv |
ICX - DEPARTAMENTO DE CIÊNCIA DA COMPUTAÇÃO ICX - DEPARTAMENTO DE ESTATÍSTICA |
publisher.none.fl_str_mv |
Universidade Federal de Minas Gerais |
dc.source.none.fl_str_mv |
reponame:Repositório Institucional da UFMG instname:Universidade Federal de Minas Gerais (UFMG) instacron:UFMG |
instname_str |
Universidade Federal de Minas Gerais (UFMG) |
instacron_str |
UFMG |
institution |
UFMG |
reponame_str |
Repositório Institucional da UFMG |
collection |
Repositório Institucional da UFMG |
bitstream.url.fl_str_mv |
https://repositorio.ufmg.br/bitstream/1843/56438/1/License.txt https://repositorio.ufmg.br/bitstream/1843/56438/2/Bias%20correction%20in%20clustered%20underreported%20data.pdf |
bitstream.checksum.fl_str_mv |
fa505098d172de0bc8864fc1287ffe22 060f432411f6707e264e711ba742a888 |
bitstream.checksumAlgorithm.fl_str_mv |
MD5 MD5 |
repository.name.fl_str_mv |
Repositório Institucional da UFMG - Universidade Federal de Minas Gerais (UFMG) |
repository.mail.fl_str_mv |
|
_version_ |
1803589373273309184 |