Please use this identifier to cite or link to this item: https://www.arca.fiocruz.br/handle/icict/25189
Title: Using Gamma and Quantile Regressions to Explore the Association between Job Strain and Adiposity in the ELSA-Brasil Study: Does Gender Matter?
Authors: Fonseca, Maria de Jesus Mendes da
Juvanhol, Leidjaira Lopes
Nobre, Aline Araújo
Griep, Rosane Härter
Alves, Márcia Guimarães de Mello
Cardoso, Letícia de Oliveira
Giatti, Luana
Nunes, Maria Angélica
Aquino, Estela M. L.
Chor, Dóra
Affilliation: Fundação Oswaldo Cruz. Escola Nacional de Saúde Pública Sergio Arouca. Departamento de Epidemiologia e Métodos Quantitativos em Saúde. Rio de Janeiro, RJ. Brasil.
Universidade Federal de Viçosa. Departamento de Nutrição e Saúde. Viçosa, MG, Brasil.
Fundação Oswaldo Cruz. Instituto Oswaldo Cruz. Laboratório de Educação em Ambiente e Saúde. Rio de Janeiro, RJ. Brasil.
Fundação Oswaldo Cruz. Programa de Computação Científica. Rio de Janeiro, RJ. Brasil.
Fundação Oswaldo Cruz. Instituto Oswaldo Cruz. Laboratório de Educação em Ambiente e Saúde. Rio de Janeiro, RJ. Brasil.
Universidade Federal Fluminense. Instituto de Saúde Coletiva. Niterói, RJ, Brasil
Fundação Oswaldo Cruz. Escola Nacional de Saúde Pública Sergio Arouca. Departamento de Epidemiologia e Métodos Quantitativos em Saúde. Rio de Janeiro, RJ. Brasil.
Universidade Federal de Minas Gerais. Faculdade de Medicina. Belo Horizonte, MG, Brasil.
Universidade Federal do Rio Grande do Sul. Escola de Medicina. Programa de Pós-Graduação em Epidemiologia. Porto Alegre, RS, Brasil.
Universidade Federal da Bahia. Instituto de Saúde Coletiva. Salvador, BA, Brasil.
Fundação Oswaldo Cruz. Escola Nacional de Saúde Pública Sergio Arouca. Departamento de Epidemiologia e Métodos Quantitativos em Saúde. Rio de Janeiro, RJ. Brasil.
Abstract: This paper explores the association between job strain and adiposity, using two statistical analysis approaches and considering the role of gender. The research evaluated 11,960 active baseline participants (2008-2010) in the ELSA-Brasil study. Job strain was evaluated through a demand-control questionnaire, while body mass index (BMI) and waist circumference (WC) were evaluated in continuous form. The associations were estimated using gamma regression models with an identity link function. Quantile regression models were also estimated from the final set of co-variables established by gamma regression. The relationship that was found varied by analytical approach and gender. Among the women, no association was observed between job strain and adiposity in the fitted gamma models. In the quantile models, a pattern of increasing effects of high strain was observed at higher BMI and WC distribution quantiles. Among the men, high strain was associated with adiposity in the gamma regression models. However, when quantile regression was used, that association was found not to be homogeneous across outcome distributions. In addition, in the quantile models an association was observed between active jobs and BMI. Our results point to an association between job strain and adiposity, which follows a heterogeneous pattern. Modelling strategies can produce different results and should, accordingly, be used to complement one another.
Keywords: quantile regression models
adiposity
job strain
body mass index
waist circumference
keywords: Adiposidade
Índice de Massa Corporal
Circunferência da Cintura
modelos de regressão quantile
esforço laboral
Issue Date: 2017
Publisher: MDPI
Citation: FONSECA, Maria de Jesus Mendes da; et al. Using Gamma and Quantile Regressions to Explore the Association between Job Strain and Adiposity in the ELSA-Brasil Study: Does Gender Matter?. Int. J. Environ. Res. Public Health, v.14, 1404, 13p, 2017.
DOI: 10.3390/ijerph14111404
ISSN: 1661-7827
Copyright: open access
Appears in Collections:IOC - Artigos de Periódicos

Files in This Item:
File Description SizeFormat 
rosane_harter_etal_IOC_2017.pdf1.21 MBAdobe PDFView/Open



FacebookTwitterDeliciousLinkedInGoogle BookmarksBibTex Format mendeley Endnote DiggMySpace

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.