Using gamma and quantile regressions to explore the association between job strain and adiposity in the elsa-brasil study: does gender matter?
Carregando...
Data
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Federal de Minas Gerais
Descrição
Tipo
Artigo de periódico
Título alternativo
Primeiro orientador
Membros da banca
Resumo
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.
Abstract
Assunto
Adiposity, Body Mass Index, Waist Circumference
Palavras-chave
Quantile regression models, Adiposity, Job strain, Body Mass Index, Waist Circumference
Citação
Departamento
Curso
Endereço externo
https://pubmed.ncbi.nlm.nih.gov/29149021/