Objectives To operationalize a definition of long working hours that overcomes limitations of existing heterogeneous definitions and to examine the temporal trends of long working hours across worker characteristics in the US.
Method We examined 25 years (1985–2010) of repeated working hour measures from a representative sample of workers in the US Panel Study of Income Dynamics. Self-reported working hours included total annual hours worked, total annual overtime worked, jobs worked, and weekly hours worked by job. An exposure assessment model was produced through the creation of a directed acyclic graph, and a corresponding multivariate model was constructed for purposes of examining long working hours as an independent risk factor for various health outcomes, including cardiovascular disease.
Results An improved measure of working hours was produced in the form of a model that included dimensions of working hour intensity and duration. Descriptive analyses evaluating the frequency and temporal trends of long working hours across demographic and occupational strata were calculated on 31 136 participants employed during this study period, with 66.8% and 32.0% who worked more than 40 and 50 h per week on average, respectively, for any year.
Conclusions The longitudinal nature of this study in a large representative sample of US workers using repeated measures of working hours allowed us to operationalize a more comprehensive definition of long working hours, addressing methodological issues identified in previous research and providing enhanced generalizability. We examined the relationship of long working hours with health outcomes while considering participants demographic and occupational characteristics.