Sun Yijia, Guardiano Megan, Saiki Mayumi, Li Jian
Department of Biostatistics, Fielding School of Public Health, University of California Los Angeles, Los Angeles, California, USA.
School of Nursing, University of California Los Angeles, Los Angeles, California, USA.
Am J Ind Med. 2025 Mar;68(3):264-272. doi: 10.1002/ajim.23686. Epub 2025 Jan 8.
Sleep disturbances are a major public health concern in the United States, leading to adverse health outcomes. In the working population, job strain has been identified as an important risk factor for sleep disturbances, but evidence from the United States remained limited. This study aimed to examine longitudinal associations between job strain and sleep disturbances in the United States, with a focus on the alternative formulations of job strain.
A total of 1721 participants were drawn from two waves of the Midlife in the United States (MIDUS) study, with an average 9-year follow-up period. Job strain was measured using Karasek's Job-Demand-Control model and operationalized in six formulations: standard quadrant, simplified quadrant, linear, quotient, logarithm quotient, and quartile based on quotient. Generalized Estimating Equations were used to estimate longitudinal associations of alternative formulations of job strain at baseline with sleep disturbances across follow-up. Corrected Quasi-likelihood Information Criterion (QICu) was used to assess the goodness of fit.
All approaches showed that higher job strain at baseline was significantly associated with an increase in sleep disturbances across follow-up. QICu scores indicated that continuous Demand-Control formulations (linear, quotient, logarithm quotient) had better model performance of 4602.66, 4604.28, and 4601.99, respectively. The logarithm quotient showed the best fit.
Our findings imply the importance of early workplace interventions in reducing job strain to improve sleep hygiene. They further show that the continuous formulations quantifying job strain were more consistent and robust, which provides suggestions for future workplace health research in the United States.
睡眠障碍是美国主要的公共卫生问题，会导致不良健康后果。在工作人群中，工作压力已被确定为睡眠障碍的一个重要风险因素，但来自美国的证据仍然有限。本研究旨在探讨美国工作压力与睡眠障碍之间的纵向关联，重点关注工作压力的替代形式。
共有1721名参与者来自美国中年（MIDUS）研究的两波调查，平均随访期为9年。工作压力采用Karasek的工作需求-控制模型进行测量，并以六种形式实施：标准象限、简化象限、线性、商数、对数商数和基于商数的四分位数。使用广义估计方程来估计基线时工作压力的替代形式与随访期间睡眠障碍之间的纵向关联。使用校正的拟似然信息准则（QICu）来评估拟合优度。
所有方法均显示，基线时较高的工作压力与随访期间睡眠障碍的增加显著相关。QICu分数表明，连续的需求-控制形式（线性、商数、对数商数）具有更好的模型性能，分别为4602.66、4604.28和4601.99。对数商数显示拟合最佳。
我们的研究结果表明早期职场干预对于减轻工作压力以改善睡眠卫生的重要性。它们还进一步表明，量化工作压力的连续形式更一致且稳健，这为美国未来的职场健康研究提供了建议。
