To our knowledge, this is the first study to examine the association between NRS and healthcare costs in Japan. Using a large longitudinal dataset of approximately 120,000 Japanese individuals over a six-year period, we analyzed both health screening and medical claims data. Our findings demonstrate that NRS is associated with significantly higher long-term healthcare costs. This association was more pronounced in men than in women.

The prevalence of NRS in countries outside of Japan ranges from 2.4% to 16.1% [18,19,20,21]. Most of these studies defined NRS as experiencing unrefreshing sleep for at least three days per week. In contrast, studies using a single yes/no item, similar to the method used in this study, have reported higher prevalence rates. For example, a U.S. study using a binary-response question reported NRS prevalence ranging from 26.3% to 42.1% [22]. In Japan, studies using the same single-item yes/no format have reported prevalence rates between 19.2% and 36.8% in the general population [7, 23], which is comparable to international findings.

However, the prevalence of NRS in the present study was notably high (45.7%). This elevated rate may be attributable to the occupational characteristics of the study population, which primarily consisted of blue-collar construction workers. Blue-collar workers are known to have shorter sleep durations than other occupational groups [24] and are more likely to report poorer sleep quality because of physically demanding labor [25]. While physical activity is generally associated with improved sleep quality [26], the effects of occupational factors, such as irregular schedules, stress, or noise, may offset these benefits. Therefore, it is plausible that the participants in this study had a higher prevalence of sleep problems, including NRS. Given the safety–critical nature of the construction industry, proactive sleep-related interventions may be particularly effective in this population.

NRS is known to increase the risk of both physical and mental health disorders, and this elevated disease risk may contribute to increased healthcare costs. NRS is associated with a higher risk of diabetes [10], hypertension [9, 10], and cardiovascular disease [27]. Several mechanisms may explain this association. First, insufficient sleep, which is a core component of NRS [28], is known to disrupt glucose metabolism. It leads to elevated blood glucose levels, partly through increased secretion of cortisol and catecholamines, and decreased leptin levels, which regulate appetite [29]. These hormonal changes promote hyperglycemia and contribute to the development of obesity and diabetes [29].

Second, NRS has been linked to systemic inflammation [30]. Inflammatory markers such as C-reactive protein, which are associated with obesity, are elevated in individuals with poor sleep [30]. Thus, inflammation may mediate the relationship between NRS and obesity-related diseases. Third, insufficient sleep and obesity are believed to share common genetic determinants [31]. Additionally, sleep apnea syndrome, a common cause of NRS, has been linked to increased risk of oral health problems [32, 33]. These conditions may further exacerbate the physical burden associated with NRS.

NRS has also been associated with an increased risk of mental health conditions [11, 12, 20]. Although the mechanisms by which NRS increases the risk of mental disorders are not fully understood, one proposed hypothesis is that individuals with NRS may experience reduced rapid eye movement (REM) sleep [6]. REM sleep plays a crucial role in emotional regulation, particularly in modulating anxiety and fear through activity in the amygdala [34]. Sleep disruption has been shown to weaken functional connectivity between the amygdala and the ventral anterior cingulate cortex, a region involved in inhibiting amygdala activity [35]. Although a direct causal relationship between NRS and reduced REM sleep has not been clearly established, NRS may represent a pathway through which emotional dysregulation and the risk of mental disorders increase.

Thus, it is conceivable that NRS contributes to increased long-term healthcare costs by increasing the risk of lifestyle-related and mental health disorders. From a policy standpoint aimed at controlling national healthcare costs, and a public health perspective focused on preserving individual health and quality of life, routine screening for NRS during medical checkups and the implementation of targeted interventions may be beneficial in reducing long-term healthcare costs. Future studies are warranted to evaluate the effectiveness and cost-efficiency of specific intervention strategies for NRS management.

This study had five primary limitations. First, the reliability and validity of the measurement of NRS may be limited. In this study, NRS was assessed using a single screening item. Although validated instruments could offer greater measurement precision, their use in large-scale occupational health checkups is impractical because of the burden on participants and the high administrative costs. Comprehensive clinical evaluations by physicians would be even more burdensome and unrealistic. Nevertheless, previous research suggests that a single-item question may be sufficient to capture the core symptoms of NRS [10]. Therefore, from the standpoints of feasibility and public health utility, the simple questionnaire used in this study may be an appropriate tool for screening NRS during workplace health examinations.

Second, the study was subject to selection bias based on participants’ occupational characteristics. As noted above, most participants were blue-collar workers, who are at a higher risk of occupational accidents, but may be at a lower risk of lifestyle-related diseases due to the amount of physical exercise than white-collar workers. Consequently, the pathways linking nonrestorative sleep to increased healthcare costs may differ from those in white-collar workers. Future studies should examine these pathways separately in blue- and white-collar populations.

Third, the dropout rate was high at 36.9%. This high dropout rate may have introduced survivorship bias, as individuals with poorer health conditions might have been more likely to leave the cohort. Additionally, this may be attributed to the withdrawal from the health insurance association due to retirement, marriage, or a change in employment, making continued follow-up impossible. Therefore, future research would benefit from the development of databases that allow longitudinal follow-up regardless of changes in health insurance coverage.

Fourth, only healthcare costs related to treatments covered by insurance benefits were used as outcomes, excluding over-the-counter (OTC) medications, in-house clinics, and self-funded medical treatments. However, because these costs are not part of public healthcare costs and account for only a small part of pharmaceutical sales in Japan, analysis from the perspective of reducing public healthcare costs may not be necessary. Nevertheless, an analysis that includes these costs is required from the perspective of individual household financing.

The fifth concern is the possibility of uncontrolled confounding factors. Sleep duration, working hours, and physical activity are reportedly associated with both sleep and healthcare costs. These variables were not examined in this study; thus, their effects could not be adjusted. Additionally, COVID-19 has been spreading in Japan since 2020, with special measures in place until May 2023, when it was reclassified into the same category as seasonal influenza. This may have significantly impacted medical consultation behavior and, consequently, healthcare costs. However, it was not considered in the analysis of this study, as it was assumed to have affected the entire target population equally. Future studies should investigate potential confounding variables at both individual and population levels and incorporate them into the model.