The association between work-family conflict and psychomotor vigilance among women working in aged care services in Japan

Abstract

Work-life imbalance is associated with occupational burnout and poor mental health outcomes, which may impair cognitive performance. However, few studies have investigated this association. Therefore, we examined the association between work-family conflict and cognitive performance using a 10-min psychomotor vigilance task (PVT) in a cross-sectional study of 437 women working in aged care services in Japan. Work-to-family conflict (WFC) and family-to-work conflict (FWC) scores were divided into low, moderate, and high. PVT measurements included mean reaction time (RT), slowest 10% RT, and lapses (number of responses >500 ms), categorized by median values. Higher scores indicated impaired cognitive performance. Odds ratios (ORs) and 95% confidence intervals (CIs) for impaired PVT performance were calculated using a multivariable logistic regression model. The ORs (95% CIs) for high WFC compared with low WFC regarding impaired PVT performance were 2.63 (1.26–5.45) for mean RT, 2.61 (1.25–5.46) for slowest 10% RT, and 3.21 (1.51–6.82) for lapses. However, no significant association was found for FWC. WFC was associated with impaired cognitive performance, whereas FWC was not. This is the first study that suggests managing WFC may improve cognitive performance. Further longitudinal studies are needed to examine causality and improve work environments for employees with high WFC.

Introduction

Work-life balance can be described as the balance between work- and family-related functions, including job schedules, sleep patterns, and the home environment, to meet both job and familial demands, which consequently supports one’s health and psychological well-being¹⁾. When the demands of work and/or family exceed available resources, the resulting imbalance leads to conflict between work and family roles, ultimately causing stress-related declines in mental and physical health²⁾. Such health effects, also known as occupational burnout, can result in unwanted labor outcomes in the workplace, including impaired performance, increased absenteeism, compromised safety, and reduced productivity. The United Nations’ Sustainable Development Goals (SDGs) discuss the importance of work-life balance as a component of sustainable work productivity³⁾. In Japan’s super-aging society, improving the sustainability of labor productivity is a pressing issue. According to a 2019 nationwide survey, women working in the aged care service sector have reported several challenges, including physical burden, mental stress, nighttime anxiety, and concerns about health risks, particularly those related to infection and injury⁴⁾.

Aged care workplaces in Japan are further characterized by long and irregular work shifts and chronic staff shortages, while the rapid aging of the population has steadily increased the demand for care services⁴⁾. National data show that women account for about 75% of the aged care workforce in Japan⁵⁾. In addition to workplace-related issues, familial conflicts also significantly affect individuals’ lives and their capacity to contribute to society in their professional roles^{6, 7)}. In particular, Japanese women often bear a higher burden due to the need to manage both work and household responsibilities^{5, 8, 9)}.

A theoretical framework is useful for understanding how competing demands from work and family create stress. Role-conflict theory states that stress arises when expectations from different roles are incompatible or ambiguous¹⁰⁾. Within this framework, burdens originating in the work and family domains are conceptually distinct⁶⁾. Work-family conflict is a critical metric for evaluating work-life balance, comprising work-to-family conflict (WFC) and family-to-work conflict (FWC), as reported in our previous study examining the impact of work-family conflict on insomnia among aged care workers¹¹⁾. Eby et al. further reported that WFC is primarily predicted by job-related stressors such as workload and role ambiguity, whereas FWC is more strongly driven by family demands and a lack of support⁶⁾. In addition, work and family demands can influence each other through cross-domain spillover processes, whereby strain in one domain affects functioning in the other¹²⁾.

Moreover, other studies have shown that heightened work-family conflicts are associated with increased stress, which can worsen health conditions and cognitive performance by affecting sleep and disturbing the autonomic nervous system^13,14,15,16). According to the Job Demands-Resources (JD-R) model, excessive job demands lead to work-family conflict, which includes both WFC and FWC; this conflict induces exhaustion and disengagement, core dimensions of burnout^16,17,18). Prolonged exposure to such stressors leads to an accumulation of physiological strain, as described in the Allostatic Load model; such strain is associated not only with cognitive impairment but also with an increased risk of cardiometabolic, neuropsychiatric, and immune-related disorders^{19, 20)}. Because cognitive impairment can diminish work performance and reduce sustainable productivity, its objective detection provides a sensitive marker of this broader health burden. Although the influence of work-life balance on operational efficiency has been recognized, its evaluation remains limited. However, the association of work-family conflict with objective performance measures has not been sufficiently investigated, despite its potential implications for occupational health.

The performance impairments caused by work-family conflict can be theoretically measured through vigilant attention using the psychomotor vigilance task (PVT) initially developed by Dinges et al. in 1985²¹⁾. PVT is commonly employed to evaluate the effects of sleep disturbances and fatigue on sustained attention^22,23,24). Over time, the application of PVT has expanded to various research domains, including studies on brain function, stress, performance evaluation, and drug-related assessments^{25, 26)}. As its use has grown, associated factors related to PVT measurements have been gradually revealed^{25, 27)}. Notably, PVT has been used to assess the risk of performance errors such as driving performance in vehicle drivers^{27, 28)}.

PVT is a relatively simple and validated tool that has been used to assess the impact of stressors on cognitive performance^{21, 29, 30)}, and it may serve as a useful method to evaluate the effect of heightened work-family conflict as a source of stress on caregiving workers’ performance in aged care services. Miyachi et al. reported a higher risk of PVT impairment among women working in aged care services in Japan, particularly in those with health conditions, such as a body mass index (BMI) over 22 kg/m^{2 31)}. However, to date, no research has shown that work-family conflict affects psychomotor vigilance, despite there being a theoretical association. Therefore, we aimed to investigate the association between work-family conflict and psychomotor vigilance among women working in aged care services in Japan. The hypothesized theoretical model is depicted in Fig. 1.

Fig. 1.

Hypothesized theoretical model of work-family conflict and adverse health outcomes.

WFC: work-to-family conflict; FWC: family-to-work conflict; PVT: psychomotor vigilance task; JD-R: job demands-resources.

Subjects and Methods

This cross-sectional study included 437 female workers aged 18 to 60 who were employed in aged care facilities in Japan. We restricted the sample to women because they constitute the majority of the aged care workforce and presumably experience comparatively high levels of WFC and FWC⁹⁾. Participants were recruited from four large-scale corporations operating aged care facilities within one private healthcare group in Japan. Initially, 745 participants provided written informed consent. Among them, 686 remained after excluding 17 who declined, 16 who did not monitor sleep-disordered breathing (SDB), and 26 with missing data. Then, 568 women participated in both the work-family conflict (WFC/FWC) questionnaire and PVT measurement between 2015 and 2016. Out of these, 451 completed the WFC/FWC questionnaire; however, 116 participants who answered “not applicable” were excluded. Finally, 14 participants were excluded for any of the analyzed PVT metrics exceeding ± 3 standard deviations from its mean. As a result, 437 participants were included in the final analysis. This study was approved by the Ethical Review Board of Juntendo University.

Work-family conflict

Work-family conflict was assessed using a self-administered questionnaire based on the methodology of the National Survey of Midlife Development in the United States (MIDUS)³²⁾. Respondents answered four questions for each domain (WFC and FWC), selecting one of three options: 0=never, 1=to some extent, 2=often. Conflict levels were quantified by calculating the total scores for WFC and FWC, each ranging from 0 (lower conflict) to 8 (higher conflict)³³⁾. Following the approach of previous studies in Japan, WFC and FWC scores were separately categorized into three conflict levels (low: 0–2, moderate: 3–4, and high: 5–8)³⁴⁾. This concise, eight-item scale, used to separately assess WFC and FWC, has shown acceptable internal consistency in Japanese samples (Cronbach’s α=0.79 for women)³⁵⁾.

Psychomotor vigilance task (PVT)

Psychomotor vigilance was assessed using a standard 10-min PVT, which is a well-established method for objectively measuring neurocognitive performance²⁶⁾. The task measured participants’ reaction time by having them press a button as soon as a visual stimulus appeared. These stimuli were presented at random intervals between 2 and 10 s. The assessment was conducted using a PVT-192 handheld portable device (CWE, Inc., Ardmore, PA, USA).

The test was performed during working hours between 8:30 and 17:00, aligning with the participants’ respective schedules. The start time of each PVT assessment was recorded for all participants. Each session began with a 1-min practice period, followed by the standard 10-min data collection procedure²⁷⁾. All reaction times (RTs) were recorded with the PVT device and analyzed using a dedicated software (PVTCOMMW and REACT) provided by the device’s manufacturer.

Five types of PVT metrics were calculated, each reflecting a different aspect of cognitive performance, as established in previous studies^{29, 30)}.

1. Mean RT (ms): the average reaction time across all responses, representing overall alertness and attentional capacity during the task.

2. Standard deviation (SD) RT (ms): the variability in reaction times, reflecting fluctuations in attention and potential periods of reduced focus.

3. Median RT (ms): the midpoint of all recorded reaction times, providing a measure of central tendency that is less influenced by outliers, such as lapses or extremely quick reactions.

4. Slowest 10% RT (ms): the average of the slowest 10% of responses, which is typically associated with attentional lapses.

5. Number of Lapses: the number of response times exceeding 500 ms, indicating significant lapses in attention.

For the analysis, each PVT variable was divided into lower and upper groups, with values below and above the median, respectively, because no established clinical cut-off points exist for PVT measures, and a median split provides an interpretable binary outcome for logistic regression. In this regard, we followed the approach of a previous study in which PVT outcomes, including lapses, were divided at the median³⁶⁾. Higher values of mean RT (265.6 ms), SD RT (56.5 ms), median RT (250.0 ms), and the slowest 10% RT (333.5 ms) were considered to reflect impaired cognitive performance. The median number of lapses was 1; therefore, values above the median were ≥2. Notably, previous studies have also suggested that two or more lapses are indicative of significant declines in alertness and attentional stability³⁶⁾.

Covariates

Height and body weight were measured with participants wearing light clothing and without footwear. BMI was calculated as the body weight (kg) divided by the height squared (m²). Blood pressure was measured twice on the right arm with a 5-min interval, with the average of these two readings used for analysis. Hypertension was defined as either systolic blood pressure ≥140 mmHg, diastolic blood pressure ≥90 mmHg, or currently taking antihypertensive drugs.

A self-administered questionnaire was used to collect data on several variables. Age, one-way commuting time, and sleep duration on working days were treated as continuous variables. Average working hours per day were categorized into three categories (<7 h, 7–<10 h, and 10–14 h). We also assessed the presence or absence of lifestyle and work-related factors, including current alcohol consumption, smoking habits, sleep duration on working days, shift work status, and living with children under 14 yr old, a partner or spouse, and/or parents. The questionnaire also included the Japanese versions of the Athens Insomnia Scale (AIS-J), with scores ≥6 indicating insomnia^{37, 38)}, and the 11-item version of the Centre for Epidemiologic Studies Depression Scale (CES-D), with scores ≥11 indicating depression³⁹⁾. Sleep-disordered breathing (SDB) was assessed using the respiratory disturbance index (RDI) and classified into three categories: <5 events/hour (no SDB), 5–<15 events/hour (mild SDB), and ≥ 15 events/hour (moderate-to-severe SDB)⁴⁰⁾. RDI was defined as the number of breathing disturbance events per hour during sleep and was measured at home using a single-channel airflow monitor (NGK Spark Plug, Nagoya, Japan) equipped with a PVDF film sensor, which detects nasal and oral airflow. Participants wore the device during sleep at home, and data were processed using dedicated software at a sleep research institute in Japan. This method has been validated against polysomnography⁴¹⁾. Based on previous studies^{12, 13, 23, 26, 27, 30, 31)} and our conceptual framework (Fig. 1), we selected as potential covariates variables that are theoretically related to both WFC/FWC and PVT performance.

Statistical analyses

Basic characteristics and PVT values were compared across WFC and FWC categories using Kruskal–Wallis tests for continuous variables and χ² tests for categorical variables. Odds ratios (ORs) and 95% confidence intervals (CIs) for impaired PVT performance were estimated using multivariable logistic regression analysis. Multivariable regression models were performed using the WFC and FWC scores as exposure variables. Covariates were determined in advance and entered simultaneously as potential confounders. The models were adjusted for potential confounders, including age (continuous), insomnia, depressive symptoms, SDB, average daily working hours (three categories: <7, 7–<10, and 10–14 h), sleep duration on working days (h), presence of shift work, presence of living with children under 14 yr old, presence of living with a partner/spouse, and the time of PVT assessment. All p-values represent overall group differences, unless noted. All statistical analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA). Two-tailed p-values <0.05 were considered statistically significant.

Results

Table 1 presents the basic characteristics of PVT values in the WFC and FWC levels. All types of the reaction times showed significant differences among WFC groups. A significant difference was observed in the proportion of participants with two or more lapses across the WFC groups. Observationally, the proportion was lowest in the low WFC group and appeared to increase in the moderate and high groups, with the proportion in the high WFC group being approximately double that in the low WFC group. However, no significant differences in PVT values were found between the FWC levels.

Table 1. Basic characteristics of PVT metrics by levels of work-family conflict (WFC/FWC).

	Work-to-Family Conflict (WFC)				Family-to-Work Conflict (FWC)
	Low (n=252)	Moderate (n=131)	High (n=54)	p-value	Low (n=280)	Moderate (n=95)	High (n=62)	p-value
Mean RT, ms (IQR)	263 (231–269)	266 (231–269)	274 (261–299)	0.02	265 (232–271)	267 (230–270)	265 (233–268)	0.94
SD RT, ms (IQR)	55 (37–77)	56 (38–74)	61 (49–78)	0.02	56 (46–69)	59 (46–72)	59 (46–72)	0.47
Median RT, ms (IQR)	248 (231–269)	250 (231–269)	261 (243–281)	0.03	249 (232–271)	251 (230–270)	253 (233–268)	0.87
Slowest 10% RT, ms (IQR)	328 (300–357)	333 (302–361)	346 (328–385)	0.01	332 (303–358)	336 (301–369)	340 (301–378)	0.75
Number of Lapses, n (IQR)	1 (0–2)	1 (0–2)	1 (0–3)	0.07	1 (0–2)	1 (0–2)	1 (0–2)	0.70
Number of Lapses ≥2 times, %	19.4	26.7	38.9	0.01	23.2	25.3	25.8	0.87

Continuous data are expressed as median (interquartile range, IQR), and categorical data as percentages (%). Differences across WFC and FWC levels were assessed using the Kruskal–Wallis test for continuous PVT variables and the chi-square test for categorical variables. PVT: Psychomotor Vigilance Task; RT: Reaction Time; ms: milliseconds; WFC: Work-to-Family Conflict; FWC: Family-to-Work Conflict; p<0.05 indicates statistical significance.

Table 2 shows the basic characteristics of participants. The number of study participants with insomnia and depressive symptoms was significantly higher in the high WFC group than in the moderate and low WFC groups. Similar trends were observed in the FWC groups, with significantly higher rates of insomnia and depressive symptoms in the high FWC group than in the moderate and low FWC groups. Working hours also varied significantly across WFC levels. Participants in the high WFC group were more likely to work longer hours in the range of 10 to 14 h per day compared to the moderate and low WFC groups (p<0.001). Participants in the moderate FWC group were oldest, and overall age differed significantly across FWC levels. However, the proportion of shift workers was higher in the low FWC group compared to the moderate and high FWC groups. Regarding household composition, participants living with children aged under 14 tended to report higher WFC and FWC. Living with a partner or spouse was most common in the moderate-FWC group, and WFC differed only slightly across groups. A statistically significant difference in sleep duration was observed among WFC groups, although the overall variation was minimal.

Table 2. Basic characteristics by levels of work-family conflict (WFC/FWC).

	Work-to-Family Conflict (WFC)				Family-to-Work Conflict (FWC)
	Low (n=252)	Moderate (n=131)	High (n=54)	p-value	Low (n=280)	Moderate (n=95)	High (n=62)	p-value
Median FWC score (IQR)	1 (0–2)	3 (0–4)	3.5 (2–6)	-	0 (0–1)	4 (3–4)	6 (5–6)	-
Median WFC score (IQR)	1 (0–2)	3 (3–4)	5 (5–6)	-	2 (1–3)	3 (2–4)	3.5 (3–5)	-
Median age, yr (IQR)	39 (28–49)	37 (29–47)	40 (30–48)	0.55	38 (26–47)	43 (35–51)	38 (31–47)	0.0007
Median BMI kg/m2 (IQR)	21.9 (20.3–24.2)	21.7 (20.2–25.1)	22.0 (20.2–24.4)	0.99	21.8 (20.2–24.4)	22.2 (20.5–24.9)	21.9 (20.0–24.2)	0.67
Working hours				<0.001				0.68
<7 h, %	13.1	8.40	5.56		9.64	10.5	16.1
7 to <10 h, %	82.1	79.4	66.7		80.7	78.9	74.2
10 to 14 h, %	4.76	12.2	27.8		9.64	10.5	9.68
Median one-way commuting time, min (IQR)	15 (10–30)	20 (10–30)	20 (15–30)	0.08	15 (10–25)	20 (10–30)	20 (15–30)	0.02
Median sleep duration on workdays, h (IQR)	6 (6–7)	6 (5–7)	6 (5–7)	0.01	6 (6–7)	6 (5–7)	6 (5–7)	0.17
Insomnia, %	19.4	37.4	57.4	<0.001	23.2	35.8	48.4	<0.001
Depressive symptoms, %	9.92	26.7	48.1	<0.001	14.3	25.3	35.5	<0.001
Sleep-Disordered Breathing (SDB)				0.68				0.38
No SDB (RDI<5/h), %	37.7	38.2	48.1		40.3	40.0	35.8
Mild SDB (5 ≤ RDI<15/h), %	51.2	51.1	44.4		53.2	50.4	48.4
Moderate-severe SDB (RDI ≥15/h), %	11.1	10.7	7.41		6.45	9.64	15.8
Hypertension, %	4.76	6.11	0.00	0.19	4.29	7.37	1.61	0.22
Alcohol intake, %	35.7	41.2	29.6	0.30	35.0	38.9	40.3	0.64
Current smoker, %	20.6	20.6	14.8	0.53	22.1	12.6	21.0	0.31
Shift work, %	30.6	30.6	20.4	0.31	35.0	20.0	17.7	0.002
Living with children under 14 yr old, %	29.8	39.7	50.0	0.01	23.9	46.3	69.4	<0.0001
Living with partner/spouse, %	53.2	58.8	68.5	0.10	43.9	81.1	77.4	< 0.0001
Living with parents, %	50.0	44.3	53.7	0.42	52.5	45.3	37.1	0.07
Median time of day for PVT measurement, hh:mm (IQR)	13:02 (10:06–14:15)	12:02 (10:03–14:12)	12:17 (9:56–14:22)	0.91	12:14 (10:05–14:13)	12:11 (10:00–14:54)	11:40 (10:15–14:26)	0.97

Continuous variables are presented as median (interquartile range, IQR). Categorical variables are presented as percentages (%) only. Between-group differences across WFC and FWC levels were assessed using the Kruskal–Wallis test for continuous variables and the chi-square test for categorical variables. WFC: Work-to-Family Conflict; FWC: Family-to-Work Conflict; BMI: Body Mass Index; RDI: respiratory disturbance index; RDI frequency expressed as events/hour; p<0.05 indicates statistical significance.

Multivariable logistic regression analysis (Table 3) revealed that high WFC was significantly associated with poorer PVT performance across multiple metrics. Specifically, participants in the high WFC group had increased odds of exhibiting higher mean RT (OR=2.63, 95% CI: 1.26–5.45), higher SD RT (OR=2.08, 95% CI: 1.02–4.24), higher median RT (OR=2.16, 95% CI: 1.06–4.40), higher slowest 10% RT (OR=2.61, 95% CI: 1.25–5.46), and higher number of lapses (OR=3.21, 95% CI: 1.51–6.82). Additionally, a significant trend was observed across increasing levels of WFC for mean RT, median RT, slowest 10% RT, and the number of lapses (p for trend=0.02, 0.046, 0.02, and 0.002, respectively). By contrast, no significant associations were identified between FWC levels and any of the PVT performance parameters (Table 4).

Table 3. Multivariable-adjusted logistic regression results for odds ratios of PVT variables by WFC levels.

	WFC level	OR (95% CI)	p for trend
High Mean RT category	Low WFC	Reference	0.02
	Moderate WFC	1.05 (0.66–1.66)
	High WFC	2.63 (1.26–5.45)
High SD RT category	Low WFC	Reference	0.06
	Moderate WFC	1.16 (0.73–1.85)
	High WFC	2.08 (1.02–4.24)
High Median RT category	Low WFC	Reference	0.046
	Moderate WFC	1.16 (0.73–1.84)
	High WFC	2.16 (1.06–4.40)
High Slowest 10% RT category	Low WFC	Reference	0.02
	Moderate WFC	1.11 (0.69–1.78)
	High WFC	2.61 (1.25–5.46)
High Number of Lapses category	Low WFC	Reference	0.002
	Moderate WFC	1.66 (0.97–2.85)
	High WFC	3.21 (1.51–6.82)

The PVT variables represent the higher half of values, divided at the median. Results are expressed as odds ratios (OR) with 95% confidence intervals (CI) obtained from multivariable logistic regression analysis. PVT: Psychomotor Vigilance Task; RT: Reaction Time; WFC: Work-to-Family Conflict. p<0.05 indicates statistical significance. Adjusted for confounding factors: age, insomnia, depressive symptoms, SDB (three categories of RDI), average daily working hours (three categories), sleep duration on working days, presence of shift work, presence of living with children under 14 yr old, presence of living with partner/spouse, and time of PVT assessment.

Table 4. Multivariable-adjusted logistic regression results for odds ratios of PVT variables by FWC levels.

	FWC level	OR (95% CI)	p for trend
High Mean RT category	Low FWC	Reference	0.64
	Moderate FWC	0.95 (0.56–1.63)
	High FWC	0.84 (0.43–1.65)
High SD RT category	Low FWC	Reference	0.44
	Moderate FWC	1.07 (0.63–1.83)
	High FWC	0.71 (0.36–1.37)
High Median RT category	Low FWC	Reference	0.77
	Moderate FWC	0.93 (0.55–1.58)
	High FWC	0.92 (0.48–1.78)
High Slowest 10% RT category	Low FWC	Reference	0.77
	Moderate FWC	0.90 (0.52–1.54)
	High FWC	0.94 (0.48–1.84)
High Number of Lapses category	Low FWC	Reference	0.51
	Moderate FWC	0.87 (0.47–1.60)
	High FWC	0.79 (0.37–1.67)

The PVT variables represent the higher half of values, divided at the median. Results are expressed as odds ratios (OR) with 95% confidence intervals (CI) obtained from multivariable logistic regression analysis. PVT: Psychomotor Vigilance Task; RT: Reaction Time; FWC: Family-to-Work Conflict. p<0.05 indicates statistical significance. Adjusted for confounding factors: age, insomnia, depressive symptoms, SDB (three categories of RDI), average daily working hours (three categories), sleep duration on working days, presence of shift work, presence of living with children under 14 yr old, presence of living with partner/spouse, and time of PVT assessment.

Discussion

We found that high WFC was significantly associated with impaired PVT performance among women working in aged care services in Japan. By contrast, FWC was not associated with PVT performance. This is a noteworthy finding, as this study is the first to evaluate the relationship between WFC, FWC, and cognitive performance using the PVT.

To our knowledge, no previous study has directly examined the association between work-family conflict and PVT performance. However, earlier studies have reported that WFC is strongly associated with chronic stress^42,43,44). WFC is primarily driven by psychosocial and medical factors related to work, not only contributing to strain but also interacting with other stressors over time⁴⁵⁾. Longitudinal analyses have shown that the relationship between WFC and strain persists in the long term, supporting the view that WFC contributes to chronic stress⁴³⁾. This chronic stress is associated with impaired cognitive performance in PVT^{17, 18, 46,47,48)}. In particular, PVT is commonly used to assess sustained attention and alertness, particularly under stress conditions, such as sleep deprivation and fatigue^{22, 23, 27)}. Another study demonstrated that induced stress is associated with impaired cognitive performance, although they did not directly investigate WFC⁴⁶⁾. Given these findings, WFC may be associated with impaired cognitive performance through stress. In the present study, WFC was found to be significantly associated with impaired PVT performance, even after adjusting for confounders, such as depressive symptoms, insomnia, and environmental stressors (i.e. length of working hours and presence of living with someone). These results suggest that WFC independently contributes to cognitive impairment.

The decline in PVT performance observed in our study may be plausibly explained by chronic stress induced by high levels of WFC, which can disrupt autonomic nervous system regulation and alter physiological states⁴⁹⁾. These physiological changes could induce reductions in cerebral blood flow and metabolism that impair brain function and contribute to impaired vigilance^{50, 51)}. Previous studies have shown that declines in PVT performance, particularly lapses and mean RT, are associated with sleep deprivation and stress, both of which are linked to decreased brain metabolism^{28, 29, 52, 53)}.

In a previous Japanese study, individuals under high levels of chronic stress exhibited reduced blood flow in brain regions, such as the ventrolateral prefrontal cortex (VLPFC), caudate nucleus, and insula. The reduced activity in the VLPFC and midbrain is thought to be associated with diminished cognitive performance⁵⁴⁾. Studies have suggested that stress could cause physiological changes, such as heightened autonomic nervous system activity, elevated cortisol levels, reduced dopamine and serotonin, changes in immune and inflammatory responses, and decreased neural plasticity, which can impair brain function^{17, 55,56,57)}. Whereas such physiological changes could also result from aging, depression, and sleep deprivation, our study suggests that high WFC may independently alter brain function, leading to reduced cognitive performance as measured by the PVT.

As mentioned in our main findings, FWC was not associated with PVT performance, which may be due to a comparatively lower stress response than that elicited by WFC. This difference between FWC and WFC may be attributed to socio-psychological factors related to differences in the nature of family and work environments⁴³⁾. Notably, previous studies from Japan and other countries reported that increased WFC has a stronger association with feelings of guilt than FWC among working mothers, which in turn relates to higher stress levels^{58, 59)}. By contrast, the home environment often provides emotional support and opportunities for mental recovery, which may buffer the impact of family demands on stress and cognitive function⁷⁾. In our study, the higher rates of living with children, and partners among women with high FWC may indicate greater opportunities for such family-based support that may mitigate the effects of conflict, although co-residence may simultaneously increase caregiving demands and stress.

Cultural norms may also play a role. In many societies including Japan, particularly among women, prioritizing family over work is often considered acceptable or even expected^58,59,60). This may reduce feelings of guilt when family responsibilities interfere with work. Additionally, a study among nursing home assistants found that demanding work schedules have been associated with depression and somatic symptoms, which may reflect chronic stress and physiological state changes⁶¹⁾. These factors could help explain why high FWC was not associated with impaired PVT performance in the present study. In fact, Sato et al. noted that insomnia is more prevalent in individuals with high WFC compared to those with high FWC¹¹⁾. Furthermore, a study reported that WFC has a more persistent, long-term impact on stress accumulation than FWC⁴³⁾. It is also possible that FWC-related stress is more effectively managed within family contexts through psychological coping or emotional buffering⁶²⁾, whereas WFC reflects more persistent stress. Therefore, improving working conditions might help improve cognitive performance and support sustainable work-life balance.

There are several limitations to this study. As a cross-sectional study, we investigated a single point of observation, which limited the ability to observe changes in impaired PVT performance and the association between high WFC over time. However, causality of those factors could not be ascertained from the cross-sectional study frame and it needs to be investigated with longitudinal study settings. In addition, the generalizability of the findings is limited. This study solely focused on female participants working in aged care services within one private healthcare group in Japan. Therefore, the results may not be applicable to men or to populations with different occupations or cultural backgrounds. Moreover, non-random recruitment from facilities within this single corporate group, which constituted a non-random sample, may have introduced selection bias. In addition, the response variability of the work–family conflict scale used in the present study may limit its sensitivity to detect individual differences and may underestimate spillover effects when compared with more psychometrically discriminative multidimensional measures⁶³⁾. Aged care workplaces in Japan are also characterized by long and irregular shifts and persistent staffing shortages, factors known to heighten occupational stress and potentially intensify work-family conflict⁴⁾. The differences between populations in work-related pressures should be considered when our findings are interpreted. Future studies should be conducted to investigate these associations in more diverse groups to confirm their generalizability. Since the measurement of PVT performance was conducted at various times during working hours between 8:30 and 17:00, fluctuations in circadian rhythm might have affected our results³¹⁾. Although the time of day when the PVT was administered was included as an adjustment factor in the multivariable analyses, residual effects of circadian variation^{24, 29)} and other unmeasured confounders such as domestic responsibilities (i.e., childcare, elder care, household labor) may still be present. Therefore, further studies with more comprehensive assessments are needed to clarify the association between WFC and PVT performance.

Conclusion

Our study found that WFC was inversely associated with PVT measures among women working in aged care services in Japan, whereas FWC was not. To our knowledge, this is the first study to demonstrate this association. These findings may contribute to further developments in this area and support efforts to create better work environments. Further longitudinal studies could examine the causal relationship and explore whether introducing flexible working arrangements, such as workplace support systems to reduce WFC, telecommuting, or adjustable working hours, could help improve work-life balance, reduce stress levels, and enhance cognitive performance among aged care workers with high WFC.

Ethics Approval

This study was approved by the Ethical Review Board of the Juntendo University Faculty of Medicine in 2014 (authorization number: 2014057). All procedures performed in studies involving human participants were conducted in accordance with ethical standards and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards.

Informed Consent

Written informed consent was obtained from all individual participants included in this study.

Author Contributions

Study concept and design: TU, AI, and TT. Acquisition of data: AI, YS, KM, NM, HW, and TT. Statistical analysis and drafting of the manuscript: TU, AI, HC and SS. Study organizational support for implementation: YK, AI and TT. Analysis, interpretation of data, and critical revision of the manuscript for important intellectual content: all authors

Funding

This study was supported in part by a Grant-in-Aid for Scientific Research (A) (Grant/Award Number: 22H00496), as well as by Grants-in-Aid for Scientific Research (B) (Nos. 18390194, 22390134, and 25293142) from the Japan Society for the Promotion of Science and the Ministry of Education, Culture, Sports, Science and Technology of Japan.

Conflict of Interest

Takeshi Tanigawa is a part-time occupational physician of Koyama Healthcare Group. All other authors declare no competing interests.

Funding Statement

This study was supported in part by a Grant-in-Aid for Scientific Research (A) (Grant/Award Number: 22H00496), as well as by Grants-in-Aid for Scientific Research (B) (Nos. 18390194, 22390134, and 25293142) from the Japan Society for the Promotion of Science and the Ministry of Education, Culture, Sports, Science and Technology of Japan.

Data Availability

The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.