Gender disparities in job flexibility, job security, psychological distress, work absenteeism, and work presenteeism among U.S. adults

Abstract

Background

While international research has examined the relationship between job characteristics and mental health, including gender differences, few studies have assessed these associations at the national level in the U.S., which has unique labor markets, health care systems, and societal structures that may exacerbate gender disparities. This study investigates gender differences in the associations between job flexibility, job security, mental health outcomes, work absenteeism, work presenteeism, and mental health care utilization among a representative sample of working U.S. adults.

Methods

We analyzed cross-sectional population-based survey data from employed adults in the 2021 National Health Interview Survey. Job characteristics included perceived job flexibility and security. Outcomes included serious psychological distress, frequency of anxiety, work absenteeism, work presenteeism, and mental health care utilization. Multivariable logistic and binomial regression analyses examined associations of interest, with statistical interaction tests conducted to assess gender differences.

Findings

The study sample included 18,112 respondents weighted to represent a population of 168,068,586 civilian, non-institutionalized working U.S. adults (47.7% female). Females with low job security had a decreased probability of serious psychological distress than males with low job security (F_(3,589) = 2.79; p = 0.040). Females with the lowest job flexibility reported more days worked while ill than males over the past 3 months, while males with higher job flexibility reported more days worked while ill than females (F_(3,589) = 4.1; p = 0.007). The average number of work days missed over 12 months was lower among females than males when job security was perceived as fairly low and higher among females than males as job security increased (F_(3,589) = 4.3; p = 0.005).

Interpretation

Findings highlight the need for policies and practices that recognize and address gender-specific workforce experiences and needs. Tailored interventions that enhance job flexibility and security, support caregiving responsibilities, and provide comprehensive mental health services can address such challenges.

Highlights

• •Women with low job security experienced less psychological distress than men.
• •Women with low job flexibility worked more days while ill than men.
• •Policies that address gender-specific needs in the workplace can enhance employee well-being.

1. Introduction

Mental health is a growing concern in the United States (U.S.), with women experiencing higher rates of poor mental health outcomes and lower rates of mental health care treatment than men (Azharuddin et al., 2023). Women are nearly twice as likely to be diagnosed with an anxiety disorder in their lifetime (23.4% vs. 14.3%) (Remes et al., 2016), experience elevated risks of depression, and have more than double the odds of having psychiatric disorders compared to men ((NCS), 2017). Despite experiencing a greater burden of mental health issues, a 2022 national survey revealed that only half of women who needed mental health appointments were able to schedule them, while 40% did not seek care at all (Diep et al., 2022).

A strong body of research, grounded in theoretical frameworks, such as the Job Demands-Resources model (Demerouti et al., 2001), provides a versatile framework for understanding how both job demands and available resources influence health across populations. Job characteristics include flexibility (the degree to which employees have control over when, where, and how they work) and security (perceived stability of one's employment) and are established determinants of mental and physical health (Bakker & Demerouti, 2017; Van den Broeck & Parker, 2017). High job demands, such as heavy workload and physical and emotional strain, can elevate the risk of stress, burnout, and poor mental health (Han, 2024). Conversely, job resources such as job flexibility, autonomy, and security can serve as protective buffers, mitigating stress and enhancing well-being (Hammig & Vetsch, 2021).

Results from several studies align with the Job Demands-Resources model, demonstrating that job resources and demands are associated with mental health. A systematic review and meta-analysis of prospective studies found job insecurity to be a significant predictor of psychological distress, anxiety, and depression (Kim & von dem Knesebeck, 2016). Interventions aimed at increasing job flexibility have been shown to yield modest improvements in mental health across randomized and non-randomized trials, cross-over studies, and prospective cohort studies (Shiri et al., 2022). Our previous cross-sectional analysis of national data reported that having greater job flexibility and security was associated with lower psychological distress and anxiety among working U.S. adults, with greater job flexibility also linked to decreased work absenteeism (Wang et al., 2024).

However, employees in the U.S. and across the globe experience distinct job characteristics, experiences, and stressors based on gender. Working women consistently report lower job flexibility and higher caregiving responsibilities at home than working men (Yucel & Fan, 2023). Job insecurity, an established risk factor for adverse mental health (Salas-Nicas et al., 2024), is disproportionately experienced by women, irrespective of educational attainment (Menendez-Espina et al., 2020). Additionally, women, particularly women of color, face greater exposure to workplace discrimination and microaggressions (Field et al., 2023), which collectively exacerbate mental health risks. The COVID-19 pandemic further worsened pre-existing gender disparities at work and at home, particularly among parents and caregivers (Lyttelton et al., 2022b).

Gender differences in the association between job flexibility, job security, and mental health have been previously reported in studies with European and Australian samples, with research cited above highlighting how gender inequities in the workplace and caregiving responsibilities may shape these disparities. Job characteristics such as flexibility and security may act as resources that buffer against stress, as posited in the Job Demands-Resources model. However, these resources are not equally accessible or effective for men and women due to structural gender inequities and societal expectations. For example, the disproportionate domestic and caregiving burdens, lower pay, reduced job flexibility and security, and limited opportunities for career advancement that women face (Herr et al., 2020; Salas-Nicas et al., 2024; Yucel & Fan, 2023) may create greater work demands and fewer resources for this population, potentially diminishing the protective health effects of job resources. A cross-sectional retrospective study of employees in England found that lower job flexibility was associated with poorer mental health, with women more affected than men (Moss et al., 2022). An Australian panel study (2002–2015) reported that improvements in job security were significantly associated with reduced depression and anxiety y symptoms, particularly for men (LaMontagne et al., 2021). Long-term panel data from the UK (2010–2021) indicated that male employees and those from higher occupational classes gained significantly more mental health benefits from work autonomy, including increased job flexibility, while female employees in lower occupational classes did not experience the same benefits (Lu et al., 2023).

While international research has explored the relationships between job characteristics, mental health, and gender, there is limited research focusing on a national U.S. sample. The U.S. labor market is characterized by the absence of universal paid parental and sick leave (de Souza et al., 2022; Eggleston et al., 2023; Wething, 2022) and fewer worker protections than other developed countries (Henderson, 2023). Additionally, the high costs of childcare and health care (Buchholz, 2024; Organisation for Economic Co-operation and Development, 2022; Wager et al., 2024), lack of universal health care and health care access (Crowley et al., 2020), and fewer comprehensive social safety net programs in the U.S. (Aizer et al., 2022) present unique challenges that may amplify gender differences in how job demands and resources shape well-being. Application of the Job Resources-Demand model in this context requires deeper consideration of gender-specific factors that shape how these demands and resources are differentially experienced by men and women, particularly in societies where caregiving burdens and access to support systems are inadequate and unequally distributed (Dubbelt et al., 2016). This study also fills a critical gap by examining data from the first year of the COVID-19 pandemic, providing an opportunity to examine the model's relevance during a time when existing gender disparities worsened across economic, social, and health indicators (Flor et al., 2022; Lyttelton et al., 2022a, Lyttelton et al., 2022b).

Using nationally representative data from the 2021 U.S. National Health Interview Survey (NHIS), we analyze gender differences in the associations between job flexibility, job security, and mental health outcomes. Drawing on the Job Demands-Resources Model (Demerouti et al., 2001) and prior research, we conceptualize gender as a moderator of the associations between job characteristics and mental health outcomes, anticipating that women's increased caregiving burdens and experiences of structural work inequities in the U.S. and in a pandemic environment may diminish the protective effects of job flexibility and security. Specifically, we examined the associations between job flexibility and job security with serious psychological distress, frequency of anxiety, work absenteeism and presenteeism (working while ill), and mental health care utilization. We previously demonstrated that job flexibility and security were negatively associated with psychological distress, anxiety, and work absenteeism in this nationally representative sample of working U.S. adults (Wang et al., 2024). Note, we use the term “gender” in the introduction and discussion to capture the social, cultural, environmental, and behavioral factors that influence gender identity and health. In contrast, we use the term “sex” in the methods and results to describe the measurement of biological sex in the data (Clayton & Tannenbaum, 2016).

For the present study, we hypothesize that negative associations between exposures (job flexibility and security) and outcomes (serious psychological distress, anxiety, work absenteeism, work presenteeism, and mental health care utilization) among U.S. adults will be less pronounced among women compared to men due to gendered differences in caregiving responsibilities and workplace stressors. Understanding gender-specific differences in how job flexibility and security may impact mental health and related outcomes, particularly during times of challenge and crisis, contributes to the scientific literature on gender health disparities associated with job characteristics, examines the application of the Job Demands-Resources model in a U.S. and pandemic context, and may inform workplace policies and interventions that promote mental well-being across the workforce.

2. Methods

2.1. Data source and study population

This study utilized data from the 2021 NHIS, an annual survey conducted by the National Center for Health Statistics via face-to-face interviews with a random sample of U.S. adults across 50 states and the District of Columbia. Using a multistage probability sampling design, the NHIS incorporates both stratification and clustering to obtain a representative sample of the U.S. population. In 2021, 29,482 individuals aged 18 years and above completed the survey, with N = 18,112 males and females reporting employment within the previous 12 months; this sample represents 168,068,586 civilian, non-institutionalized working U.S. adults. Additional details describing the study population and methods are described previously (Wang et al., 2024).

2.2. Exposure measures

Job flexibility was assessed by creating a summative variable using responses from the following questions: “How easy or difficult is/was it for you to change your work schedule to do things that are important to you or your family? Would you say very easy, somewhat easy, somewhat difficult, or very difficult?” (1 = very or somewhat easy; 0 = somewhat or very difficult); “Did your work schedule at your main job change on a regular basis?” (yes = 0; no = 1); “Approximately how far in advance does your employer usually tell you/do you usually know/did your employer usually tell you/did you usually know—do you/your employer tell you the hours that you—needed/will need—schedule to work on any given day?” (less than a week = 0; more than a week = 1). The summative variable ranged from 0 to 3, with higher scores indicating greater job flexibility.

Job security was assessed by the question: "Regarding the next 12 months, how likely do you believe it is that you will lose your job or be laid off?" Responses were coded as follows: 3 = not at all likely; 2 = somewhat likely; 1 = fairly likely; 0 = very likely, with higher scores indicating greater perceived job security. Both job flexibility and security were examined as continuous variables to allow for a more nuanced exploration of their associations with the three outcome measures.

2.3. Outcome measures

Mental health: Serious psychological distress was measured using the Kessler-6 (K6) scale, which asks about experiences of depression, nervousness, hopelessness, restlessness, worthlessness, and feeling that everything was an effort over the past 30 days (Kessler et al., 2002). Each item utilized a five-point Likert scale (e.g., 0 = none of the time; 4 = all of the time); scores were then summed to yield a quasi-interval scale (range of 0–24). The clinically validated threshold of 13 was used to classify those with serious psychological distress (scores of 13+) or not (scores of 12 or lower) (Kessler et al., 2003). Frequency of anxiety was measured by asking participants “How often do you feel worried, nervous, or anxious?” with daily, weekly, monthly, a few times a year, or never as response options.

Work absenteeism was assessed using two survey items: (1) Number of work days missed while feeling ill over the past 3 months (0-90); (2) Number of work days missed in the past 12 months (0–130+). Work presenteeism was assessed by asking respondents the number of days they worked while feeling ill over the past 3 months (0-90). Mental health care utilization was assessed by asking respondents if they received mental health counseling over the past 12 months (yes/no); and (2) if they currently receive mental health counseling (yes/no).

2.4. Covariates

The following covariates were included in the adjusted models to account for factors that could influence job flexibility, job security, and mental health outcomes, thereby reducing potential bias in the estimated associations. Age (years), race, ethnicity, federal poverty level, and education level were included to control for sociodemographic characteristics that may shape access to job opportunities and related health disparities. Marital status, number of children and adults in the household, and place of birth were considered to account for family dynamics and social contexts that could impact job-related stressors and mental health. Health insurance coverage, disability status, and self-reported chronic health conditions were included to adjust for underlying health factors that might confound the relationship between job characteristics and mental health outcomes.

We additionally controlled for depression diagnosis as depression and anxiety often co-occur (Belzer & Schneier, 2004), and lack of job security and flexibility have been shown to be linked with depressive symptoms (Kim & von dem Knesebeck, 2016; Shiri et al., 2022). If not controlled for, the associations between job flexibility, job security, and the mental health outcomes we examined in this present study may be overestimated by pre-existing depression. Self-reported sex (male or female) was examined as an effect modifier. The study was exempt from human subjects research by the Institutional Review Board (protocol number 275805), as the NHIS data are de-identified and publicly accessible. We followed STROBE reporting guidelines for observational studies for this research.

2.5. Statistical analysis

Exposure, outcome, and covariate measures were analyzed for females and males using means for continuous variables and percentages for categorical variables, with 95% confidence intervals (CI). Rao-Scott Chi-Square test of independence was computed to assess significant sex-based differences in exposure and outcome measures.

A series of regression models were fitted according to the measurement scale of the outcome variables: (1) multivariable logistic regression for serious psychological distress; (2) multinomial logistic regression for frequency of anxiety; (3) negative binomial regressions for work absenteeism; and (4) multivariable logistic regression for mental health utilization. The magnitude of the associations was assessed using odds ratios (OR) for logit models and incidence rate ratios (IRR) for count models along 95% CIs. Given that interpreting estimates of multiplicative interaction terms is complex in non-linear models (Karaca-Mandic et al., 2012), we opted to present graphical representations of effect modifications by sex in the associations of interest, with the probabilities of outcomes plotted for statistically significant interactions identified through adjusted Wald tests. All regression models adjusted for the listed covariates, and the plotted estimated probabilities were derived from these fully adjusted models to account for the effects of sociodemographic and health-related variables.

All descriptive and regression analyses accounted for the NHIS complex survey design, incorporating sampling weights, primary sampling units, and stratification to generate population-representative estimates. Missing data accounted for <5% with imputed income data included, rendering complete case data analysis possible and statistically robust due to the low level of missingness. All analyses were performed with STATA/SE 17 using the svy prefix (StataCorp, 2021). Statistical significance was set a priori at alpha 0.05.

3. Results

3.1. Descriptive analysis

3.1.1. Sample characteristics by sex

The 18,112 respondents represent a population of 168,068,586 civilian, non-institutionalized working U.S. adults (47.7% female with a mean age of 42.1 years (95% CI: 41.6–42.5) and 52.3% males with a mean age of 42.4 years (95% CI: 42.0–42.8). See Table 1 for additional sociodemographic data by sex.

Table 1.

Study population sociodemographic characteristics and health conditions by sex (N = 18,112).

Measures	Weighted % (95% CI)
	Male (n = 8965)a	Female (n = 9147)a
Biological sex	52.3 (51.5–53.2)a	47.7 (46.8–48.5)a
Age (18–85+ years): mean	42.4 (42.0–42.8)	42.1 (41.6–42.5)
Race/ethnicity
White, non-Hispanic	33.1 (32.0–34.1)	28.8 (27.8–29.9)
Black, non-Hispanic	5.3 (4.8–5.8)	6.1 (5.6–6.8)
Hispanic	9.4 (8.6–10.3)	8.3 (7.6–9.1)
Other, non-Hispanic	4.6 (4.2–5.1)	4.4 (4.0–4.8)
Education
≤ Bachelor's degree	32.1 (31.1–33.0)	27.5 (26.6–28.4)
> Bachelor's degree	20.3 (19.5–21.1)	20.2 (19.5–20.9)
Marital status
Single/divorced/widowed	18.7 (18.0–19.4)	19.5 (18.8–20.3)
Married/partnered	33.7 (32.9–34.5)	28.1 (27.4–28.9)
Born in the US
No	10.2 (9.4–10.9)	8.7 (8.1–9.3)
Yes	42.2 (41.3–43.2)	39.0 (38.1–39.9)
Household composition
Number of adults in household: mean	2.10 (2.08–2.12)	2.08 (2.06–2.10)
Number of children in household: mean	0.66 (0.64–0.69)	0.70 (0.67–0.72)
Health insurance coverage
No	6.1 (5.6–6.6)	3.6 (3.2–4.0)
Yes	46.3 (45.4–47.2)	44.1 (43.2–44.9)
Disability status
Not disabled	50.6 (49.8–51.5)	45.6 (44.8–46.4)
Disabled	1.7(1.5–1.9)	2.1 (1.8–2.3)
Federal Poverty Level (%FPL)
< 100	3.3 (2.9–3.7)	3.9 (3.5–4.3)
100 - 199	7.1 (6.6–7.6)	7.7 (7.2–8.2)
200 - 299	3.7 (3.4–4.1)	3.8 (3.5–4.2)
300 - 399	7.6 (7.1–8.0)	6.9 (6.5–7.4)
≥ 400	30.7 (29.9–31.5)	25.3 (24.5–26.1)
Ever diagnosed with depression
No	46.9 (46.1–47.8)	37.7 (36.9–38.6)
Yes	5.4 (5.0–5.8)	9.9 (9.4–10.5)
Number of chronic conditions other than depression (1-10): mean	0.69 (0.66–0.71)	0.70 (0.68–0.73)

Note: Analytical study sample size (n = 18,112) and weighted sample size estimated at a representative population (N = 168,068,586) of adult civilians living in the United States. Estimates are weighted. The NCHS top-codes all individuals aged 85+ to protect the confidentiality among the oldest adults.

^aFrequencies for female and male subgroups “n” are unweighted and represent the number of participants in the analytic sample, whereas the corresponding percentages are weighted to reflect population estimates.

Source: 2021 National Health Interview Survey – National Center of Health Statistics (NCHS)

Table 2 presents bivariate analyses examining outcome and exposure measures by sex. Twice as many females experienced serious psychological distress than males (3.9 vs. 1.8; p < 0.001). Females reported daily, weekly, and monthly anxiety at significantly higher rates than males (15.9% vs. 8.9; 18.6% vs. 13.8%; 14.6% vs. 12.1%, respectively; p's < 0.001).

Table 2.

Results of bivariate associations of outcomes and exposures by sex (N = 18,112).

Measures	Weighted% (95% CI)
	Male	Female	p-values
Outcomes
Psychological distress			<0.001
Did not experience serious psychological distress (Kessler 6 < 13)	98.2 (97.8–98.5)	96.1 (95.5–96.5)
Experienced serious psychological distress (Kessler 6 ≥ 13)	1.8 (1.5–2.2)	3.9 (3.5–4.5)
Frequency of anxiety			<0.001
Never	33.6 (32.9–34.9)	19.6 (18.6–20.7)
A few times a year	31.5 (30.5–32.6)	31.2 (30.0–32.4)
Monthly	12.1 (11.4–12.9)	14.6 (13.7–15.6)
Weekly	13.8 (13.1–14.7)	18.6 (17.6–19.7)
Daily	8.9 (8.2–9.6)	15.9 (15.0–16.9)
Mental health utilization
Received mental health counseling over the past 12 months (No)	91.7 (91.0–92.4)	85.2 (84.2–86.1)	<0.001
Received mental health counseling over the past 12 months (Yes)	8.3 (7.6–9.0)	14.8 (13.9–15.8)
Currently receiving mental health counseling (No)	40.3 (36.3–44.4)	38.8 (35.5–42.1)	0.556
Currently receiving mental health counseling (Yes)	59.7 (55.6–63.7)	61.2 (57.9–64.5)
Work absenteeisma
Number of work days missed while feeling ill over the past 3 months (0)	83.1 (82.1–84.1)	80.3 (79.2–81.3)	<0.001
Number of work days missed while feeling ill over the past 3 months (1+)	16.9 (15.9–17.9)	19.7 (18.7–20.8)
Number of work days missed while feeling ill over the past 3 months (1+): mean	7.8 (7.0–8.6)	7.6 (6.8–8.4)	0.728
Number of days worked while feeling ill over the past 3 months (0)	82.5 (81.6–83.5)	80.5 (79.4–81.5)	0.003
Number of days worked while feeling ill over the past 3 months (1+)	17.5 (16.5–18.4)	19.5 (18.5–20.6)
Number of days worked while feeling ill over the past 3 months (1+): mean	10.4 (8.9–11.8)	10.8 (9.6–12.1)	0.593
Number of work days missed in the past 12 months (0)	63.2 (62.1–64.4)	56.9 (55.6–58.2)	<0.001
Number of work days missed in the past 12 months (1+)	36.8 (35.6–37.9)	43.1 (41.8–44.4)
Number of work days missed in the past 12 months (1+): mean	11.2 (10.3–12.1)	11.9 (11.0–12.7)	0.275
Exposures
Job flexibility: mean	1.06 (1.04–1.08)	1.06 (1.04–1.08)	0.827
Job security: mean	2.78 (2.76–7.80)	2.78 (2.76–2.80)	0.935

Note: CI = confidence intervals; alpha = 0.05; p = probability of type 1 error; p-values in bold indicate statistical significance.

^aWork absenteeism variables measured in continuous days was dichotomized as zero and ≥1 day.

Source: 2021 National Health Interview Survey – National Center of Health Statistics (NCHS).

A higher percentage of females missed at least one day of work while ill than males (19.7% vs. 16.9%; p < 0.001). Similarly, a higher percentage of females vs. males worked at least one day while ill (19.5 vs. 17.5%; p = 0.003). No sex differences emerged in the average number of work days missed while ill over the past 3 months (mean difference: −0.19; p = 0.728) or the average number of days worked while ill over the past 3 months (mean difference: 0.48; p = 0.593). Although the proportion of 12-month work absenteeism was higher among females vs. males (43.1% vs. 36.8%; p < 0.001), the average number of work days missed over the past 12 months did not differ by sex (mean difference: 0.65; p = 0.275).

More females (14.8%) reported receiving mental health counseling in the past year than males (8.3%); p < 0.001. At the time of survey administration, 61.2% of females and 59.7% of males reported receiving mental health counseling (p = 0.556). Job flexibility and security measures did not vary by sex (mean differences: 0.003 and 0.001; p = 0.827 and p = 0.935, respectively).

3.2. Regression analysis

We interpreted estimates of effect modification regression analyses (i.e., multiplicative interactions) plotted as probabilities for each outcome of interest, controlling for confounders and accounting for NHIS complex survey design. All p-values were derived from the complex design-adjusted Wald test of the multiplicative interactions. Estimates for significant associations are included as supplemental tables.

3.2.1. Serious psychological distress

Associations between perceived job security and serious psychological distress varied by sex (p = 0.040). Notably, females reporting job loss as “very likely” had a decreased probability of experiencing serious psychological distress than males reporting job loss as very likely. These patterns changed with increasing levels of job security, with females experiencing greater serious psychological distress than males when they rated job loss as “fairly likely,” “somewhat likely,” or “not at all” (Fig. 1).

Fig. 1.

Association between Job Security (likelihood of job loss) and Serious Psychological Distress: Effect Modification of Sex. Note: Wald Test F_(3,589) = 2.79; p = 0.040 adjusted for complex survey design. Associations are adjusted for age (years), race, ethnicity, federal poverty level, education level, marital status, number of children and adults in the household, place of birth, health insurance coverage, disability status, self-reported chronic health conditions, and depression diagnosis.

Associations between job flexibility and serious psychological distress did not significantly vary by sex, with both males and females exhibiting a higher probability of serious psychological distress at the lowest level of job flexibility and a decreasing probability of serious psychological distress as job flexibility increased (all p-values>0.05) (Supplemental Table 1).

3.2.2. Frequency of anxiety

Associations between differences in job flexibility and anxiety and in job security and anxiety did not significantly vary by sex. Although females consistently demonstrated a slightly higher probability of experiencing anxiety on a daily, weekly, or monthly basis than males across job flexibility levels, the associations between job characteristics and anxiety were similar across sex.

3.2.3. Work presenteeism and absenteeism

Females with the lowest job flexibility (score of 0) reported more days worked while ill than men, while males with higher job flexibility (scores of 2 and 3) reported more days worked while ill than women (p = 0.007) (Fig. 2; Supplemental Table 2). The association between job security and 12-month work absenteeism also differed by sex. The average number of work days missed over 12 months was 5.1 days among females vs. 10.0 days for males when job loss was perceived as “fairly likely”; 6.9 days among females and 3.8 days among males when job loss was perceived as “somewhat likely; and 4.8 days among females and 3.8 days among males when job loss was perceived as “not at all likely” (p = 0.005) (Fig. 3; Supplemental Table 3). When job loss was perceived as very likely, females and males reported nearly equal numbers of missed work days (4.2 and 4.5) over the past 12 months.

Fig. 2.

Association between Job Flexibility and Work Presenteeism (number of days worked while feeling ill over the past 3 months): Effect Modification of Sex. Note: Wald Test F_(3,589) = 4.1; p = 0.007 adjusted for complex survey design. Associations are adjusted for age (years), race, ethnicity, federal poverty level, education level, marital status, number of children and adults in the household, place of birth, health insurance coverage, disability status, self-reported chronic health conditions, and depression diagnosis.

Fig. 3.

Association between Job Security (likelihood of job loss) and Work Absenteeism (number of work days missed in the past 12 months): Effect Modification of Sex. Note: Wald Test F_(3,589) = 4.3; p = 0.005 adjusted for complex survey design. Associations are adjusted for age (years), race, ethnicity, federal poverty level, education level, marital status, number of children and adults in the household, place of birth, health insurance coverage, disability status, self-reported chronic health conditions, and depression diagnosis.

The associations between job flexibility, job security, and other measures of work absenteeism did not differ by sex. No significant sex differences were found in the associations between job characteristics (job flexibility and security) and receiving mental health counseling, either currently or in the past 12 months. As part of a sensitivity analysis, we conducted sex-stratified adjusted models (Supplemental Table 4), and results were consistent with those reported in the primary models.

4. Discussion

Our study demonstrated notable gender differences in the association between job security and psychological distress, as well as job flexibility, work absenteeism, and work presenteeism, among a nationally representative sample of working U.S. adults during the COVID-19 pandemic. Contrary to our hypotheses, the protective effects of job flexibility and security on mental health were not observed to be overall less pronounced among women than men. Women with low job security exhibited lower rates of psychological distress than men, while women with the least job flexibility were more likely to work while ill than men with higher flexibility. In terms of work absenteeism, when job security was perceived as low, women reported fewer missed workdays compared to men. However, as job security increased, women missed more work days than men.

Gender differences in the association between job flexibility and security and mental health have been previously demonstrated in studies conducted in other countries, though prior findings are distinct from the ones that emerged in our study. A study in England reported that lower job flexibility was associated with poorer mental health, with greater negative effects among women than men (Moss et al., 2022). An Australian panel study showed that job security improvements were associated with reduced depression and anxiety, particularly for men (LaMontagne et al., 2021). Similarly, UK panel data indicated that men and higher occupational classes gained more mental health benefits from greater job flexibility, while women in lower classes did not see similar benefits (Lu et al., 2023). The lack of similar findings in our study suggests the importance of considering how national labor policies, gender roles, work culture, and societal expectations may shape the experience and impact of job flexibility and security on mental health and work absenteeism differently across countries.

The observed lower rates of psychological distress among women with low job security may be attributed to several mechanisms, including differences in how men and women respond to mental health challenges. Nearly twice as many women in our study reported receiving mental health counseling in the past year than men (14.8% vs. 8.3%). Research shows that women tend to seek mental health support and utilize mental health services more often than men (Nam et al., 2010), which can lead to lower levels of psychological distress. Additionally, gender differences in financial responsibility and societal expectations can contribute to these disparities. Even in dual-income households, men often feel greater pressure to be the primary income-earner, leading to greater stress when job loss is likely (Kham, 2020). Women also tend to have stronger support networks, seek support more often than men (Matud, 2004), and are more likely to use a wider array of coping strategies to manage stress (Tamres et al., 2002). In contrast, men are more likely to rely on fewer coping strategies, such as social withdrawal, which may exacerbate distress in times of job insecurity (Menendez-Espina et al., 2019).

Our study also showed important gender distinctions in work absenteeism and presenteeism associated with job characteristics. Women with the lowest job flexibility reported more days worked while ill, while men with the highest job flexibility reported more days worked while ill. When job security was perceived as low, women reported fewer missed workdays compared to men. However, as job security increased, women missed more workdays than men. Women with low job flexibility and low job security may prioritize work over their health due to several factors, including gender differences in professions and family responsibilities (Herr et al., 2020), fear of negative job evaluations or missed career advancement opportunities (Glass, 1990), and greater financial loss when taking leave (Herr et al., 2020). These structural disadvantages are particularly pronounced for women in part-time or lower-wage positions with limited flexibility and benefits (Landivar et al., 2022), which may compel them to prioritize work over their health. These challenges were further exacerbated during the COVID-19 pandemic as working mothers balanced homeschooling with concerns about long-term impacts on their families and labor market participation (Schaeffer, 2022).

On the other hand, men with higher job flexibility may work more while ill for different reasons. High flexibility can increase the expectation that employees should remain available, even when sick, as they have more control over their schedules. The economic pressures and uncertainty caused by the pandemic likely heightened the expectation of constant availability, particularly for those in flexible work arrangements. Studies have demonstrated gender differences in work presenteeism prior to (Gustafsson Senden et al., 2016; Kwon, 2020) and during the pandemic (Biron et al., 2021; Magalhaes et al., 2022), with financial reasons cited as a reason to continue working even when ill. Men may also feel societal pressure to meet work expectations, driven by traditional norms that encourage them to prioritize work and financial stability (Kham, 2020). This pressure, combined with perceived expectations of constant availability due to increased job flexibility in the context of a pandemic, may explain why men in our study with high job flexibility were more likely to work while unwell.

Though we did not find gender differences in the association between job flexibility and anxiety, women consistently reported higher anxiety than men across all levels of job flexibility. This may be due to the pre-existing disproportionate burden of caregiving responsibilities working women shoulder (Yucel & Fan, 2023), which worsened as mothers managed more childcare and homeschooling than fathers during pandemic closures and may have led to greater anxiety (Schaeffer, 2022). Women generally reported higher social anxiety due to the COVID-19 pandemic than men (Kindred & Bates, 2023), which aligns with the broader trend of increased depression, anxiety, and stress levels among women during the first few years of the pandemic (Malkawi et al., 2021).

The absence of gender differences in the association between job flexibility and mental health care counseling, as well as job security and mental health care counseling, may stem from various factors. Firstly, mental health care utilization is influenced by multifaceted factors beyond job characteristics alone, including socioeconomic status, health insurance coverage, access to mental health care services, and perceived stigma (Coombs et al., 2021). Additionally, workplace factors, such as organizational culture, support systems, and available resources may play a crucial role in shaping employees' decisions to seek mental health care (Monteiro & James, 2023). Therefore, while job flexibility and security are important determinants of mental health, their direct impact on receiving mental health care counseling may be influenced by other contextual factors.

Our findings contribute to the scientific literature by highlighting how gender influences the associations between job resources, mental health, work absenteeism, and work presenteeism in the U.S. during the COVID-19 pandemic, a period in which gender disparities across work, health, and family domains were exacerbated (Flor et al., 2022; Herr et al., 2020; Lyttelton et al., 2022a, Lyttelton et al., 2022b; Yucel & Fan, 2023). Our study also offers a more nuanced application of the Job Demands-Resources model by explicitly considering how gender may differentially shape the effects of job demands and resources, particularly in contexts where caregiving burdens and access to support systems are unevenly distributed and inadequate (Aizer et al., 2022; de Souza et al., 2022; Dubbelt et al., 2016; Harknett & Schneider, 2022; Wething, 2022). By acknowledging how gender may substantially shape the link between work and health in U.S. culture and society in ways distinct from other countries, our study offers insights into the intersection of job characteristics, gender, and mental health in the American workforce.

Study findings underscore the critical need for gender-sensitive workplace policies that address job flexibility and security as well as the broader structural inequities that disproportionately affect women at work and at home. Policies to promote mental health in the workplace could include strategies to improve job design for all employees, with particular attention to positions with low flexibility and security, which are disproportionately held by women and people of color (Menendez-Espina et al., 2020; Yucel & Fan, 2023). This could involve experimenting with flexible work arrangements, such as telecommuting or flexible scheduling, to accommodate diverse personal and caregiving responsibilities and promote work-life balance (Chatterjee et al., 2022; Ray & Pana-Cryan, 2021), as well as expanding paid parental and sick leave and including mental health coverage in benefits plans (Harknett & Schneider, 2022; Schneider & Harknett, 2019; Wething, 2022). Future studies can employ longitudinal designs to explore how changes in job flexibility and security influence mental health over time, particularly across different gender and racial/ethnic groups in the U.S. Overall, addressing gender disparities in the U.S. workplace requires a multifaceted approach that combines policy, practice, and research efforts to promote mental well-being for all employees.

4.1. Limitations

While the associations observed in this study provide valuable insights, several study limitations must be acknowledged. First, the cross-sectional design limits our ability to establish causality or infer directionality, though a number of prospective studies and randomized and quasi-experimental studies have demonstrated that job flexibility and security predict mental health (Kim & von dem Knesebeck, 2015; 2016; Shiri et al., 2022). Individuals with poorer mental health might select into or be restricted to jobs with lower flexibility or security, rather than these job characteristics directly causing poor mental health outcomes. Second, our reliance on self-reported data for mental health, work absenteeism, work presenteeism, and use of mental health counseling introduces potential reporting bias, as participants may either overestimate or underestimate their experiences. Third, our measure of mental health care utilization focuses solely on counseling, excluding other potentially relevant treatment modalities such as pharmacological interventions that influence mental health. Fourth, confounding by occupational status, hours worked, and hours spent on domestic labor presents additional challenges, as women spend more time on domestic labor and childcare than their male peers and are overrepresented in part-time, lower-wage positions that may offer less flexibility and security. Measuring and adjusting for occupational and domestic labor differences in future studies may clarify the complex interplay between job characteristics, gender, and mental health. Finally, survey measures in this study did not assess gender identity, type of occupation (e.g., professional or managerial vs. non-professional), or specify whether the reported anxiety, distress, work absenteeism, or counseling utilization were directly linked to job characteristics. Future studies using longitudinal, experimental, or qualitative methods can address these gaps to gain a more nuanced understanding of how job characteristics influence mental health, work absenteeism, and work presenteeism.

5. Conclusion

This study provides evidence of significant gender disparities in the relationship between job characteristics, mental health, and work absenteeism among working U.S. adults. These findings emphasize the importance of designing workplace interventions prioritizing job flexibility and security while acknowledging and addressing structural gender inequities in work. Future research should further explore these relationships, focusing on caregiving responsibilities and workplace culture, to inform more effective interventions and policies that promote well-being across the workforce.

CRediT authorship contribution statement

Monica L. Wang: Writing – review & editing, Writing – original draft, Conceptualization. Marie-Rachelle Narcisse: Writing – review & editing, Visualization, Formal analysis, Conceptualization. Kate Rodriguez: Writing – review & editing. Pearl A. McElfish: Writing – review & editing.

Data sharing statement

The 2021 National Health Interview Survey (NHIS) is openly made available by the Centers for Diseases Control and Prevention (CDC) at https://www.cdc.gov/nchs/nhis/2021nhis.htm. No datasets were generated for this study. Any analysis, interpretation, and/or conclusion based on the NHIS 2021 data is solely that of the authors. Opinions, conclusions, and recommendations expressed herein do not necessarily represent those of the National Center for Health Statistics or CDC, which are responsible for the data.

Ethical statement

The study was exempt from human subjects research by the Institutional Review Board (protocol number 275805), as the NHIS data are de-identified and publicly accessible. We followed STROBE reporting guidelines for observational studies for this research.

Funding

The work was supported by University of Arkansas for Medical Sciences Translational Research Institute funding awarded through the National Center for Advancing Translational Sciences of the National Institutes of Health (NIH) (UL1 TR003107). Dr. Wang is supported in by part by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) Grant #R01DK120713 (PI: Wang). Dr. Narcisse is supported by the Bradley Hospital COBRE Center for Sleep and Circadian Rhythms in Child and Adolescent Mental Health funded by the National Institute of General Medical Sciences (NIGMS) grant number P20GM139743. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. Funding sources had no role in the study design; in the collection, analysis, or interpretation of data; in the writing of the report; or in the decision to submit the paper for publication.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article:

Data availability

NHIS data are de-identified and publicly accessible