Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Nov 21.
Published in final edited form as: Stress Health. 2022 Dec 1;39(3):614–626. doi: 10.1002/smi.3210

Metabolic Health Disparities Driven by Financial Stress: Behavioral Adaptation or Modification?

Wan-Chin Kuo 1, Lisa C Bratzke 1, Erika W Hagen 2, Lauren Hale 3, Roger L Brown 1, Jodi H Barnet 2, Paul E Peppard 2
PMCID: PMC12632441  NIHMSID: NIHMS2103944  PMID: 36413205

Abstract

Financial stress has been linked to an increased risk of metabolic syndrome, yet, it remains unclear whether suboptimal sleep duration and physical inactivity are the adaptive responses to financial stress or effect modifiers in the association between financial stress and metabolic syndrome. Hence, this study aims to examine whether physical activity and sleep duration mediate or moderate the bivariate association between financial stress and metabolic syndrome. A prospective secondary analysis was conducted using data from the Wisconsin Sleep Cohort (WSC) Study (N=445, mean [SD] age = 64 [7] years). Baseline moderation effect was examined using subgroup analysis with model constraints; prospective mediation model was examined using bias-corrected bootstrap confidence intervals. Results indicate that participants with higher financial stress were less likely to meet physical activity and sleep recommendations. Baseline moderation analysis indicates that meeting current recommendations of sleep duration and physical activity attenuated the association between financial stress and metabolic syndrome. In the prospective mediation analysis, weekly physical activity levels partially mediated the relationship between financial stress and metabolic syndrome, but sleep duration did not mediate this relationship. In conclusion, the joint effect of optimal sleep duration and physical activity disassociates financial stress from the risk of metabolic syndrome. Future interventions addressing metabolic risk might achieve better outcomes if clinicians and researchers factor in the behavioral adaptation of physical inactivity in financially stressed adults. (Clinical Trial Registration: NCT00005557).

Keywords: sleep duration, physical activity, financial stress, metabolic syndrome, mediation, moderation

1|. Introduction

Financial stress is defined as the perception of having inadequate financial resources to meet a typical standard of living (Steen & MacKenzie, 2013). A standard of living refers to the average services and material goods that are available and affordable for a person or population to maintain daily expenditure, often indicated by gross domestic product (GDP) per capita (Rao & Min, 2018). However, individuals might expect various standards of living depending on personal values, demographics, beliefs, and living environment (Easterlin, 2000). As a subjective perception, financial stress comprises three main constructs, including perceived inability to afford life’s necessities, perceived inability to pay monthly bills, and worrying about current or future financial status as an individual or household (Agrigoroaei et al., 2017; Kuo et al., 2021).

Financial stress not only dictates adults’ daily consumer behaviors, but also impacts individuals’ access to preventive health services and healthy choices, which, in turn, can affect health and well-being, including metabolic outcomes (Epel, 2009). Metabolic health disparities refer to the preventable differences in the etiology and burden of cardiometabolic diseases across subgroups, which are driven by various social determinants of health, including racial, ethnic, or sexual minorities, financial disadvantages, socioeconomic statuses, and occupational exposures (Caceres et al., 2022; Commodore-Mensah et al., 2021). In the past ten years, the prevalence of metabolic syndrome in the U.S. has increased from 25% to 34% (Aguilar et al., 2015; Moore et al., 2017), and the prevalence is disproportionately higher among low-income and financially deprived populations, compared to those who live in better financial conditions with adequate financial resources (Loucks et al., 2016). However, for more than three decades, research on the etiology of metabolic syndrome has predominantly focused on the proximal causal factors, such as energy imbalance, unhealthy diet, physical inactivity, and suboptimal sleep (Hostinar et al., 2017). There is a dearth of evidence examining how socioeconomic contexts shape these proximal causal factors and metabolic health disparities.

Financial stress is a form of chronic adversity, which cannot be simply explained by acute stress response. Acute stress activates physical and psychological arousal through the activation of hypothalamic-pituitary-adrenal (HPA) and sympathetic-adreno-medullar (SAM) axes, during which individuals implement stress responses (e.g., fight or flight) to resolve or avoid the threats (Umberson et al., 2008). When life stressors cannot be resolved or avoided through stress responses, these stressors become chronic (e.g., financial stress). The chronic activation of negative feedback and compensatory downregulation can result in HPA dysregulation in neurotransmitters and hormones (e.g., abnormal levels of dopamine or DHEA/Cortisol Ratio) (Epel et al., 2018). HPA dysregulation not only results in allostatic load with physiological adaptation but also induces behavioral adaptation through a shift from goal-directed cognitive system to habitual responding system (Meier et al., 2022; Schwabe & Wolf, 2009; Wood et al., 2022). Habits are behavioral adaptations shaped by three elements: (1) environmental cues, (2) repetition, and (3) reward. Habit helps individuals reserve cognitive efforts and therefore, reduce perceived stress (Wood et al., 2022). Unfortunately, due to environmental injustice, those with financial stress are more likely to live in food deserts or unsafe neighborhood (i.e., environment cues) and experience stress proliferation (i.e., repetition of stress responses) (Crowe et al., 2018). The behavioral adaptation is often strengthened by the feeling of momentary satisfaction as rewards (e.g., engaging in sedentary behaviors or tasting palatable food). Despite the theoretical premises explaining the behavioral adaptation under chronic stress, existing evidence examining behavioral adaptation of financial stress and its mediation role in metabolic health disparities is limited and inconsistent. For instance, Beenackers et al. found that financial stress is associated with low levels physical activity (Beenackers et al., 2018), but Ensel found no association between life stressors and physical activity levels (Ensel & Lin, 2004).

A potential approach to alleviate the metabolic health disparities driven by financial stress is to characterize how the relationship between financial stress and metabolic outcomes varies across subsets of behavioral patterns and design interventions to modify risk behaviors in those who are mostly affected by financial stress (Epel et al., 2018; Hamer, 2012; MacKinnon, 2011). Behavioral modification fosters individuals’ protective factors, which, in turn, buffer the negative influence of chronic stress on health outcomes. However, existing evidence that examined the stress-buffering effects of sleep and physical activity on metabolic outcomes has shown mixed findings depending on sources of stress, metabolic phenotypes, and study populations (Du et al., 2022; Puterman et al., 2010; Schilling et al., 2020). For instance, Puterman et al. found that physical activity protects financially stressed adults from insulin resistance (Puterman E et al., 2012), protects post-menopausal stressed women from premature cellular aging (Puterman et al., 2010), and protects stressed teenage girls from overweight and obesity (Puterman et al., 2016). Yet, Schilling et al. and Du et al. found that physical activity and sleep duration did not moderate the stress-metabolism relation among police officers (Schilling et al., 2020) or college students (Du et al., 2022). Furthermore, these studies have predominantly focused on the stress-buffering effect of single behavior (i.e., either sleep duration or physical activities), without further investigation of the joint stress-buffering effects of adequate sleep and physical activity.

Researchers in prevention science have proposed that protective behaviors (e.g., optimal sleep duration and physical activity) might both moderate and mediate the bivariate association between the exposure and disease outcomes (D’Lima et al., 2012; Karazsia & Berlin, 2018). However, there is a lack of empirical evidence to support the mediation and moderation effects of physical activity and sleep duration in the association between financial stress and metabolic syndrome. What remains unclear is whether and how sleep duration and physical activity explain the metabolic health disparities experienced by persons with financial stress.

Hence, this study has two specific aims:

  1. To examine whether and to what extent optimal sleep duration and physical activity moderate the association between financial stress and the prevalence of metabolic syndrome.

  2. To examine whether physical activity and sleep duration partially mediate the association between financial stress and the prevalence of metabolic syndrome.

2|. Methods

Study design, sample, and setting

This secondary data analysis used the deidentified data from the Retirement and Sleep Trajectories (REST) study, an ancillary study of the Wisconsin Sleep Cohort (WSC) study. The WSC study is an ongoing longitudinal study of sleep habits and sleep disorders in adults followed prospectively since 1988. The design of the WSC has been described previously (Peppard et al., 2000). Briefly, participants in the WSC were selected from a random sampling pool of payroll records for State of Wisconsin employees aged between 30 and 60 years old in 1988. In 2010, WSC participants who completed either 3 WSC mailed surveys or 1 WSC in-laboratory study protocol and 1 WSC survey prior 2010 were invited to participate in the REST study. The REST study is a longitudinal study consisting of four annual mailed surveys, sent between 2010 and 2015 (Wave 1 to Wave 4). REST surveys collected information regarding retirement status, health behaviors, sleep health, chronic stressors, and other psychosocial factors (Hagen et al., 2016).

In this secondary data analysis, we included 445 participants who met the following inclusion criteria: (1) participated in REST Wave 1 survey without missing data in financial stress, physical activity, and sleep duration, and (2) completed at least one WSC in-laboratory assessment after 2011 without missing data in metabolic biomarkers. The flow chart of final sample is presented in Figure 1.

Figure 1.

Figure 1.

Open in a new tab

The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: Flow chart of the sample size in the present study.

For the prospective mediation analysis, we adopted the prospective timeframe illustrated in Figure 2. Specifically, for each participant who completed a WSC in-laboratory assessment, we traced the participant back to the nearest physical activity and sleep data collected between REST Wave 1 and Wave 4. Therefore, the exposure variable (financial stress) and hypothesized mediators (weekly physical activity and sleep duration per night) occurred before the outcome variable (metabolic syndrome) to ensure a temporal sequential order. The time differences in data collection between the exposure variable and outcome variable were transformed into months and included in the mediation and moderation models as a covariate. The SAS coding for data pre-processing in prospective mediation analysis is provided in Supplement 1. This secondary data analysis and the parent studies were approved by the Health Sciences Institutional Review Board (IRB) at the University of Wisconsin-Madison.

Figure 2.

Figure 2.

Open in a new tab

Schematic of timeline and epochs of variables used in this prospective data analysis.

Note. For each participant with an WSC in-laboratory visit, we traced the participant back to the nearest sleep/PA data. Therefore, the exposure variable (FS) and health behaviors (sleep and PA) occurred before the outcome variable (MetS) to ensure a temporal sequential order.

Abbreviations. FS=Financial stress. PA=Physical activity. MetS=Metabolic syndrome. DV=Dependent variable. IV=Independent variable.

Measures

Metabolic syndrome.

Metabolic syndrome was identified through medication reconciliation and in-laboratory assessment of metabolic biomarkers, including waist circumference, resting blood pressure, and blood tests that measured fasting glucose, triglycerides, and high-density lipoprotein cholesterols (HDL). Medication reconciliation was performed through individual interview and pill-bottle assessments to identify participants who took prescribed medication to control blood pressure (e.g., beta-blockers, diuretics, ACE inhibitors, angiotensin II receptor blockers, or calcium channel blockers), dyslipidemia (e.g., statin, fibrates, niacin, or ezetimibe), and hyperglycemia (e.g., metformin, sulfonylureas, DPP-4 inhibitors, or insulin). Diagnostic criteria of metabolic syndrome were based on current guidelines provided by the American Heart Association National Cholesterol Education Program (see Table 1).

Table 1.

Criteria for metabolic syndrome in current study

Outcome Criteria Guidelines
Metabolic syndrome Met at least three of the criteria below:
 • BP ≥130/85 mmHg or medicated
 • FPG ≥100 mg/dL or medicated
 • TG ≥150 mg/dL or medicated
 • HDL <40 mg/dL (male), <50 mg/dL (female), or medicated
 • WC >40 inches (male), >35 inches (female)		 NCEP-ATP III, NHLBI, AHA
Open in a new tab

Note. T2DM=Type 2 diabetes Mellitus. BP=Blood pressure. FPG=Fasting plasma glucose. TG=Triglycerides. HDL=High-density lipoprotein. WC=Waist circumference. NCEP-ATP III=National Cholesterol Education Program, Adult Treatment Panel III. AHA=American Heart Association.

Financial stress.

Financial stress was collected through REST mailed survey during 2010–2011, using a three-item financial stress scale adapted from the Boston Longitudinal Study (BOLOS), a Boston area oversample of the Midlife in the United States (MIDUS) study (Agrigoroaei et al., 2017). This adapted scale assesses participants’ financial stress on three different constructs, including the ability to meet life necessities, the ability to pay monthly bills, and an assessment of current financial situations. These three items are on Likert-type scales: (1) “How would you rate your current financial situation?” Response options range from 0 (worst) to 10 (best). (2) “In general, would you say you (and your family living with you) have more money than you need, just enough for your needs, or not enough to meet your needs?” Response options range from 1 (money more than you need) to 3 (not enough money). (3) “How difficult is it for you (and your family) to pay your monthly bills?” Response options range from 1 (very difficult) to 4 (not at all difficult). The first and the third items were coded in reverse, so that higher scores indicate higher levels of financial stress for each construct (Agrigoroaei et al., 2017). This instrument has good internal consistency with standardized Cronbach’s alpha of 0.82 (Kuo et al., 2020).

Sleep Duration.

During each of the four waves of REST mailed surveys between 2010 and 2015, participants were asked how many hours and minutes they sleep on a typical workday and a typical nonwork day (Hagen et al., 2016). The average daily sleep duration was computed as ([5 × workday sleep] + [2 × non-workday sleep])/7. According to the National Sleep Foundation’s recommendations for middle-aged and older adults (Hirshkowitz et al., 2015), participants were categorized into optimal sleep duration (7–9 hours per night) versus suboptimal sleep duration (<7 or >9 hours per night).

Physical Activity.

Participants’ light, moderate, and vigorous physical activities, were assessed at each of the four waves of REST mailed surveys during 2010–2015 using the modified version of the Paffenbarger Physical Activity Questionnaire (Paffenbarger et al., 1995). On each mailed survey, participants were asked to recall the frequency, intensity, duration, and types of physical activities they had actively participated during the past year, including walking, regular physical activities, and recreational physical activities (Mesas et al., 2018). The weekly energy expenditure on light, moderate, and vigorous physical activities was transformed into weekly metabolic equivalent of tasks (MET-minutes/week). Weekly energy expenditures were further divided into two categories: (1) met the weekly physical activity recommendation (≥500 MET-minutes/week), (2) did not meet the weekly physical activity recommendation (<500 MET-minute/week). This cutoff point is chosen based on current physical activity recommendation for middle-aged and older adults: 150 minutes of moderate intensity of physical activity per week, which is equivalent to a value of 500 MET-minutes/week (150*3.33=500) (Kaminsky & Montoye, 2014; Mesas et al., 2018). Physical inactivity was defined as those who did not meet current physical activity recommendation.

Covariates.

Demographics and socioeconomic statuses were obtained from REST mailed surveys, including age, sex, work status, education, occupational types, and subjective financial control. Occupational types were categorized into four categories based on self-reported job titles collected from REST mailed surveys or payroll records as they enrolled in the WSC: (1) management/supervisor, (2) human service, (3) sales/office, and (4) blue-collar or skilled labor. This categorial scheme for occupational types has been used in the Multi-Ethnic Study of Atherosclerosis (MESA) study with concurrent validity (Fujishiro et al., 2011; Landsbergis et al., 2015). For those who were retired, we asked for their job titles before they stopped working. For homemakers, we obtained their job titles based on their payroll records when they enrolled in the WSC. To quantify the perceived control over finance, we used a one-item question to ask participants’ subjective financial control on a Liker-type scale, ranging from 1 (none) to 10 (very much) (Johnson & Krueger, 2006). Health statuses were obtained from WSC in-laboratory assessments to control for potential confounders, including number of comorbidities and depressive symptoms. Comorbidities included myocardial infarction, heart failure, cancer, Parkinson’s disease, Alzheimer’s disease, chronic obstructive pulmonary disease, glaucoma, restless leg syndrome, emphysema, and kidney disease (Plante et al., 2017). Depressive symptoms were assessed using the Center for Epidemiologic Studies-Depression (CES-D) Scale (Lewinsohn et al., 1997). As shown in Figure 2, the differences in data collection time between exposure variable (financial stress) and outcome variable (metabolic syndrome) were transformed into months and included in the moderation and mediation models as a covariate.

Statistical Analyses

Sample characteristics were summarized using descriptive statistics. Statistical analyses were performed using SAS software version 9.4 (SAS Institute Inc., 2020) and Mplus software version 8 (Muthén & Muthén, 2017). All reported p-values were two-tailed with p-values less than 0.05 considered as significant. Missing data in covariates were imputed using multiple imputations, under the assumption that missing data in the covariates (including education, work status, occupational types, and depressive symptoms) were unrelated to sleep duration, physical activity, and metabolic syndrome. Multiple imputations were performed using Monte Carlo Markov Chain (MCMC) method based on 100 MCMC iterations with 5 imputed datasets (Asparouhov & Muthén, 2010).

Baseline Moderation Analysis.

To examine whether and to what extent meeting sleep and physical activity recommendations modify the association between financial stress and the prevalence of metabolic syndrome (Aim 1), we treated sleep duration and physical activity as binary variables based on current sleep and physical activity recommendations and categorized the participants into four subgroups based on which recommendations they met: (1) met zero recommendation, (2) met sleep recommendation only, (3) met physical activity recommendation only, and (4) met both sleep and physical activity recommendations. Structural equation modeling with subgroup analysis was used to estimate the association between financial stress and metabolic syndrome across the four recommendation-type subgroups (Muthén & Muthén, 2015). Probit model with weighted least squares mean and variance (WLSMV) was used to model binary outcome variable (Li, 2016). The differences in probit coefficients among subgroups (Δβ) were tested for statistical significance (two-tailed p-values <0.05) using model constraints. Probit coefficients were further transformed into odds ratios (ORs) to quantify the magnitude of the association between financial stress and the prevalence of metabolic syndrome across the recommendation-type subgroups.

Prospective Mediation Analysis.

To examine whether weekly physical activity and sleep duration per night partially mediate the association between financial stress and the prevalence of metabolic syndrome (Aim 2), we used Mackinnon’s asymmetric confidence intervals (CI) to examine the indirect effect (Mackinnon et al., 2004). Following Mackinnon’s mediation approach, the direct, indirect, and total effects were estimated using bias-corrected bootstrap CI (bootstrap=5000); indirect effects were considered significant if the 95% confidence intervals do not contain 0 (Zhao et al., 2010). Weekly physical activity was log-transformed to reduce the skewness and normalize the distribution based on previous recommendation (Beyler, 2010; Bruneau Jr et al., 2018). Again, probit model with WLSMV was used due to the binary outcome of metabolic syndrome. To ensure the estimated model is reasonably consistent with the current sample, we used McKelvey and Zavoina Pseudo R2 as the goodness-of-fit index, due to the binary nature of outcome variable (McKelvey & Zavoina, 1975). An indirect effect is subjected to confounding bias when common causes of mediator and outcome are not controlled for or not measured (Loeys et al., 2015). Hence, following Loeys and colleagues’ recommendation, we performed a sensitivity analysis using Baron and Kenny’s approach to examine the direct and indirect effects and prevent potential inflation in Type 1 Error (Baron & Kenny, 1986; Loeys et al., 2015).

3|. Results

Sample characteristics

In this sample of middle-aged and older adults (N=445), the mean age was 63.8 (SD=6.9), 243 (55%) participants were men, 390 (96%) participants were White, and 212 (48%) participants had a college degree or higher. Subgroups according to physical activity and sleep recommendations are presented in Table 2. Specifically, those with higher educational attainments, working in management positions, and reporting higher levels of depressive symptoms were more likely to meet both physical activity and sleep recommendations (Table 2). As shown in Supplement 2, missing values were identified in four covariates, including work status, education, occupational types, and depressive symptoms. The diagnosis of missing patterns with Little’s test for MCAR did not reject the null hypothesis, which supported the approach of multiple imputations (Chi-square=26.76; p=0.084).

Table 2.

Baseline characteristics for the total sample and subgroups based on the type of recommendation (REC) the participants met

Total sample Subgroups according to sleep and PA recommendations (N=445)
N=445a Met zero REC a Met sleep REC a Met PA REC a Met both RECs a p-value b
(N=62) (N=82) (N=116) (N=185)
Age (mean, SD) 64 (6.9) 63 (6.6) 66 (8.1) 63 (6.4) 64 (6.5) 0.004
Sex (%) 0.002
 Male 55% 37% 46% 64% 58%
 Female 45% 63% 54% 36% 42%
Education (%) <0.001
 High school or less 19% 31% 28% 16% 13%
 Some college 33% 35% 42% 40% 23%
 College or more 48% 34% 30% 44% 64%
Work status (%) 0.402
 Working full-time 31% 39% 23% 35% 29%
 Working part-time 18% 13% 21% 18% 19%
 Not working for pay 51% 48% 56% 47% 52%
Occupational types (n, %) 0.030
 Management/supervisor 16% 8% 16% 15% 21%
 Human services 25% 26% 17% 22% 29%
 Sales/office 39% 55% 43% 40% 32%
 Blue-collar and skilled labors 20% 11% 24% 23% 18%
Perceived financial control
 (mean, SD)
 [range 0–10]		 6.3
(2.5)		 5.8
(2.5)		 5.9
(2.8)		 6.2
(2.4)		 6.6
(2.5)		 0.051
Depressive symptoms [CES-D]
(mean, SD) [range 0–53]		 7.9
(9.3)		 12.4
(11.2)		 7.3
(8.8)		 8.1
(9.0)		 6.4
(8.5)		 <0.001
Number of comorbidities c (%) 0.011
 None 64% 68% 54% 71% 64%
 One 29% 27% 32% 28% 28%
 Two or more 7% 5% 15% 1% 8%
Open in a new tab
a

The descriptive results for subgroups are based on the imputed data.

b

Chi-square test was used for categorical variables. The Friedman’s non-parametric test was used for non-normally distributed data (i.e., perceived financial control and depressive symptoms). One-way ANOVA was used for normally distributed data (i.e., age). P-values less than 0.05 were considered significant.

c

Comorbidities included myocardial infarction, heart failure, cancer, Parkinson’s disease, Alzheimer’s disease, chronic obstructive pulmonary disease, glaucoma, restless leg syndrome, emphysema, and kidney disease.

Abbreviations. SD=standard deviation; PA=physical activity; REC=recommendation

Financial Stress Is Associated with Suboptimal Sleep Duration and Physical Inactivity

As shown in Figure 3, higher levels of financial stress were associated with higher odds of physical inactivity (<500 MET-minutes/week) and suboptimal sleep (<7 hours or >9 hours/night). In the unadjusted logistic regression model, a one-unit increase in financial stress score was associated with 52% higher odds of suboptimal sleep duration (OR=1.52; 95% CI=[1.23, 1.88]; p<0.001) and 48% higher odds of physical inactivity (OR=1.48; 95% CI=[1.19, 1.83]; p<0.001). After the adjustment of demographics and socioeconomic statuses (including age, sex, work status, education, occupational types, and subjective financial control) and health statuses (including depressive symptoms and number of comorbidities), a one-unit increase in financial stress score was related to 58% higher odds of suboptimal sleep (OR=1.58, 95% CI=[1.20, 2.09]; p=0.001) and 40% higher odds of physical inactivity (OR= 1.40; 95% CI=[1.05, 1.86]; p=0.023).

Figure 3.

Figure 3.

Open in a new tab

Estimated probabilities of suboptimal sleep duration and physical inactivity as the function of financial stress based on the unadjusted logistic regression model with 95% confidence intervals. Note. Physical inactivity is defined by the self-report of less than 150 minutes of moderate physical activities per week (<500 MET-minutes/week); suboptimal sleep duration is defined by self-report of inadequate sleep (<7 hours/night) or excessive sleep (>9 hours/night). Financial stress is the composite score of three items, including difficulty paying bills, difficulty paying life necessities, and current financial strain. Higher score indicates higher level of financial stress.

Behavioral Modification: Meeting Both Physical Activity and Sleep Recommendations Attenuates the Association between Financial Stress and Metabolic Syndrome

Participants were divided into four groups based on the number of recommendations each participant met. As shown in Table 3, after the adjustment of age, sex, work status, education, occupational types, subjective financial control, depressive symptoms, and number of comorbidities, financial stress was significantly associated with the prevalence of metabolic syndrome among those who met zero recommendations (OR=5.40; 95% CI=[1.55, 18.81]; p=0.008). Figure 4 illustrates the significant differences in probit coefficients (Δβ) among subgroups after the adjustment of covariates using model constraints. As shown in Figure 4, interaction effect was observed between those who met zero recommendation versus those who met both recommendations (Figure 4).

Table 3.

The association between financial stress and the prevalence of metabolic syndrome stratified by the type of recommendation the participants met

Model 1 a
 Model 2 b
OR 95% CI p-value Interaction (Δβ) OR 95% CI p-value Interaction (Δβ)
Group 1: Met zero recommendation (N=62) 3.10 [1.31, 7.33] 0.010 Ref 5.40 [1.55, 18.81] 0.008 Ref
Group 2: Met sleep recommendation only (N=82) 1.41 [0.90, 2.22] 0.138 −0.46 (p=0.111) 1.56 [0.46, 5.24] 0.472 −0.73 (0.161)
Group 3: Met PA recommendation only (N=116) 1.06 [0.68, 1.65] 0.797 −0.63 (p=0.029) 1.70 [0.77, 3.73] 0.186 −0.68 (p=0.124)
Group4: Met both sleep and PA recommendations (N=185) 1.18 [0.84, 1.67] 0.336 −0.57 (p=0.041) 1.33 [0.84, 2.10] 0.221 −0.83 (p=0.039)
Open in a new tab
a

Model 1: Probit regression model without the adjustment of covariates (Crude Model).

b

Model 2: Adjusted for demographics and socioeconomic variables (age, sex, work status, education, types of occupation, and subjective financial control), health statuses (depressive symptoms and numbers of comorbidities), and time differences between REST mailed surveys and WSC in-laboratory assessment.

Abbreviations. PA=physical activity; OR=odds ratio; 95%CI=95% confidence interval

Figure 4.

Figure 4.

Open in a new tab

Odds ratios of metabolic syndrome as a function of financial stress stratified by meeting sleep and physical activity recommendations with interaction terms (Δβ). Notes: The levels of histogram refer to odds ratio; the whiskers on the top of each histogram refer to 95% confidence intervals; the bold font indicates that the p-value is less than 0.05; the probit models were adjusted for age, sex, education, work status, occupational types, subjective financial control, depressive symptoms, number of comorbidities, and the time differences between REST mailed surveys and WSC in-laboratory assessment. Abbreviations. PA=physical activity. REC=Recommendation.

Behavioral Adaptation: Weekly Physical Activity Partially Mediates the Association between Financial Stress and Metabolic Syndrome

Following Mackinnon’s approach with bias-corrected bootstrap confidence intervals, weekly physical activity mediated the relationship between financial stress and prevalence of metabolic syndrome with significant indirect effect (β=0.056, 95% CI=[0.005, 0.168), but sleep duration did not mediate this relationship (β=−0.004, 95% CI=[−0.048, 0.011]). Table 4 describes the mediation results with bias-corrected bootstrap confidence intervals. All the mediation models were adjusted for age, sex, work status, education, occupational types, subjective financial control, depressive symptoms, and number of comorbidities.

Table 4.

Mediation analysis of weekly physical activity and sleep duration on the relationship between financial stress and prevalence of metabolic syndrome based on bias-corrected bootstrap confidence intervals

Pathways β SE P value
FS→MetS 0.205 0.071 0.004
Log PA→MetS −0.186 0.080 0.020
Sleep duration→MetS 0.056 0.054 0.302
FS→Log PA −0.138 0.062 0.026
FS→Sleep duration −0.029 0.065 0.654
Indirect effect Effect SE Bias-corrected bootstrap 95% CI
FS→Log PA→MetS 0.056 0.040 [0.005, 0.168]
FS→Sleep duration→MetS −0.004 0.012 [−0.048, 0.011]
Open in a new tab

Abbreviations. FS=financial stress; MetS=metabolic syndrome; Log PA=log-transformed physical activity; 95% CI=95% confidence interval.

Note. Age, sex, education, work status, types of occupation, subjective financial control, depressive symptoms, number of comorbidities, and time differences between REST mailed surveys and WSC in-laboratory assessment were controlled as covariates.

Sensitivity analysis was performed using Baron and Kenny’s mediation approach. Financial stress was associated with the prevalence of metabolic syndrome (Figure 5 [A]). Weekly physical activity partially mediated the relationship between financial stress and the prevalence of metabolic syndrome with a significant indirect effect (OR=1.05; 95%CI= [1.01, 1.10]; p=0.030). Yet, the indirect effect through sleep duration was not detected (p=0.643). The Pseudo R2 suggested that the mediational model provides slightly better model fit (Figure 5 [B], Pseudo R2=0.21; p<0.001), than the model without mediators (Figure 5 [A], Pseudo R2=0.18; p<0.001). Supplement 3 describes the mediation results based on Baron and Kenny’s mediation analysis. Both Mackinnon’s and Baron and Kenny’s mediation approaches yielded similar mediation results in current study.

Figure 5.

Figure 5.

Open in a new tab

Mediation model following Baron and Kenny’s approach, including total effect (c), indirect effect (ab), and direct effect (c’). (A) The top panel portrays the total relation between financial stress and prevalence of metabolic syndrome (B) The bottom panel portrays the indirect effect of financial stress on the prevalence of metabolic syndrome through the mediator (physical activity). Abbreviations & symbols: FS=financial stress; MetS=metabolic syndrome; Log PA=log-transformed physical activity; β=probit coefficient. Note. Age, sex, education, work status, occupational types, subjective financial control, depressive symptoms, number of comorbidities, and time differences between REST mailed surveys and WSC in-laboratory assessment were controlled as covariates. The asterisk (*) indicates that the p-value is less than 0.05 (*<0.05; **<0.01; ***<0.001).

4|. Discussion

Despite the growing evidence demonstrating high risk of metabolic syndrome in financially disadvantages populations (Arnett et al., 2019; St-Onge et al., 2016), there is a lack of evidence examining the behavioral modification and adaptation in the association between financial stress and metabolic syndrome among middle-aged and older adults. In this study, we demonstrated that meeting both physical activity and sleep recommendations dissociates the link between financial stress and metabolic syndrome. We further found that weekly physical activity partially mediated the association between financial stress and the prevalence of metabolic syndrome. Results from this study highlight the complex behavioral mechanisms embedded in the metabolic health disparities driven by financial stress.

We found a significant indirect effect of physical activity in the association between financial stress and the prevalence of metabolic syndrome, but the indirect effect through sleep duration was not detected (Figure 5). Our findings are consistent with previous mediation study conducted in the general population (18–79 years of age), where Montano found that occupational types, income, and educational attainment are the major social determinants of metabolic syndrome, and these associations are partially mediated by physical activity. However, Montano did not examine the role of sleep duration (Montano, 2017). Although more studies are needed to pinpoint the underlying mechanisms, several hypotheses have been proposed to explain the behavioral adaption of physical inactivity under financial stress: (1) switch from goal-directed system to habit-responding system, (2) development of scarcity mindset, and (3) environmental injustice. First, behavioral adaptation tends to occur under stressful contexts, during which habit-responding system often surpasses the goal-directed system (Meier et al., 2022; Schwabe & Wolf, 2009). Secondly, long-term exposure to financial stress might alter adults’ perception about what they need most and what they should prioritize, which compromises their intrinsic motivations toward the enjoyment and accomplishment in physical activities (Shah et al., 2015). Under the assumption of scarcity mindset, adults with financial stress might place their unmet financial needs as top priority, instead of physical fitness and personal health (De Bruijn & Antonides, 2021). Third, mounting evidence has shown that safety concerns arising from unfavorable built environments (including unsafe surroundings, no access to sidewalks, and lack of winter recreation centers) could counteract both intrinsic and extrinsic motivations toward regular physical activities (Suminski et al., 2005). Future studies are needed to test whether person-centered interventions assisted with Just-in-Time Adaptive Interventions (JITAIs) (e.g., situational nudge) could stabilize the behavioral variations that are conditional on the execution context and stressful environment and examine the effectiveness to break the cycle of habit formation and behavioral maladaptation (Forberger et al., 2019; Maher et al., 2021).

As shown in Table 3, financial stress was significantly associated with the prevalence of metabolic syndrome among those who met zero recommendations, with significant interaction effect identified between those who met both recommendations versus those who met zero recommendation (Δβ=−0.83, p=0.039). Our findings highlight the synergistic benefits of meeting both sleep and physical recommendation in modifying the risk of metabolic syndrome among financially stressed adults. From an energy expenditure perspective, adults typically spend each 24-hour period on three types of activity: (1) sleep, (2) sedentary activities, and (3) physical activities (Rosenberger et al., 2019). Promoting optimal sleep duration and physical activity could simultaneously decrease individuals’ time spent in sedentary activities (Foti et al., 2011). Since 2019, the American College of Cardiology and American Heart Association (ACC/AHA) have recommended routinely assessing and optimizing patients’ sleep and physical activity during healthcare visits, given the joint benefits of optimal sleep and physical activity in the primary prevention of atherosclerotic cardiovascular disease (ASCVD) (Arnett et al., 2019; Kuo et al., 2020).

Although we observed the synergistic benefits of meeting both physical activity and sleep recommendation through the significant interaction effect (Figure 3), we are not able to conclude whether meeting sleep recommendation alone is sufficient to modify the effect of financial stress on the prevalence of metabolic syndrome, due to the lack of significant interaction effect. However, in the crude model (Table 3), we did identify significant interaction effect between those who met physical activity recommendation only versus who met zero recommendation (Δβ=−0.63, p=0.029). In other words, meeting physical activity recommendation alone might dissociate the link between financial stress and metabolic syndrome, but the moderation effect could be attenuated by demographics, socioeconomic statuses, and health statuses. Finally, detecting moderation effects based on simple subgroup analysis is risky, because the heterogeneity of effect sizes observed across subgroups might be due to small sample sizes. To overcome this challenge, we examined the interaction effects with model constraints based on Wang and Ware’s recommendations (Stride et al., 2015; Wang & Ware, 2013). Nevertheless, a representative national sample with a larger sample size is warranted to further confirm the generalizability of our findings in middle-aged and older Americans.

In the current study, 60% of the participants met sleep recommendation, which is similar to recent populational studies with prevalence rates between 64% to 67% (Sabia et al., 2021; Winer et al., 2021). On the contrary, the prevalence rate of meeting physical activity recommendation in current sample is 67.6%, which is higher than the general older population in the U.S. (55.2%), according to the 2019 national surveillance system (Behavioral Risk Factor Surveillance System, 2019). Such difference could be explained by the variations in physical activity measurements and sampling frames. Particularly, physical activity in the current study is assessed by Paffenbarger Physical Activity Questionnaire, which counts every minute of moderate to vigorous physical activities (MVPA), while the physical activity question in the BRFSS counts MVPA in bouts of at least 10 min in duration (Zenko et al., 2019). In terms of sampling frame, previous reports of WSC study indicated that WSC participants had a slightly healthier profile and lower death rate, compared to the entire sampling frame of the Wisconsin (Young et al., 2008).

Findings from this study must be interpreted with respect to the limitations. First, causal inferences and recursive relationships between the financial stress and metabolic syndrome cannot be concluded because the outcome focuses on the prevalence of metabolic syndrome rather than the cumulative incidence of metabolic syndrome. To accommodate this limitation, we ensured a temporal sequential order between the exposure and outcome variables in conjunction with a prospective design (Figure 2). Secondly, genetic predisposition, dietary patterns, and the availability of healthy food choices play important roles in the metabolic health disparities (Montano, 2017), but these potential confounders were not adjusted for due to the lack of data availability. Being aware of these unmeasured confounding effects could help improve future research to predict behavioral and metabolic responses to financial stress. Third, given the status of transition to retirement in this sample and systematic bias in self-reported annual income identified in previous literature (Fukuoka et al., 2007), we did not include income as a covariate. To overcome this weakness, we adjusted for the educational attainment, self-reported financial control, and type of occupation as the proxy indicators for socioeconomic statuses. Finally, the current sample is based on the WSC study, which has limited racial diversity (96% of participants were white) (Peppard et al., 2000). Thus, caution should be taken when generalizing the results to other racial/ethnic backgrounds.

Conclusion

In this sample of middle-aged and older adults, we found that meeting current recommendations of sleep duration and physical activity disassociates financial stress from the risk of metabolic syndrome. Yet, we also found that those with financial stress were less likely to have optimal sleep duration and physical activity, which highlights current challenges in promoting sleep health and physical activity in financially stressed adults. Future interventions addressing metabolic health disparities might achieve better outcomes if clinicians and researchers factor in the behavioral adaptation of physical inactivity under financial stress.

Supplementary Material

Supplement 1-3

References

  1. Agrigoroaei S, Lee-Attardo A, & Lachman ME (2017). Stress and subjective age: Those with greater financial stress look older. Research on Aging, 39(10), 1075–1099. 10.1177/0164027516658502 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Aguilar M, Bhuket T, Torres S, Liu B, & Wong RJ (2015). Prevalence of the metabolic syndrome in the United States, 2003–2012. JAMA, 313(19), 1973–1974. 10.1001/jama.2015.4260 [DOI] [PubMed] [Google Scholar]
  3. Arnett DK, Blumenthal RS, Albert MA, Buroker AB, Goldberger ZD, Hahn EJ, Himmelfarb CD, Khera A, Lloyd-Jones D, McEvoy JW, Michos ED, Miedema MD, Munoz D, Smith SCJ, Virani SS, Williams KAS, Yeboah J, & Ziaeian B (2019). 2019 ACC/AHA Guideline on the Primary Prevention of Cardiovascular Disease: A Report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines. Circulation, 140(11), e596–e646. 10.1161/CIR.0000000000000678 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Asparouhov T, & Muthén B (2010). Multiple imputation with Mplus. MPlus Web Notes, 238–246. [Google Scholar]
  5. Baron RM, & Kenny DA (1986). The moderator–mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51(6), 1173. [DOI] [PubMed] [Google Scholar]
  6. Beenackers MA, Oude Groeniger J, van Lenthe FJ, & Kamphuis CB (2018). The role of financial strain and self-control in explaining health behaviours: The GLOBE study. The European Journal of Public Health, 28(4), 597–603. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Behavioral Risk Factor Surveillance System. (2019). Nutrition, Physical Activity, and Obesity: Data, Trends and Maps. https://nccd.cdc.gov/dnpao_dtm/rdPage.aspx?rdReport=DNPAO_DTM.ExploreByLocation&rdRequestForwarding=Form
  8. Beyler NK (2010). Statistical methods for analyzing physical activity data. Iowa State University. [Google Scholar]
  9. Bruneau M Jr, Walsh S, Selinsky E, Ash G, Angelopoulos TJ, Clarkson P, Gordon P, Moyna N, Visich P, & Zoeller R (2018). A genetic variant in IL‐15Rα correlates with physical activity among European–American adults. Molecular Genetics & Genomic Medicine, 6(3), 401–408. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Caceres BA, Ancheta AJ, Dorsen C, Newlin-Lew K, Edmondson D, & Hughes TL (2022). A population-based study of the intersection of sexual identity and race/ethnicity on physiological risk factors for CVD among U.S. adults (ages 18–59). Ethnicity & Health, 27(3), 617–638. 10.1080/13557858.2020.1740174 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Commodore-Mensah Y, Turkson-Ocran R-A, Foti K, Cooper LA, & Himmelfarb CD (2021). Associations between social determinants and hypertension, stage 2 hypertension, and controlled blood pressure among men and women in the United States. American Journal of Hypertension, 34(7), 707–717. 10.1093/ajh/hpab011 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Crowe J, Lacy C, & Columbus Y (2018). Barriers to food security and community stress in an urban food desert. Urban Science, 2(2), 46. 10.3390/urbansci2020046 [DOI] [Google Scholar]
  13. De Bruijn E-J, & Antonides G (2021). Poverty and economic decision making: A review of scarcity theory. Theory and Decision, 1–33. [Google Scholar]
  14. D’Lima GM, Pearson MR, & Kelley ML (2012). Protective behavioral strategies as a mediator and moderator of the relationship between self-regulation and alcohol-related consequences in first-year college students. Psychology of Addictive Behaviors: Journal of the Society of Psychologists in Addictive Behaviors, 26(2), 330–337. 10.1037/a0026942 [DOI] [PubMed] [Google Scholar]
  15. Du C, Adjepong M, Zan MCH, Cho MJ, Fenton JI, Hsiao PY, Keaver L, Lee H, Ludy M-J, Shen W, Swee WCS, Thrivikraman J, Amoah-Agyei F, de Kanter E, Wang W, & Tucker RM (2022). Gender differences in the relationships between perceived stress, eating behaviors, sleep, dietary risk, and body mass index. Nutrients, 14(5). 10.3390/nu14051045 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Easterlin RA (2000). The worldwide standard of living since 1800. Journal of Economic Perspectives, 14(1), 7–26. [Google Scholar]
  17. Ensel WM, & Lin N (2004). Physical fitness and the stress process. Journal of Community Psychology, 32(1), 81–101. [Google Scholar]
  18. Epel ES, Crosswell AD, Mayer SE, Prather AA, Slavich GM, Puterman E, & Mendes WB (2018). More than a feeling: A unified view of stress measurement for population science. Frontiers in Neuroendocrinology, 49, 146–169. 10.1016/j.yfrne.2018.03.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Forberger S, Reisch L, Kampfmann T, & Zeeb H (2019). Nudging to move: A scoping review of the use of choice architecture interventions to promote physical activity in the general population. The International Journal of Behavioral Nutrition and Physical Activity, 16(1), 77. 10.1186/s12966-019-0844-z [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Foti KE, Eaton DK, Lowry R, & McKnight-Ely LR (2011). Sufficient sleep, physical activity, and sedentary behaviors. American Journal of Preventive Medicine, 41(6), 596–602. 10.1016/j.amepre.2011.08.009 [DOI] [PubMed] [Google Scholar]
  21. Fujishiro K, Diez Roux AV, Landsbergis P, Baron S, Barr RG, Kaufman JD, Polak JF, & Stukovsky KH (2011). Associations of occupation, job control and job demands with intima-media thickness: The Multi-Ethnic Study of Atherosclerosis (MESA). Occupational and Environmental Medicine, 68(5), 319–326. 10.1136/oem.2010.055582 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Fukuoka Y, Rankin SH, & Carroll DL (2007). Systematic bias in self-reported annual household incomes among unpartnered elderly cardiac patients. Applied Nursing Research: ANR, 20(4), 205–209. 10.1016/j.apnr.2007.01.010 [DOI] [PubMed] [Google Scholar]
  23. Hagen EW, Barnet JH, Hale L, & Peppard PE (2016). Changes in sleep duration and sleep timing associated with retirement transitions. Sleep, 39(3), 665–673. 10.5665/sleep.5548 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Hamer M (2012). Psychosocial stress and cardiovascular disease risk: The role of physical activity. Psychosomatic Medicine, 74(9), 896–903. 10.1097/PSY.0b013e31827457f4 [DOI] [PubMed] [Google Scholar]
  25. Hirshkowitz M, Whiton K, Albert SM, Alessi C, Bruni O, DonCarlos L, Hazen N, Herman J, Katz ES, Kheirandish-Gozal L, Neubauer DN, O’Donnell AE, Ohayon M, Peever J, Rawding R, Sachdeva RC, Setters B, Vitiello MV, Ware JC, & Adams Hillard PJ (2015). National Sleep Foundation’s sleep time duration recommendations: Methodology and results summary. Sleep Health, 1(1), 40–43. 10.1016/j.sleh.2014.12.010 [DOI] [PubMed] [Google Scholar]
  26. Hostinar CE, Ross KM, Chen E, & Miller GE (2017). Early-life socioeconomic disadvantage and metabolic health disparities. Psychosomatic Medicine, 79(5), 514–523. 10.1097/PSY.0000000000000455 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Johnson W, & Krueger RF (2006). How money buys happiness: Genetic and environmental processes linking finances and life satisfaction. Journal of Personality and Social Psychology, 90(4), 680. [DOI] [PubMed] [Google Scholar]
  28. Kaminsky LA, & Montoye AH (2014). Physical activity and health: What is the best dose? [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Karazsia BT, & Berlin KS (2018). Can a mediator moderate? Considering the role of time and change in the mediator-moderator distinction. Behavior Therapy, 49(1), 12–20. [DOI] [PubMed] [Google Scholar]
  30. Kuo W-C, Oakley LD, Brown RL, Hagen EW, Barnet JH, Peppard PE, & Bratzke LC (2021). Gender differences in the relationship between financial stress and metabolic abnormalities. Nursing Research, 70(2), 123–131. 10.1097/NNR.0000000000000489 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Kuo W-C, Stevens JM, Ersig AL, Johnson HM, Tung T-H, & Bratzke LC (2020). Does 24-h activity cycle influence plasma PCSK9 concentration? A systematic review and meta-analysis. Current Atherosclerosis Reports, 22(7), 30. 10.1007/s11883-020-00843-x [DOI] [PubMed] [Google Scholar]
  32. Landsbergis PA, Diez-Roux AV, Fujishiro K, Baron S, Kaufman JD, Meyer JD, Koutsouras G, Shimbo D, Shrager S, Stukovsky KH, & Szklo M (2015). Job strain, occupational category, systolic blood pressure, and hypertension prevalence: The multi-ethnic study of atherosclerosis. Journal of Occupational and Environmental Medicine, 57(11), 1178–1184. PsycINFO. 10.1097/JOM.0000000000000533 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Lewinsohn PM, Seeley JR, Roberts RE, & Allen NB (1997). Center for Epidemiologic Studies Depression Scale (CES-D) as a screening instrument for depression among community-residing older adults. Psychology and Aging, 12(2), 277. [DOI] [PubMed] [Google Scholar]
  34. Li C-H (2016). Confirmatory factor analysis with ordinal data: Comparing robust maximum likelihood and diagonally weighted least squares. Behavior Research Methods, 48(3), 936–949. 10.3758/s13428-015-0619-7 [DOI] [PubMed] [Google Scholar]
  35. Loucks EB, Gilman SE, Britton WB, Gutman R, Eaton CB, & Buka SL (2016). Associations of mindfulness with glucose regulation and diabetes. American Journal of Health Behavior, 40(2), 258–267. PsycINFO. 10.5993/AJHB.40.2.11 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. MacKinnon DP (2011). Integrating mediators and moderators in research design. Research on Social Work Practice, 21(6), 675–681. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Mackinnon DP, Lockwood CM, & Williams J (2004). Confidence limits for the indirect effect: Distribution of the product and resampling methods. Multivariate Behavioral Research, 39(1), 99–99. PubMed. 10.1207/s15327906mbr3901_4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Maher JP, Rebar AL, & Dunton GF (2021). The influence of context stability on physical activity and sedentary behaviour habit and behaviour: An ecological momentary assessment study. British Journal of Health Psychology, 26(3), 861–881. 10.1111/bjhp.12509 [DOI] [PubMed] [Google Scholar]
  39. McKelvey RD, & Zavoina W (1975). A statistical model for the analysis of ordinal level dependent variables. Journal of Mathematical Sociology, 4(1), 103–120. [Google Scholar]
  40. Meier JK, Staresina BP, & Schwabe L (2022). Stress diminishes outcome but enhances response representations during instrumental learning. Elife, 11, e67517. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Mesas AE, Hagen EW, & Peppard PE (2018). The bidirectional association between physical activity and sleep in middle-aged and older adults: A prospective study based on polysomnography. Sleep, 41(9). 10.1093/sleep/zsy114 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Montano D (2017). Association between socioeconomic determinants and the metabolic syndrome in the German Health Interview and Examination Survey for Adults (DEGS1)—A mediation analysis. The Review of Diabetic Studies: RDS, 14(2–3), 279–294. 10.1900/RDS.2017.14.279 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Moore JX, Chaudhary N, & Akinyemiju T (2017). Metabolic syndrome prevalence by race/ethnicity and sex in the United States, National Health and Nutrition Examination Survey, 1988–2012. Preventing Chronic Disease, 14, E24. 10.5888/pcd14.160287 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Muthén LK, & Muthén BO (2015). Ch14: Special Modeling Issues. In Mplus User’s Guide (Seventh Edition). https://www.statmodel.com/download/usersguide/MplusUserGuideVer_7.pdf [Google Scholar]
  45. Muthén LK, & Muthén BO (2017). Mplus Statistical Analysis With Latent Variables. Los Angeles, CA: Muthén & Muthén, Version 8. https://www.statmodel.com/download/usersguide/MplusUserGuideVer_8.pdf [Google Scholar]
  46. Paffenbarger RSJ, Wing AL, & Hyde RT (1995). Physical activity as an index of heart attack risk in college alumni. 1978. American Journal of Epidemiology, 142(9), 889–903; discussion 887–888. [DOI] [PubMed] [Google Scholar]
  47. Peppard PE, Young T, Palta M, & Skatrud J (2000). Prospective study of the association between sleep-disordered breathing and hypertension. The New England Journal of Medicine, 342(19), 1378–1384. 10.1056/NEJM200005113421901 [DOI] [PubMed] [Google Scholar]
  48. Plante DT, Finn LA, Hagen EW, Mignot E, & Peppard PE (2017). Longitudinal associations of hypersomnolence and depression in the Wisconsin Sleep Cohort Study. Journal of Affective Disorders, 207, 197–202. 10.1016/j.jad.2016.08.039 [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Puterman E, Adler N, Matthews KA, & Epel E. (2012). Financial strain and impaired fasting glucose: The moderating role of physical activity in the coronary artery risk development in young adults study. Psychosomatic Medicine, 74(2), 187–192. rzh. [DOI] [PubMed] [Google Scholar]
  50. Puterman E, Lin J, Blackburn E, O’Donovan A, Adler N, & Epel E (2010). The power of exercise: Buffering the effect of chronic stress on telomere length. PloS One, 5(5), e10837. 10.1371/journal.pone.0010837 [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Puterman E, Prather AA, Epel ES, Loharuka S, Adler NE, Laraia B, & Tomiyama AJ (2016). Exercise mitigates cumulative associations between stress and BMI in girls age 10 to 19. Health Psychology: Official Journal of the Division of Health Psychology, American Psychological Association, 35(2), 191–194. 10.1037/hea0000258 [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Rao ND, & Min J (2018). Decent living standards: Material prerequisites for human wellbeing. Social Indicators Research, 138(1), 225–244. 10.1007/s11205-017-1650-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Rosenberger ME, Fulton JE, Buman MP, Troiano RP, Grandner MA, Buchner DM, & Haskell WL (2019). The 24-hour activity cycle: A new paradigm for physical activity. Medicine and Science in Sports and Exercise, 51(3), 454–464. 10.1249/MSS.0000000000001811 [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Sabia S, Fayosse A, Dumurgier J, van Hees VT, Paquet C, Sommerlad A, Kivimäki M, Dugravot A, & Singh-Manoux A (2021). Association of sleep duration in middle and old age with incidence of dementia. Nature Communications, 12(1), 2289. 10.1038/s41467-021-22354-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. SAS Institute Inc. (2020). SAS/IML®9.4 Programmer’s Bookshelf. SAS Institute Inc. [Google Scholar]
  56. Schilling R, Colledge F, Pühse U, & Gerber M (2020). Stress-buffering effects of physical activity and cardiorespiratory fitness on metabolic syndrome: A prospective study in police officers. PloS One, 15(7), e0236526. 10.1371/journal.pone.0236526 [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Schwabe L, & Wolf OT (2009). Stress prompts habit behavior in humans. Journal of Neuroscience, 29(22), 7191–7198. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Shah AK, Shafir E, & Mullainathan S (2015). Scarcity frames value. Psychological Science, 26(4), 402–412. 10.1177/0956797614563958 [DOI] [PubMed] [Google Scholar]
  59. Steen A, & MacKenzie D (2013). Financial stress, financial literacy, counselling and the risk of homelessness. Australasian Accounting, Business and Finance Journal, 7(3), 31–48. [Google Scholar]
  60. St-Onge M-P, Grandner MA, Brown D, Conroy MB, Jean-Louis G, Coons M, & Bhatt DL (2016). Sleep duration and quality: Impact on lifestyle behaviors and cardiometabolic health: A scientific statement from the American Heart Association. Circulation, 134(18), e367–e386. 10.1161/CIR.0000000000000444 [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Stride C, Gardner S, Catley N, & Thomas F (2015). Mplus code for mediation, moderation, and moderated mediation models. [Google Scholar]
  62. Suminski RR, Poston WSC, Petosa RL, Stevens E, & Katzenmoyer LM (2005). Features of the neighborhood environment and walking by US adults. American Journal of Preventive Medicine, 28(2), 149–155. [DOI] [PubMed] [Google Scholar]
  63. Umberson D, Liu H, & Reczek C (2008). Stress and health behaviour over the life course. Advances in Life Course Research, 13, 19–44. [Google Scholar]
  64. Wang R, & Ware JH (2013). Detecting moderator effects using subgroup analyses. Prevention Science: The Official Journal of the Society for Prevention Research, 14(2), 111–120. 10.1007/s11121-011-0221-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  65. Winer JR, Deters KD, Kennedy G, Jin M, Goldstein-Piekarski A, Poston KL, & Mormino EC (2021). Association of short and long sleep duration with amyloid-β burden and cognition in aging. JAMA Neurology, 78(10), 1187–1196. 10.1001/jamaneurol.2021.2876 [DOI] [PMC free article] [PubMed] [Google Scholar]
  66. Wood W, Mazar A, & Neal DT (2022). Habits and goals in human behavior: Separate but interacting systems. Perspectives on Psychological Science, 17(2), 590–605. 10.1177/1745691621994226 [DOI] [PubMed] [Google Scholar]
  67. Young T, Finn L, Peppard PE, Szklo-Coxe M, Austin D, Nieto FJ, Stubbs R, & Hla KM (2008). Sleep disordered breathing and mortality: Eighteen-year follow-up of the Wisconsin sleep cohort. Sleep, 31(8), 1071–1078. [PMC free article] [PubMed] [Google Scholar]
  68. Zenko Z, Willis EA, & White DA (2019). Proportion of adults meeting the 2018 physical activity guidelines for Americans according to accelerometers. Frontiers in Public Health, 7, 135. 10.3389/fpubh.2019.00135 [DOI] [PMC free article] [PubMed] [Google Scholar]
  69. Zhao X, Lynch JG Jr, & Chen Q (2010). Reconsidering Baron and Kenny: Myths and truths about mediation analysis. Journal of Consumer Research, 37(2), 197–206. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplement 1-3
NIHMS2103944-supplement-Supplement_1-3.docx (24.3KB, docx)

RESOURCES