How Midlife Chronic Stress Combines with Stressful Life Events to Influence Later Life Mental and Physical Health for Husbands and Wives in Enduring Marriages

Abstract

Objectives: To investigate how midlife chronic stress (40–50 years) and subsequent acute stressful events (50–65 years) influence husbands’ and wives’ later life mental and physical health (65+ years). Methods: Dyadic structural equation modeling was used with prospective data over 25 years from 256 husbands and wives in enduring marriages. Results: For both spouses, midlife chronic stress influenced the occurrence of subsequent acute stressful life events, which in turn influenced both depressive symptoms and poor physical health in later years. For both spouses, midlife chronic stress also directly influenced depressive symptoms. Wives’ midlife chronic stress also directly influenced their own poor physical health and husbands’ depressive symptoms in later years. Discussion: These findings enhance our knowledge about the long-term joint influence of midlife chronic stress and acute stressful life events on health outcomes of husbands and wives in later years and identify dependencies between spouses, which can inform health preventive intervention programs.

Keywords

chronic stress acute stress marriage health midlife

Past research has indicated that health problems that develop in later life may be attributed in part to exposure to stressful life circumstances over the adult years (Lovallo, 2005; Wickrama et al., 2019). Previous studies have documented that both chronic and acute stress play important roles in the development of mental and physical health problems in this period (Hammen et al., 2009; Lantz et al., 2005; Lorenz et al., 2006), and that older adults are more vulnerable to the physiological and psychological effects of stress (Zaidi, 2014). However, less is known about how midlife chronic stress (40–50 years) and acute stressful life events in mid-later years (50–65 years) combine to influence health outcomes in later life (>65 years).

Chronic stress can be defined as continuing and ongoing stressful circumstances, such as conflictual marital and parental relations and persistent economic hardship; whereas, acute stress can be defined as individual stressful events that have relatively discrete beginnings and endings over a short period, such as a child dropping out of school, a child having unwanted pregnancy, loss of a home, and serious injury. There is some evidence that chronic stress, rather than acute stress should be the central focus of stress-health investigations (McGonagle & Kessler, 1990). For instance, previous research has shown that stressful situations are more likely to be attributed to a stable cause, which persist despite efforts to avoid or ameliorate them, have a stronger health effect than acute stressors and can be considered as potent predictors of health (McGonagle & Kessler, 1990; Pearlin et al., 2007). This persistent influence on health outcomes operates through psychophysiological mechanisms (Hammen et al. 2009; Pearlin et al., 2007). Conversely, the intensity of an acute event, with identifiable beginnings and endings, is presumed to peak quickly and then recede precipitously (Wheaton, 1996). Therefore, the health effect of acute stressful events may not be as persistent as the effects of chronic stress. Thus, chronic stress is such an important source of variability in the stress-health association, and it is important to assess acute stressful events in the context of chronic stress.

However, less is known about how chronic stress combines with acute stress in different ways to influence the mental and physical health of husbands and wives in the couple context, especially in the area of aging and health. This dearth of previous research is particularly important for couples in enduring (continuing or consistent) marriages who are parents, as stressful events occur within the context of persistent life conditions, such as family financial difficulties and conflictual relationships with a partner or with children (Caspi et al., 2003; Turner & Turner, 2005).

Acute Stress Mediates the Health Influence of Chronic Stress: The Mediation Hypothesis

Drawing from the life course stress trajectory model (Kuh et al., 2003), we expect that health risk follows a “sticky” trajectory over the life course. This mediation process is consistent with the stress proliferation hypothesis from the stress process model (Pearlin and Skaff, 1996) which contends that initial stressors (i.e., primary stressors) give rise to later stressors (i.e., secondary stressors) in subsequent life stages. Consistent with this argument, chronic difficulties and stressful contextual factors may contribute to the occurrence of stressful acute events in different life domains leading to health problems later in life (Pearlin et al., 2005; Thoits, 2010; Turner & Turner, 2005). For example, chronic stressors, like family economic and marital stress, may proliferate into acute subsequent stressful life events such as home foreclosure, change in residency and children’s off-time life transitions. Thus, we hypothesize that the persistent health influence of chronic stress in midlife is partly mediated by stressful life events in mid-later years (mediation hypothesis, see Hammen et al., 2009).

Both Chronic and Acute Stress Cumulatively Influence Health Outcomes: The Cumulative Hypothesis

Furthermore, according to the life course cumulative model (Kuh et al., 2003), each period of stress exposure continues to uniquely influence health outcomes later in life. That is, even when an individual is subsequently removed from the early stressor, its health effect persists. Early stress exposure may damage biological and psychological systems in persistent ways that manifest as health problems in later years (Glymour et al., 2010). Accordingly, both chronic stress and acute stress cumulatively and uniquely influence health outcomes in later years, even after accounting for the previously discussed mediation process (Hammen et al. 2009; Lantz et al., 2005; McGonagle & Kessler, 1990). For example, chronic stressors, like family economic stress, and later acute stressors, like children’s off-time transition events, may influence the later life health of husbands and wives independently of each other (cumulative hypothesis).

Health Effect of Acute Stress in Chronic Stress Context: The Maturation Hypothesis

In addition, chronic stress and acute stressful events may interact to influence mental and physical health outcomes. Previous research has provided mixed findings about such interaction effects. Some studies have shown a sensitization effect of chronic stress with chronic stress increasing individuals’ physiological and psychological susceptibility to acute stressful events (Hammen et al., 2009). Previous studies have shown that early stressful experiences interfere with the stress-response system, and youth who experience early chronic stress often develop an impaired or overactive physiological response to acute stressful events (Das, 2019; Loman & Gunnar, 2010). However, most of these studies have focused on childhood and young adulthood.

Conversely, other studies have shown that chronic stress decreases the susceptibility—or increases the resilience—of individuals against acute stressful events (Cairney et al., 2003). For example, chronic economic hardship may reduce susceptibility to acute stressful events such as a temporary pay cut. More specifically, McGonagle and Kessler (1990) suggest that “anticipation and reappraisal may reduce the stressfulness of stressful events by making its meaning more benign” (681p). Thus, people are “steeled” or “hardened” by ongoing chronic stress, such that later acute negative events have less impact (Rutter, 2012). Stress researchers also attribute this hardening effect to a sense of maturation and personal growth in response to chronic stress, including material gain, changes in perspective, stronger social bonds, increased coping skills, and increased self-knowledge (maturation hypothesis, see Aldwin & Sutton, 1998; Liu, 2015; Schilling & Dieihl, 2015).

We contend that the nature of this interaction effect between chronic and acute stress (i.e., sensitization or maturation) on mental and physical health may vary depending on the individual’s life stage, the stress-response period, study duration, and perhaps the nature of the stressors. Most previous stress studies have not adequately investigated stressful life experiences prospectively and comprehensively over an extended period of time (Lynch et al., 1997; Wickrama et al., 2019). Thus, in the present study, we expect that chronic stress decreases the susceptibility—or increases the resilience—of husbands and wives against acute stressful events in mid-later years: the maturation hypothesis.

Dyadic Associations between Spouses in Enduring Marriages

Consistent with the life course “linked lives” notion (Elder & Geile, 2009) and the relational perspective (Berscheid & Ammazzalorso, 2001), husbands and wives live in an intimate relational context influencing each other’s behaviors, beliefs, and feelings. Their behaviors and emotions are closely tied and mutually influential at the dyadic level. This dyadic process may be particularly salient for husbands and wives in enduring marriages who have a long shared history. Accordingly, this dependency of stress experiences and stress responses of husbands and wives in enduring marriages may result in contemporaneous associations or correlations and cross-over or partner effects between spouses.

Particularly, partner effects may exist because stress experiences and perceptions are transmitted between spouses. For example, if a husband experiences chronic economic hardship, this is communicated within the interdependent marital context to his wife, whose emotional response, in turn, may include the development of feelings of distress (Berscheild & Amazzalorso, 2001). Moreover, consistent with the emotional contagion perspective, feelings of distress can be “induced” and can “spread” from one partner to the other (Cacioppo et al., 2009).

Some research suggests that these partner effects may vary in magnitude depending on the spouse’s gender, with stronger effects noted for partner effects between wives’ distress and husbands’ mental health. For example, impaired marital role performance of wives due to their chronic marital stress may have stronger consequences for their husbands’ well-being. These husband–wife dependencies combined with the increased likelihood of biased parameter estimates when conducting separate analyses of husbands and wives (Kenny et al., 2006) necessitate dyadic analyses of husbands and wives over an extended period of time.

Present Study

As depicted in Figure 1, using prospective data over 25 years from a sample of 256 husbands and wives in enduring marriages (>40 years of marriage), the current study will investigate a dyadic life course model examining how midlife chronic stress and subsequent acute stressful life events in mid-later years (average age 40–65, 1991–2015) influence husbands’ and wives’ later life mental and physical health outcomes (>65 years in 2015). This investigation will enhance knowledge about the long-term joint influences of midlife chronic stress and acute stressful life events on health outcomes in later years. This enhanced knowledge can inform health preventive intervention programs targeting middle-aged adults in enduring marriages.

Figure 1.

Theoretical framework: Midlife chronic stress, stressful life events, and later life health outcomes of husbands and wives in enduring marriages.

The following specific hypotheses were drawn /from the theoretical framework presented in Figure 1:

Mediation Hypotheses:

Individuals’ chronic midlife stress (1991–2001) will be associated with their experiences of stressful acute life events in their mid-later years (2001–2015).

Individuals’ acute stressful life events over their mid-later years (2001–2015) will be associated with their health outcomes later life (i.e., depressive symptoms and poor physical health in 2015).

Acute stressful life events over the mid-later years (2001–2015) will mediate the association between chronic midlife stress (1991–2001) and later life health outcomes.

Cumulative Hypothesis:

Individuals’ chronic midlife family stress (2001) and subsequent acute stress independently and additively influence their own later life health outcomes (i.e., depressive symptoms and poor physical health in 2015).

3. Maturation Hypotheses: Individuals’ chronic midlife stress (2001–2015) will weaken the association between acute stressful life events in their mid-later years (2001–2015) and later life health outcomes (i.e., depressive symptoms and poor physical health in 2015).

Methods

Participants and Procedures

The data used to evaluate these hypotheses are from the Iowa Youth and Family Project (IYFP, 1989–1994), which was later continued as two panel studies: the Midlife Transitions Project in 2001) and the Later Adulthood Study in 2015 and 2017. Together, these projects provide data over 27 years on rural families from a cluster of eight counties in north-central Iowa that closely mirror the economic diversity of the rural Midwest. The IYFP began as a study of rural couples with children, at least one of whom was a seventh grader in 1989 (Conger & Elder, 1994).

The 256 couples in the current study are those who participated in 1991, 1994, 2001, and 2015 data collections and were consistently married throughout the study period. Data collected in 1991, rather than 1989, were used as the first time point of the current study due to the availability of study variables. The attrition rate for respondents was 30% from 1991 to 2015. The majority of couples not included in these analyses were divorced or separated by 2015. An attrition analysis compared the current analytic sample of consistently married couples who were excluded from the current analyses due to divorce and study attrition on demographic characteristics (i.e., age, education level, economic hardship measured by counts of economic cutbacks, and divorce proneness (Booth et al., 1983) and study variables (e.g., depressive symptoms) in 1991. The only significant difference noted was for divorce proneness in 1991, with higher scores reported for couples who were excluded from the current analysis. In 1991, spouses were in their early middle years.

In the study sample, the average ages of husbands and wives were 42 and 40 years, respectively, and their ages ranged from 33 to 59 years for husbands and 31 to 55 years for wives. On average, the couples had been married for 19 years and had three children. The median age of the youngest child was 12 years. In 1989, the average number of years of education for husbands and wives was 13.68 and 13.54 years, respectively. The median yearly income was $22,000 and $10,000 for husbands and wives, respectively. Almost all of the men were employed (99.6%), and 66.0% of the women were employed. The occupational status of the men in this sample included 22% farmers; 28% professionals and managers, including sales occupations; 40% technicians and skilled workers, including service occupations; 3% office workers; and 7% unskilled workers and other workers. The women’s occupational status included 30% clerks and secretaries; 27% professionals, including teachers; 9% technicians and skilled workers; 7% sales workers; and 27% service and household workers. Because there are very few minorities in the rural area studied, all participating families were white.

Measures

Midlife chronic stress

The measure of exposure to cumulative family stress captured multiple stress experiences (i.e., marital, parental, and family economic stress) over three occasions during their middle years (1991, 1994, and 2001) in the middle years (40–50 years). Based on previous cumulative adversity measures, scores were computed to address three important dimensions of cumulative stress exposure, including the intensity, frequency, and duration of stress (Bartley & Blane, 2009; Evans & Kim, 2010). First, standardized measures of different domains (i.e., family economic stress, marital stress, and parental stress) were summed separately for each of the three measurement occasions (1991, 1994, and 2001). Summing different domains of stress for each occasion captures the accumulation of family stress at each occasion. Next, for each measurement occasion, a mean split was utilized to distinguish “high” and “low” cumulative scores (0 = low stress and 1 = high stress), which created eight possible patterns of stress over the three occasions (e.g., 0 0 0 representing low stress at all three occasions, 1 0 0 representing high stress at the first measurement occasion only, and 0 1 1 representing high stress at the second and third occasions). Last, a cumulative midlife score capturing the duration of stress was assigned for each pattern. For instance, the pattern with 0s for all three occasions (0 0 0) was assigned a value of 0 to represent zero measurement occasions with above-average stress experience (1991, 1994, and 2001). The pattern with 1s for all three occasions (1 1 1) was assigned a value of 10 to represent 10 years of above-average stress (1991–2001). Other patterns were similarly assigned scores that were proportionate to the duration of stress experienced. Patterns with two adjacent, above-average stress values (i.e., 1 1 0 and 0 1 1 representing stress in 1991/1994 and 1994/2001, respectively) were assigned values of 3 and 7, respectively, which assumes the stress continued between the two occasions). For individuals with nonadjacent above-average stress values (1 0 1), a score of five was given to reflect half the number of years between the high-stress occasions. Patterns with a single above-average stress report were given scores equivalent to the half of number of years between the high-stress occasion and the closest low-stress occasion (e.g., a value of 1.5 for stress pattern 1 0 0 reflects half of the three-year distance between 1991 and 1994 measurements). Assigning scores to different patterns of stress based on the intensity, frequency, and duration of stress experienced captures the chronic exposure (and/or repetition) of cumulative family stress. Midlife chronic stress measures used to compute chronic family stress measure are as follows.

Family economic problems

The list of economic problems was adapted from Dohrenwend et al. (1978) to capture families’ economic circumstances. Separately for husbands and wives, for 1991, 1994, and 2001, “yes” responses to each of the 27 items were summed to indicate economic problems experienced by the family during the previous year (1 = yes and 0 = no). The list of economic problems included items such as “borrowed money to help pay bills,” “sold possessions or cashed in life insurance,” and “changed food shopping or eating habits to save money.” Means (and standard deviations, SDs) for the family economic problems in 1991, 1994, and 2001 were 4.84 (5.21), 3.76 (4.50), and 3.10 (3.60) for husbands and 5.13 (5.12), 4.80 (4.50), and 3.75 (4.10) for wives.

Marital stress

For 1991, 1994, and 2001, a composite measure of marital stress was calculated by summing standardized scores of spouses’ perceived hostile behavior, destructive conflict resolution behavior, and marital instability. For hostile marital behavior, at each time point, spouses indicated how often (1 = always and 7 = never) during the past month his/her partner engaged in seven hostile behaviors (Matthews et al., 1996). Sample items included “get angry at you,” “shout or yell at you,” and “make you feel guilty.” Higher scores represent a higher level of hostility. The range of Cronbach’s α was .89–.91 for husbands and wives. For destructive conflict resolution behavior, eight items assessed participants’ reports of their spouse’s destructive conflict resolution behaviors at each time point (e.g., “criticizes you or your ideas for solving the problem” and “ignores the problem,”). Items were scored on a 7-point Likert scale ranging from 1 = never to 7 = always. The internal consistencies varied from .80 to .91 for husbands and wives. A modified five-item short form version of the Marital Instability Index was utilized to assess marital instability. The scale’s validity and reliability have been demonstrated (Booth et al. 1983). The sample items asked how recently either of them suggested getting divorce and thought that their marriage might be in trouble (1 = not in the last year and 4 = within the last three months). The internal consistencies varied from .62 to .72 for husbands and wives. For husbands and wives, a composite marital stress measure was then constructed by summing standardized scores of the three indicators. Means (and SDs) for the measure of husbands’ marital stress in 1991, 1994, and 2001 were −.14 (5.50), −.10 (6.20), and .01 (5.91). For wives, means (and SD) in 1991, 1994, and 2001 were −.02 (2.01), −.17 (1.60), and −.03 (1.50).

Parental stress

In 1991 and 1994, mothers and fathers reported on the hostile behaviors of their child. This 10-item assessment was developed for the IYFP (Spoth et al., 2008). Sample items include how often in the past year children had “ignored you when you tried to talk to him or her,” and “insulted or swore at you.” Internal consistencies on this scale varied from .80 and .88 for mothers and fathers (M = 1.57 and SD = .59 for husbands and M = 1.61 and SD = .59 for wives in 1991 and M = 1.58 and SD = .60 for husbands and M = 1.59 and SD = .55 for wives in 1994). Given children’s life stage, in 2001, the 10-item assessment was less appropriate as a measure of parental stress; instead, parents were asked three items capturing how much “conflict, tension, and disagreement” they felt between themselves and their child on a scale from 1 (none) to 4 (a lot). Internal consistencies were .60–.65 for mothers and fathers, respectively (M = 1.51 and SD = .71 for husbands and M = 1.50 and SD = .64 for wives). To be consistent with the marital stress measure, a measure of parental stress in 2001 was calculated for husbands and wives by summing standardized scores of the three indicators.

Stressful acute life events (acute stress) in mid-later years

Respondents’ stressful midlife events from 2001 to 2015 (50–65 years, referred as acute stress in mid-later years) were obtained using a life events calendar reported in 2015. The calendar provided an extensive list of 45 life events that were classified into mutually exclusive domains, according to whether the events related to marriage (e.g., live separately due to work, school, and have had a “falling out”), children (e.g., dropped out of high school, unwanted pregnancy, and get arrested), or self and family (e.g., lose home, major one-time expense, moved to worse residence, and had a serious injury). Two items (i.e., became ill with a life-threatening disease and saw a doctor about an emotional problem) were deleted from the list because of possible contamination with the measures of physical and mental health. Then, for each domain, interviewers presented the respondent with the list of possible events and asked whether each event occurred once or more. The severity of each life event was captured by asking “How much did this event affect your life (i.e., goals, plans, and close relationships) on 4-category ordinal scale (1 = not at all to 4 = a lot). The average number of events recorded by women was higher than the number reported by men. Each event was multiplied by the severity score before obtaining the sum score of stressful events for each spouse.

Psychological distress

In their middle years (2001) and later life (2015) (65+ years), 13 items from the Symptoms Checklist 90-R (SCL-90-R; see Derogatis & Melisaratos, 1983) assessed husbands’ and wives’ depressive symptoms in the previous week. Items were scored on a 5-point Likert scale from 1 = not at all to 5 = extremely. Sample items include, “thoughts of ending your life,” “feelings of worthlessness,” and “feeling hopeless about the future.” The internal consistencies were greater than .90 for husbands and wives.

Poor physical health

A measure of poor physical health was created for 2001 (middle years) and later life (2015) (65+ years) by adding standardized measures of self-assessed poor global health, physical limitations, and bodily pain. Self-assessments of poor global health were obtained using two items. The first item asked participants to indicate on a scale from 1 = excellent to 5 = poor: “How would you rate your overall physical health?” The second item asked participants to indicate on a scale from 1 = much better to 5 = much worse: “Compared to one year ago, how would you rate your physical health in general now?” The mean of these two items was calculated with higher scores representing poorer physical health. These two items were highly correlated (r > .70 for each spouse at every measurement occasion). The degree of physical limitation was measured by the 10-item physical impairment scale of the Rand 36-Item Health Survey 1.0 (Hays et al., 1993). Respondents were asked to indicate on a 3-point scale ranging from 1 = no, not limited at all to 3 = yes, limited a lot how much they were limited by physical impairments. The scale captures impairment for vigorous (e.g., running or lifting heavy objects) and moderate activities (e.g., moving a table, pushing a vacuum cleaner, and lifting or carrying groceries). Responses were averaged to create a measure of overall physical limitation. Cronbach’s alpha varied from .80 to .91 for husbands and wives. The degree of bodily pain was assessed by two items from the Rand Health Science Program in Health Survey 1.0 (Hays et al., 1993). On a 6-point scale ranging from 1 = none to 6 = very severe, respondents indicated how much bodily pain they experienced in the four preceding weeks. Also, on a 5-point scale ranging from 1 = not at all to 5 = extremely, respondents indicated “how much pain interfered with normal work.” Responses were standardized and averaged. The correlation between the two items was .73 and .70 for husbands and wives, respectively.

Statistical Analyses

Within a structural equation modeling framework, a single, comprehensive dyadic model was used to test our theoretical model involving husbands’ and wives’ midlife chronic stress (1991–2001), stressful life events (2001–2015), and later life mental health and physical health (2015) after controlling for lagged health measures in 2001. All models were tested using Mplus, version 8 (Muthen & Muthen, 1998–2018).

Out of the sample of 256 couples, some cases were unavailable for a specific wave of data collection (nearly 9% of the data). Full information maximum likelihood was utilized to test the hypotheses with all available data. A range of fit indices was used to evaluate the model fit of the models including the chi-square statistic, cumulative fit indices (CFI), root mean square error of approximation (RMSEA), and the Tucker-Lewis Index (TLI). For the chi-square fit statistic, the model is thought to fit the data well when the chi-square divided by the degrees of freedom is below 3.0 (Carmines & McIver, 1981). When the CFI value is near or greater than .95 and the RMSEA value is close to or less than .06, this suggests that the model fits the data well (Hu & Bentler, 1999). The TLI, similar to the CFI, is an incremental fit index that assesses the improvement in the fit of a model over a baseline model, with values near or greater than .95 suggesting the model fits the data well (Kline, 2016).

Results

Table 1 presents descriptive statistics of the study variables. In general, husbands and wives reported moderate levels of midlife chronic stress in their middle years (a cumulative measure capturing 1991–2001) (M = 3.56 and 4.52, respectively). Husbands and wives averaged a similar number of acute stressful life events (M = 5.04 and 5.10, respectively). Zero-order correlations were in the expected directions (see Table 1). Husbands’ physical health and depressive symptoms were correlated in 2001 and 2015 as were wives physical health and depressive symptoms. Husbands’ and wives’ perceptions of midlife chronic stress were associated with their own physical health and depressive symptoms in 2001 and 2015 (p < .05).

Table 1.

Descriptive Statistics and Zero-Order Correlations.

Husband’s chronic stress 1991–2001

1.00

Wife’s chronic stress 1991–2001

0.37***

1.00

III

Husband’s acute life events 2001–2015

.23**

.07

1.00

Wife’s acute life events 2001–2015

.13+

.17

.59***

1.00

Husband’s chronic stress X acute life events

.09

.00

.50***

.28***

1.00

Wife’s chronic stress X acute life events

−.07

.05

.03

.32***

.23**

1.00

VII

Husband’s physical health index 2015

.34***

.11

.12+

.10

−.12+

−.11

1.00

VIII

Wife’s physical health index 2015

.16*

.33***

.09

.21**

−.08

.00

.23***

1.00

Husband’s depressive symptoms 2015

.30***

.10

.13+

−.14+

−.02

.41***

.17**

1.00

Wife’s depressive symptoms 2015

.16*

.37***

.16*

.24***

.10

.16*

.44***

.33***

1.00

Husband’s physical health index 2001

.29***

.15*

.06

−.07

.01

−.09

.58***

.18**

.29***

.19**

1.00

XII

Wife’s physical health index 2001

.09

.23***

.00

.13

−.04

.13+

.15*

.62***

.17***

.40***

.18**

1.00

XIII

Husband’s depressive symptoms 2001

.38***

.18**

.03

.06

−.08

.00

.32***

.08

.57***

.20**

.35***

.07

1.00

XIV

Wife’s depressive symptoms 2001

.15*

.38***

.06

.14

.09

.19**

.10

.30***

.25***

.53***

.14*

.34***

.18**

1.00

III

VII

VIII

XII

XIII

XIV

Mean

3.56

4.52

5.04

5.10

8.80

7.17

.00

18.47

19.60

.00

18.27

19.68

Standard deviation

3.58

3.37

10.74

11.60

42.86

39.09

.83

5.97

6.45

.79

.81

5.72

6.39

Note. Variables with an “X” represent interaction terms; +p < .10, *p < .05, **p < .01, ***p < .001.

Testing the Hypothesized Models of Depressive Symptoms as Outcomes

Overall, the model assessing how chronic and acute stress predicted depressive symptoms in later adulthood (see Figure 2) fit the data reasonably well (CFI = .95, RMSEA = .04, and χ² (df) = 48.57(32)). The model explained 39% and 32% of the variance in husbands’ and wives’ depressive symptoms, respectively. The results for specific study hypotheses are as follows.

Figure 2.

Results from a structural equation model with depressive symptoms as outcomes.

Mediation hypothesis

Mediation Hypothesis 1a posited that husbands’ and wives’ midlife chronic stress over a decade (four measurement occasions from 1991 to 2001) during their middle years would be related to their experiences of acute stressful life events during their mid-later years (2001–2015). Results showed that midlife chronic stress significantly predicted the occurrence of acute stressful life events during their mid-later years for both husbands and wives (β = .20 and .18, respectively, p < .01). Consistent with the Mediation Hypothesis 1b, husbands and wives who experienced more stressful acute life events in mid-later years averaged more depressive symptoms in later years (2015) after controlling for early depressive symptoms (2001) (β = .14 and .13, respectively, p < .05). There was also support for Mediation Hypothesis 1c (i.e., acute events would mediate the association between midlife chronic stress and depressive symptoms). The indirect effects from midlife chronic stress to later life depressive symptoms were statistically significant for both husbands and wives (β = .03 and .02, respectively, p < .05).

Cumulative hypothesis

There was mixed evidence for Hypothesis 2 given that midlife chronic stress was directly associated with later life depressive symptoms for wives (β = .15, p < .05) but not for husbands (p > .05). Additionally, in this model, wives’ midlife chronic stress directly influenced husbands’ later life depressive symptoms (2015) after controlling for depressive symptoms in 2001 (β = .16, p < .05), which indicates a partner effect.

Maturation hypothesis

As shown in panel A of Figure 4, consistent with Hypothesis 3, for husbands, there was a statistically significant, negative interaction between midlife chronic stress and acute stress influencing later life depressive symptoms (β = −.18, p < .01). Consistent with the maturation notion, this interaction effect indicates that earlier exposure to midlife chronic stress in midlife reduced the influence of acute stress on later life depressive symptoms for husbands. This interaction effect was not statistically significant for wives (panel B, Figure 4).

Contemporaneous dependencies were evident between spouses’ experiences of midlife chronic stress as well as their subsequent acute stressful life events and depressive symptoms. We tested the statistical significance of observed gender differences in hypothesized effects using inequality tests. The results showed that the gender difference in the interaction between chronic and acute stress on mental health (maturation effect) was significantly different (Δχ²(Δdf) = 5.30(1), p < .05). The gender difference in the effect of spouse’s midlife chronic stress on partner’s mental health was significantly different (Δχ²(Δdf) = 3.86(1), p < .05). The gender difference in effect of spouse’s midlife chronic stress on own mental health was only marginally different ((Δχ²(Δdf) = 3.00(1), p < .10).

Physical Health Symptoms as Outcomes

The model assessing how chronic and acute stress predicted physical health in later adulthood (see Figure 3) also fit the data well (CFI = .98, RMSEA = .03 and χ² (df) = 44.91(32)). The model explained 42% and 47% of the variance in husbands’ and wives’ later life physical health, respectively. Husbands’ and wives’ chronic stress as well as their acute stressful life events were correlated, and their acute stressful life events, which is evidence of contemporaneous dependencies between spouses.

Figure 3.

Association between acute stress and depressive symptoms and self-reported health for low (20th percentile) and high (80th percentile) chronic stress groups.

Mediation hypothesis

The results showed that midlife chronic stress significantly influenced the occurrence of acute stressful life events for both husbands and wives (β = .20 and .19, respectively, p < .01) (Mediation Hypothesis 1a), and, consistent with Mediation Hypothesis 1b, stressful acute life events significantly predicted poor physical health in later years (2015) (β = .13 and .14, respectively, p < .05) after controlling for poor physical health in 2001. Consistent with Mediation Hypothesis 1c, the indirect effects from midlife chronic stress to later life physical health were significant for both spouses (β = .03 for husbands and wives, respectively, p < .05).

Cumulative hypothesis

Consistent with the Hypothesis 2, midlife chronic stress was directly associated with poor physical health in later life poor for both husbands and wives (β = .19 and .14, respectively, p < .05). There was no evidence for partner effects between midlife chronic stress and poor physical health.

Maturation hypothesis

As shown in panels C and D of Figure 4, for both husbands and wives, there were statistically significant interactions between midlife chronic stress and acute stress for later life physical health (β = −.20 and −.14, respectively, p < .05) (Hypothesis 3). Consistent with the maturation notion, plotting these interaction effects indicated that midlife chronic stress reduced the detrimental effects of acute stress on later life physical health for both spouses.

Figure 4.

Results from a structural equation model with physical health outcomes. (a) Predicted values of husband depression scale at high and low values of chronic strain and acute events, (b) predicted values of wife depression scale at high and low values of chronic strain and acute events, (c) predicted values of husband physical health at high and low values of chronic strain and acute events, and (d) predicted values of wife physical health at high and low values of chronic strain and acute events.

Discussion

Greater individual variability in mental and physical health begins to manifest in midlife and continues into later years. Previous studies have documented that both chronic and acute stress play important roles in the development of health problems in this period (Hammen et al., 2009; Lantz et al., 2005; Lovallo, 2005), and older adults are more vulnerable to the physiological and psychological effects of stress (Zaidi, 2014). However, less is known about how midlife chronic stress and acute stressful life events in mid-later years combine to influence health outcomes in later life. Building on the life course mediation and cumulative models of health risks, we hypothesized a model incorporating both mediation and cumulative hypotheses as well as a maturation hypothesis to explore how chronic and acute stress processes explain variation in husbands’ and wives’ mental and physical health outcomes in later years. In summary, all three hypotheses were supported. Consistent with the mediation hypothesis, husbands’ and wives’ midlife chronic stress influenced their own stressful life events, which were implicated in both depressive symptoms and poor physical health in later years. Consistent with the cumulative hypothesis, direct effects of midlife chronic stress on later life physical and mental health were also noted. Last, consistent with the maturation hypothesis, for husbands and wives, midlife chronic stress reduced the health effect of stressful life events.

Regarding the findings that supported the mediation hypothesis, for both husbands and wives, chronic midlife stress influenced stressful life events in mid-later years. That is, we found that chronic difficulties and contextual factors contributed to the occurrence of additional stressful acute events (Turner & Turner, 2005). This proliferation of early midlife chronic stressors into later acute stressor demonstrates stress accumulation over the life course and the formation of a “chain of risk” (Kuh et al., 2003). The present investigation utilized a comprehensive midlife chronic stress measure and a total count measure of acute stressful life events. These measures did not allow for an examination of the association between midlife chronic stress and acute events separately for each domain (e.g., family, marriage, and child relation), although such an analysis would provide deeper insight into the occurrence and impacts of distinct stressful life events. Future research should investigate such domain-specific associations. A strength of the measure of stressful life events was weighting events by their seriousness, as reported by spouses. Weighting acute events by seriousness allowed for an assessment of these events based on the importance respondents assigned to them.

Consistent with the cumulative hypothesis, the influence of midlife chronic stress on later life health remained significant over an extended period of time (15–25 years) even after including acute stressful life events, which provides evidence for the unique influences of chronic and acute stresses and suggests distinct long-term and short-term etiological processes stemming from stress exposure. The chronic family stress measure was comprehensive and captured several domains of stress. Furthermore, based on previous cumulative adversity measures (Bartley & Blane, 2009; Evans & Kim, 2010), using prospective data over a decade, our chronic measure captured important aspects of cumulative stress exposure, including the intensity, frequency, and duration of stress exposure. Although both husbands and wives reported difficulties in the same domains, there were unique actor effects on later life health outcomes, which suggest that their reports diverged in meaningful ways from one another. To address potential reverse causation, analyses controlled for lagged health measures in 2001.

Because we measured midlife chronic stress from 1991 to 2001, it is possible that some of the chronic difficulties resolved after 2001. Nevertheless, for husbands’ physical health, the persistent health influence of midlife chronic stress remained statistically significant, suggesting either (a) chronic stressors persisted after 2001 or (b) the health effects of biological and psychological damage caused by chronic stress before 2001 manifested in later years. For wives’ mental health, the persistent mental health influence of chronic stress remained statistically significant, suggesting either (a) chronic stressors persisted after 2001 and/or (b) wives’ mental health susceptibility to midlife chronic stress is greater than husbands’ susceptibility. Variations in susceptibility may be partly attributed to their burden caregiving for family members and aging parents.

Importantly, consistent with maturation hypothesis, for both husbands and wives, midlife chronic stress in early midlife was associated with a reduction in detrimental health effects stemming stressful acute life events in later life (Cairney et al., 2003; McGonagle & Kessler, 1990). Researchers have provided several explanations for this type of maturation effect, including (a) experiencing and overcoming stress makes the meaning of stress more benign, (b) people who experience stress develop individual resources that protect them from later stress (“steeling” or “hardiness”), and (c) overcoming stress is associated with gaining more material and social resources and increased self-knowledge (Aldwin & Sutton, 1998; Schilling & Dieihl, 2015). The lengthy follow-up period in the current study provided sufficient time for these couples to gain material and other resources and to embrace more benign meanings for stressful life events. That is, such maturation effects may not be evident in studies following individuals over shorter periods of time because such effects may only be evident when sufficient time has passed. Results of this study suggest this is true for both husbands’ and wives’ physical health, but only for husbands’ mental health. The lack of a maturation effect for wives’ mental health may suggest greater psychological sensitivity to stress for women compared to men (Sandanger et al., 2004).

In sum, the findings support the proposition that chronic stress should be the central focus of stress-health investigations (McGonagle & Kessler, 1990) because midlife chronic stress showed direct, indirect, and interactive effects on mental and physical health outcomes in later life. In addition, our results showed that the occurrence of acute stressful events is influenced by the context of chronic stress for husbands and wives in enduring marriages. Thus, health programs and interventions for older adults should have a greater focus on reducing the chronic stress that they bring to their later years.

The dyadic analytical framework utilized in the study allowed us to consider spouses’ contemporaneous correlations and partner effects within a single analytical framework (Kenny et al., 2006). The results revealed differences between husbands’ and wives’ direct effects, indirect effects, and interactive effects as well as transactional partner effects involving the effects of midlife chronic stress and subsequent stressful acute events on health outcomes in later years. The results revealed that wives’ midlife chronic stress influenced husbands’ depressive symptoms in later years. Perhaps wives who experience more stress provide less spousal support in their mid-later years, which affects their husbands’ mental health. This possibility is consonant with research demonstrating that (1) wives generally provide more spousal support than husbands and (2) husbands benefit more from spousal support than wives (Cutrona, 1996).

These findings would not have observed in separate analyses for husbands and wives. Husbands’ and wives’ behaviors and emotions are closely tied and mutually influential at the dyadic level (Berscheid & Amazzalorso, 2001), and this dyadic process may be particularly salient for husbands and wives in enduring marriages (Wickrama et al., 2018). This dependency between husbands and wives in enduring marriages may result in dyadic associations between spouses that necessitate dyadic analyses over an extended period of time. An added benefit of dyadic analyses is the reduction of parameter estimate bias compared to single-gender analyses (Kenny et al., 2006).

There are limitations to the current study that should be noted. First, the assessment of health outcomes relied on self-report measures. Although self-reported health problems are important to research, it is also important to also examine health outcomes using clinical measures. This research can be extended to incorporate assessments of midlife chronic stress using objective measures (e.g., tax records to assess family economic hardship and law enforcement records to assess stressful life events). Second, retrospective reports of stressful life events may be affected by memory problems, especially for older adults. Third, the generalizability of the current findings is limited by the sample’s composition of European American individuals who lived in rural Iowa during the farm crisis of the 1980s. Although the farm crisis provided the opportunity to research relatively widespread economic stress, future studies testing similar models with a more diverse population are needed. For instance, future samples should include multiple ethnicities, greater variation in the length of marriage, and other geographic locations. Finally, the current study limited its investigation to economic, marital, and parental stress, but other life course stressors may also have long-term health effects and warrant investigation, such as stressful work and caring for aging parents.

Despite these limitations, the current study contributes to the enhancement of knowledge about the long-term influences of husbands’ and wives’ chronic and acute stress in their middle years on mental and physical health problems in later years. These findings provide support for the value and necessity of national- and state-level policies aimed at alleviating chronic family difficulties including persistent economic hardship and relational difficulties. These results are also important for consideration by mental health professionals and counselors as future interventions should not overlook earlier (i.e., midlife) stress experiences, particularly chronic stress, as a potential cause of health problems in later years. Clinical implications also include the improving interpersonal processes as a way to protect from the negative consequences of midlife chronic and acute stress.

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research is currently supported by a grant from the National Institute on Aging (AG043599, Kandauda A. S. Wickrama, PI). The content is solely the responsibility of the authors and does not necessarily represent the official views of the funding agencies. Support for earlier years of the study also came from multiple sources, including the National Institute of Mental Health (MH00567, MH19734, MH43270, MH59355, MH62989, MH48165, MH051361), the National Institute on Drug Abuse (DA05347), the National Institute of Child Health and Human Development (HD027724, HD051746, HD047573, HD064687), the Maternal and Child Health Bureau (MCJ-109572), and the MacArthur Foundation Research Network on Successful Adolescent Development Among Youth in High-Risk Settings. (AG043599 – NIA – Couple Relationships and Health during Transition to Later Adulthood)

Research Ethics

This research was approved by the University of Georgia Institutional Review Board. Reference number: ID#STUDY00001438

ORCID iD

Eric T. Klopack

References

Aldwin

C. M.

Sutton

K. J.

(1998). A developmental perspective on posttraumatic growth. In Tedeschi

R. G.

Park

C. L.

Calhoun

L. G.

(Eds.), The LEA series in personality and clinical psychology. Posttraumatic growth: Positive changes in the aftermath of crisis (pp. 43-63). Lawrence Erlbaum Associates Publishers.

Bartley

Blane

(2009). Life-course influences on health at older ages. In Graham

(Ed.), Understanding health inequalities (pp. 48-65). McGraw-Hill Education.

Berscheid

Ammazzalorso

(2001). Emotional experience in close relationships. In Fletcher

G. J.

Clark

M. S.

(Eds.), Blackwell handbook of social psychology: Interpersonal process (pp. 308-330). Blackwell.

Booth

Johnson

Edwards

J. N.

(1983). Measuring marital instability. Journal of Marriage and the Family, 45(2), 387-394. doi:10.2307/351516

Cacioppo

J. T.

Fowler

J. H.

Christakis

N. A.

(2009). Alone in the crowd: The structure and spread of loneliness in a large social network. Journal of Personality and Social Psychology, 97(6), 977-991.

Cairney

Boyle

Offord

D. R.

Racine

(2003). Stress, social support and depression in single and married mothers. Social Psychiatry and Psychiatric Epidemiology, 38(8), 442-449. doi:10.1007/s00127-003-0661-0

Carmines

E. G.

McIver

J. P.

(1981). Analyzing models with unobserved variables: Analysis of covariance structures. In Bohmstedt

G. W.

Borgatta

E. F.

(Eds.), Social measurement (pp. 65-115). Sage.

Caspi

Sugden

Moffitt

T. E.

Taylor

Craig

I. W.

Harrington

McClay

Mill

Martin

Braithwaite

Poulton

(2003). Influence of life stress on depression: Moderation by a polymorphism in the 5-HTT gene. Science, 301(5631), 386-389.

Conger

Elder

G. H.

(1994). Families in troubled times: Adapting to change in rural America. Aldine de Gruyter.

10.

Cutrona

C. E.

(1996). Social support in couples. Sage.

11.

Das

(2019). Chronic ongoing stressors and C-reactive protein: A within-person study. Journal of Aging and Health, 08982643198624122. Advance online publication. doi:10.1177/0898264319862419

12.

Derogatis

L. R.

Melisaratos

(1983). The brief symptom inventory: An introductory report. Psychological Medicine, 13(3), 595-605.

13.

Dohrenwend

B. S.

Askenasy

A. R.

Krasnoff

Dohrenwend

B. P.

(1978). Exemplification of a method for scaling life events: The PERI life events scale. Journal of Health and Social Behavior, 19(2), 205-229. doi:10.2307/2136536

14.

Elder

G. H

Jr Giele

J. Z.

(2009). The craft of life course research. The Guildford Press.

15.

Evans

G. W.

Kim

(2010). Multiple risk exposure as a potential explanatory mechanism for the socioeconomic status-health gradient. Annals of the New York Academy of Sciences, 1186(1), 174-189.

16.

Glymour

M. M.

Errel

K. A.

Berkman

(2010). What can life-course epidemiology tell us about health inequalities in old age? In Antonucci

T. C.

Jackson

J. S

(Eds.), Annual review of gerentology and geriatrics (pp. 77-97). Springer.

17.

Hammen

Kim

E. Y.

Eberhart

N. K.

Brennan

P. A.

(2009). Chronic and acute stress and the prediction of major depression in women. Depression and Anxiety, 26(8), 718-723.

18.

Hays

R. D.

Sherbourne

C. D.

Mazel

R. M.

(1993). The rand 36-item health survey 1.0. Health Economics, 2(3), 217-227.

19.

L. t.

Bentler

P. M.

(1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6, 1-55.

20.

Kenny

D. A.

Kashy

D. A.

Cook

W. L.

(2006). Dyadic data analysis. Guilford Press.

21.

Kline

R. B.

(2016). Principles and practice of structural equation modeling (4th ed.). The Guilford Press.

22.

Kuh

Ben-Shlomo

Lynch

Hallqvist

Power

(2003). Life course epidemiology. Journal of Epidemiology and Community Health, 57(10), 778-783.

23.

Lantz

P. M.

House

J. S.

Mero

R. P.

Williams

D. R.

(2005). Stress, life events, and socioeconomic disparities in health: Results from the Americans’ changing lives study. Journal of Health and Social Behavior, 46(3), 274-288.

24.

Liu

R. T.

(2015). A developmentally informed perspective on the relation between stress and psychopathology: When the problem with stress is that there is not enough. Journal of Abnormal Psychology, 124(1), 80-92. doi:10.1037/abn0000043

25.

Loman

M. M.

Gunnar

M. R.

(2010). Early experience and the development of stress reactivity and regulation in children. Neuroscience and Biobehavioral Reviews, 34, 867-876. doi:10.1016/j.neubiorev.2009.05.007

26.

Lorenz

F. O.

Wickrama

K. A. S.

Conger

R. D.

Elder

G. H.

Jr. (2006). The short-term and decade-long effects of divorce on women's midlife health. Journal of Health and Social Behavior, 47, 111-125.

27.

Lovallo

W. R.

(2005). Stress and health: Biological and psychological interaction (2nd ed.). Thousand Oaks, CA: Sage Publications.

28.

Lynch

J. W.

Kaplan

G. A.

Shema

S. J.

(1997). Cumulative impact of sustained economic hardship on physical, cognitive, psychological, and social functioning. New England Journal of Medicine, 337(26), 1889-1895.

29.

Matthews

L. S.

Wickrama

K. A. S.

Conger

R. D.

(1996). Predicting marital instability from spouse and observer reports of marital interaction. Journal of Marriage and the Family, 58(3), 641-655 doi:10.2307/353725

30.

McGonagle

K. A.

Kessler

R. C.

(1990). Chronic stress, acute stress, and depressive symptoms. American Journal of Community Psychology, 18(5), 681-706.

31.

Muthén

L. K.

Muthén

B. O.

(1998-2018). Mplus user’s guide (8th ed.). Muthén & Muthén.

32.

Pearlin

L. I.

Nguyen

K. B.

Schieman

Milkie

M. A.

(2007). The life-course origins of mastery among older people. Journal of Health and Social Behavior, 48(2), 164-179.

33.

Pearlin

L. I.

Schieman

Fazio

E. M.

Meersman

S. C.

(2005). Stress, health, and the life course: Some conceptual perspectives. Journal of Health and Social Behavior, 46(2), 205-219.

34.

Pearlin

L. I.

Skaff

M. M.

(1996). Stress and the life course: A paradigmatic alliance. The Gerontologist, 36(2), 239-247. doi:10.1093/geront/36.2.239

35.

Rutter

(2012). Resilience as a dynamic concept. Development and Psychopathology, 24(2), 335-344.

36.

Sandanger

Nygård

J. F.

Sørensen

Moum

(2004). Is women’s mental health more susceptible than men’s to the influence of surrounding stress? Social Psychiatry and Psychiatric Epidemiology, 39(3), 177-184.

37.

Schilling

O. K.

Diehl

(2015). Psychological vulnerability to daily stressors in old age. Zeitschrift für Gerontologie und Geriatrie, 48(6), 517-523.

38.

Spoth

Randall

G. K.

Shin

(2008). Increasing school success through partnership-based family competency training: Experimental study of long-term outcomes. School Psychology Quarterly, 23(1), 70-89. doi:10.1037/1045-3830.23.1.70

39.

Thoits

P. A.

(2010). Stress and health: Major findings and policy implications. Journal of Health and Social Behavior, 51(Suppl 1), S41-S53.

40.

Turner

H. A.

Turner

R. J.

(2005). Understanding variations in exposure to social stress. Health: An Interdisciplinary Journal for the Social Study of Health, Illness and Medicine, 9(2), 209-240.

41.

Wheaton

(1996). The Domains and boundaries of stress concepts. In Howard

B. K.

(Ed.), Psychosocial stress: Perspectives on structure, theory, life-course, and methods (pp. 29-70). Academic Press.

42.

Wickrama

K. A. S.

Lee

Klopack

E. T.

Wickrama

(2019). Stressful work conditions, positive affect, and physical health of middle‐aged couples: A dyadic analysis. Stress and Health, 35, 382, 395. doi:10.1002/smi.2866

43.

Wickrama

K. A. S.

O'Neal

C. W.

Klopack

E. T.

Neppl

T. K.

(2018). Life course trajectories of negative and positive marital experiences and loneliness in later years: Exploring differential associations. Family process, 59, 142, 157. doi:10.1111/famp.12410

44.

Zaidi

(2014). Life cycle transistions and vulnerabilities in old age: A review. Human development report office. United Nations Development Programme.