Psychological and Physical Health in Widowhood: Does Marital Quality Make a Difference?

Abstract

We investigate how preloss marital quality is associated with changes in psychological distress and physical health among older widow(er)s. Using prospective data with a 2-year follow-up from the Health and Retirement Study, we selected 546 respondents who transitioned into widowhood. Respondents were classified as supportive, ambivalent, aversive, or neutral groups. The supportive and ambivalent group experienced greater increase in depressive symptoms compared to the aversive group, in widowhood. The aversive group showed greater increase in chronic conditions compared to the supportive group. Findings indicated that spousal loss may result in more psychological distress for those with supportive and ambivalent marital relationship. Yet, those with mostly negative accounts of their marriage may experience worsened physical health, albeit no increase in psychological distress. Understanding different benefits and challenges facing older individuals after a positive or negative marriage may help direct support and interventions efforts toward older couples during marriage and in widowhood.

Keywords

widowhood marital quality depressive symptoms chronic conditions

Widowhood is one of the most stressful life events that is likely to occur in later life (Carr & Utz, 2001; Sasson & Umberson, 2014; Schaan, 2013), resulting in abrupt and gradual alterations in older widow(er)’s everyday lives and overall well-being (Stroebe et al., 2007; Wilcox et al., 2003; Williams, 2004). This is because marriage often functions as a primary source of support and companionship and as a protective factor for psychological and physical health and well-being (Choi & Ha, 2011; Walker & Luszcz, 2009). However, research also finds that negative features of one’s marriage, such as poor marital quality and caregiving-related burden and stressors, have deleterious effects on well-being (Choi & Marks, 2008; Sandberg et al., 2013; Umberson & Williams, 2005; Whisman & Uebelacker, 2006). The meaning of a spousal loss, therefore, can be manifold. It could mean losing a critical source of support and companionship, which may also lead to changes in daily routines and social roles. For individuals strained by poor marital relationships, however, spousal loss can mark an end of a problematic relationship (Pruchno et al., 2009).

Spousal loss often results in a period of intense grief and has been found to increase older adults’ risk of morbidity and mortality (Lillard & Waite, 1995; Rendall et al., 2011; Stroebe et al., 2007; Wilcox et al., 2003; Williams, 2004). Yet, studies also indicate that most widowed individuals find ways to deal with the grief and successfully adjust to their new lives (Bonanno et al., 2002; Carr et al., 2000), and that the majority of widowed persons do not experience long-lasting negative consequences of the loss (Stroebe et al., 2007).

Key factors that explain the various ways in which widowed persons respond to spousal loss can largely be attributed to characteristics of the deceased and the bereaved, or those of the marriage (Carr et al., 2000). However, earlier findings pertaining to the role of the marital relationship in particular have not been consistent, and often not in line with theoretical predictions (Carr, 2004; Carr & Boerner, 2009; Carr et al., 2000; Choi & Vasunilashorn, 2014; Itzhar-Nabarro & Smoski, 2012). The objective of this study was to examine the link between preloss assessments of marital quality and subsequent changes in health among older adults who have transitioned into widowhood.

Theoretical Perspectives

Two main theoretical arguments on the topic of marital quality and well-being in widowhood come from the psychoanalytic perspective and attachment theory. The psychoanalytic perspective holds that the loss of a conflicted or troubled marital relationship is likely to result in more intense, or even pathological grief; that is, a conflicted or troubled relationship is more difficult to let go, as the widowed person may be stuck with unresolved feelings such as anger or guilt (Abakoumkin et al., 2010). Another pertinent line of thought comes from attachment theory, from which two distinct predictions can be derived. On the one hand, Bowlby’s original framework (1980) suggests that relationship dissolution thorough spousal loss can cause separation distress: the closer the emotional bond, the more intense the separation distress. On the other hand, another specification of attachment theory (e.g., Mikulincer & Shaver, 2008) is more in line with the psychoanalytic perspective, in that individuals with an insecure attachment (i.e., anxious, ambivalent attachment) are postulated to be more likely to experience adjustment difficulties following spousal loss, compared to those with a secure attachment. This is because, unlike securely attached persons who maintain confident attitude and expectations about available support, these individuals may have difficulty stipulating emotional challenges during and after the adversities, that may intensify the level of distress and impair one’s ability to seek support (Mikulincer & Shaver, 2008). They may also proactively reject or disapprove available support which in turn can lead to psychological and somatic symptoms (Mikulincer & Shaver, 2008).

Inconsistencies in Empirical Evidence: Methodological and Conceptual Challenges

Empirical evidence from studies on the topic is reflective of the mixed theoretical explanations. For example, a number of earlier studies supported the psychoanalytic prediction; that is, widowed persons who evaluated their marriage as conflicted were more likely to report higher levels of anxiety, guilt, and depression, even after several years following the death of their spouse (Parkes, 1988; Parkes & Weiss, 1983; Pruchno et al., 2009). Conversely, findings from a study by Futterman and colleagues (1990) found an opposite pattern, where greater marital satisfaction was associated with more distress over the loss. To add to the ambiguity, several other studies found no association between marital satisfaction and postloss distress (Bonanno et al., 2004; Ott et al., 2007).

Such inconsistencies may in part be attributable to several methodological shortcomings. First, when longitudinal population-based research was scarce, many studies relied on retrospective assessments of the marital relationship, often assessed after the loss (e.g., Futterman et al., 1990; Mancini et al., 2009). Such assessments could be negatively biased as widowed individuals suffering from psychological distress are more likely to report negative evaluations of their marriage, or could be positively biased when they engage in sanctification as a way of honoring the deceased (Lopata, 1981). To overcome this issue, researchers have increasingly invested in conducting studies using prospective data, which have allowed them to separate preloss conditions from postloss outcomes. Previous studies based on prospective study designs found that widowed persons reporting conflicted marriages showed lower levels of distress while those with more positively assessed marital relations experienced higher levels of distress (Bonanno et al., 2002; Carr et al., 2000), as well as higher health care expenditures after spousal loss (Prigerson et al., 2000).

The second issue is related to how marital quality was conceptualized in earlier studies. Scholars widely acknowledge that marital quality is a complex concept, as reflected in its interchangeable usage with other concepts such as marital dependency, closeness, satisfaction, and conflict. Earlier studies on widowhood did not consider the complex nature of marital relationships, as they have typically utilized a unidimensional marital quality measure. An exception is a study by Abakoumkin and colleagues (2010), which utilized separate measures of positive and negative marital quality to create an indicator of ambivalence. This study found that higher negative quality ratings were associated with higher distress levels during the marriage; however, this association was no longer in effect in widowhood. Positive quality ratings were associated with lower distress levels in both marriage and widowhood, whereas ambivalent ratings were unrelated to distress at all.

Finally, the issue of how to adequately analyze the relationship between marital quality and health in widowhood merits more consideration. Studies indicate that strained marital relationships are associated with poor psychological and physical health, whereas individuals reporting satisfactory marriage quality often show better health (Umberson et al., 2006). The stress-process model (Pearlin et al., 1981; Thoits, 2010) provides a useful framework for understanding troubled marital quality or spousal loss as primary stressors and other psychological and social deficits induced by the primary stressors as secondary stressors, which in turn may facilitate behavior changes (e.g., adoption of health risk behaviors, disengagement from health-promoting behavior) that eventually lead to poorer physical health outcomes. Regardless of the issue of causal direction between marital quality and health, disruption of marriage may have differential health consequences depending upon individuals’ health status prior to the spousal loss. For example, individuals with good health and a positive marital relationship may be more susceptible to experiencing negative consequences with the disruption of the status quo, whereas those with poor health and a negative marital relationship may experience the opposite. To take into consideration such bidirectional relationship between health and marriage, preloss self-reported health is conceptualized as a confounder for the relationship between preloss marital quality and subsequent well-being in widowhood, and thus used as a control in this study (Fayers & Sprangers, 2002; Pruchno et al., 2009).

In summary, the empirical evidence on the role of marital quality on psychological outcomes in widowhood remains ambiguous. What can be concluded from the state of the empirical literature is that the theoretical view which holds that troubled marriages lead to a more difficult postloss adjustment has found little support. What needs more clarification is how, if at all, marital quality affects health outcomes in widowhood, and which marital quality aspects may matter for which type of outcomes. Importantly, little is known about the association between marital quality and physical health in widowhood, as the topic remains largely unexplored (Choi & Vasunilashorn, 2014; Wilcox et al., 2003). Considering the well-documented morbidity and mortality risks associated with widowhood (Choi & Ha, 2011; Stroebe et al., 2007; Wilcox et al., 2003), this is an important gap to fill.

Current Study

In this study, we draw from the concept of ambivalence to capture the possible coexistence of positive and negative aspects of marital relationships (Lüscher & Pillemer, 1998; Suitor et al., 2011). We classified married individuals into four distinct typologies: namely, individuals who rate their marriage as 1) both highly positive and negative (ambivalent), 2) highly positive and not negative (supportive), 3) not positive and highly negative (aversive), or 4) neither positive nor negative (indifferent/neutral) (Uchino et al., 2004). Previous studies indicate that older adults expressing an ambivalent feeling toward their social network members report a lower level of well-being (Fingerman et al., 2004). To our knowledge, however, no study to date has examined the association between marital quality and health in widowhood based on these four distinct groups.

Based on our review of the literature, we postulated that marital quality would be associated with different patterns of changes in psychological distress and physical health for individuals who transition into widowhood. Given the fragmented evidence, we offer no formal hypothesis on the association between each subtype of marital quality and changes in physical health. However, we expected that individuals who rated their marriage as solely negative (i.e., aversive) would report better psychological distress outcomes (i.e., steady or less precipitous changes) compared to those who did not, especially in relation to those with positive accounts of the marital quality.

Methods

Data Sources

We used data from the Health and Retirement Study (HRS), a nationally representative longitudinal panel survey of individuals 51 and older, initiated in 1992 with follow-up interviews every 2 years thereafter (Servais, 2010). Data for this study were primarily taken from the core public-use files; additionally, variables for health measures, household income, and marital history were taken from the file constructed by the RAND Corporation.

A set of key variables for the study, including measures on marital quality (MQ), was collected through the Leave-Behind Questionnaire (LBQ), a self-administered survey designed to collect detailed information on psychological well-being and social relationships every 4 years. The LBQ was first introduced and distributed to a random half of the HRS sample in 2006 (n = 7,635), and the remaining half (n = 6,576) was administered in 2008. Accordingly, we treated the first wave (2006, 2008, 2010, or 2012) in which an individual participated in the LBQ as the baseline, and the subsequent wave (2008, 2010, 2012, or 2014) as the follow-up.

To identify the analytic sample, we first excluded proxy respondents and selected those who completed the LBQ at baseline (n = 6,914 (in 2006); 5,935 (in 2008); 7,431 (in 2010); 6,483 (in 2012)). We then selected respondents who were married and lived with spouse at baseline and widowed at follow-up: married in 2006 and widowed in 2008 (n = 163), married in 2008 and widowed in 2010 (n = 158), married in 2010 and widowed in 2012 (n = 136), and married in 2012 and widowed in 2014 (n = 122). Pooled together, the sample consisted of 579 individuals. Of the 579 respondents, we excluded respondents who were younger than 55 at baseline (n = 26) and had missing data on key study variables (e.g., marital quality measures) (n = 7), yielding the final analytic sample of 546 respondents.

Measures

Health outcomes

Psychological distress was assessed at baseline and follow-up with the short form of the Center for Epidemiologic Studies Depression (CES-D) Scale (Andresen et al., 1994). This version of the scale included eight items asking whether the following statements had been true for respondents much of the time during the past week: 1) was depressed, 2) everything was an effort, 3) sleep was restless, 4) felt lonely, 5) felt sad, 6) could not get going, 7) was happy, and 8) enjoyed life (1 = yes). Two positive items (i.e., “was happy” and “enjoyed life”) were reverse coded. Affirmative responses for the eight items were summed, with higher scores indicating greater symptoms of depression.

Physical health was assessed at baseline and follow-up with an index variable for the number of chronic conditions that a doctor had ever told the respondent that he or she had. A total of seven chronic conditions was considered: 1) high blood pressure or hypertension, 2) diabetes, 3) cancer, 4) lung disease, 5) heart conditions, 6) stroke, and 7) arthritis or rheumatism. Consequently, the number of chronic conditions ranged from 0 to 7, with higher scores indicating worse physical health.

As noted earlier, the focus of this study was to examine changes in health among those who transitioned to widowhood. One possible way of acknowledging systematic patterns in examining the link between marital quality and health is to analyze the change score of the (health) variable of interest, often calculated by the difference in pre-post measures (Fitzmaurice, 2001). This approach is deemed especially pertinent when the initial levels of variable under scrutiny differ across groups, as is the case for this study (for a detailed discussion on this issue, refer to Fitzmaurice, 2001). Therefore, change scores (i.e., follow-up value subtracted by baseline value) were constructed for both health outcomes and were used as dependent variables in the analytic models, where observed range for changes in depressive symptoms was −8 to 8, whereas changes in chronic conditions ranged from 0 (i.e., no change) to 2.

Marital quality

MQ at baseline was assessed using two separate scales for subjective evaluations of positive and negative marital relationship quality. The positive scale included three items: 1) “how much does your spouse really understand the way you feel about things,” 2) “how much can you rely on your spouse if you have a serious problem,” and 3) “how much can you open up to your spouse if you need to talk about your worries.” The negative scale contained four items: 1) “how often does your spouse make too many demands on you,” 2) “how much does your spouse criticize you,” 3) “how much does your spouse let you down when you are counting on him or her,” and 4) “how much does your spouse get on your nerves.” Response options for each item included 1) a lot, 2) some, 3) a little, and 4) not at all.

We first created an indicator variable denoting a “high” level of positive quality, using the number of “a lot” and “some” responses to items in the positive scales: respondents with three “a lot” or “some” responses to the three items in the positive scale were identified as “having a high level of positive MQ (= 1),” while those with less than three were treated as “not having a high level of positive MQ (or a low level of positive MQ) (= 0).” Similarly, an indicator for a high level of negative MQ was coded as 1 for at least three “a lot” or “some” responses in the negative scale and 0 for others. MQ indicators were treated as missing if respondents had missing values for two or more items on the positive scale or three or more items on the negative scale (Clarke et al., 2007). We then used combinations of the positive and negative MQ indicators to classify respondents’ MQ as supportive (high positive/low negative; n = 318 (58.2%)), ambivalent (high positive/high negative; n = 43 (7.9%)), aversive (low positive/high negative; n = 75 (13.7%)), and neutral (low positive/low negative; n = 110 (20.1%)). It should be noted that we refrained from using a widely used approach to computing ambivalence scores (Griffin’s Similarity and Intensity of Components formula; Thompson et al., 1995) for two reasons: 1) ambivalence scores tend to be more strongly affected by scores for the negative scale than by those of the positive scale, and 2) ambivalence scores using Griffin’s formula do not differentiate between ambivalent and neutral relationships (Windsor & Butterworth, 2010).

Covariates

A set of demographic, socio-economic, health, marriage-related, and social integration variables known to be associated with depressive symptoms and chronic conditions in widowhood were included in the multivariate analyses. All covariates were measured at baseline except for living arrangement, which was assessed at follow-up. Demographic characteristics included age (in years), gender, and race/ethnicity (non-Hispanic White = 1). Socio-economic variables included education (in years), and household income (transformed by the natural log).

Self-rated health was assessed as fair or poor health (= 1) and good, very good, and excellent health (= 0). We included two dummy variables indicating whether respondents received functional assistance from the spouse (received help = 1) and provided such assistance (provided help = 1). Previous studies showed that spousal care exchange can influence both marital quality and health outcomes (Choi & Marks, 2006; Freedman et al., 2014). As marriage-related measures, a variable denoting the duration of current marriage (in years, range: 0.3–69.9) and a variable indicating whether the current marriage was the first one (first marriage = 1; others = 0) were included. Social integration in widowhood was assessed with the living arrangement in widowhood, coded as living alone (= 1) and living with others (= 0).

Analytic Plan

We first examined descriptive characteristics of the full study sample and subsamples stratified by MQ types, focusing on the differences in the average numbers of depressive symptoms and chronic conditions at baseline, as well as at follow-up and the changes thereof, using a series of paired t-tests and analysis of variance (ANOVA). We then estimated a series of ordinary least squares (OLS) regression models to predict changes in depressive symptoms and Poisson regression models to predict changes in chronic conditions; the first model included the set of indicators for MQ types only, and the adjusted model included all covariates. Several covariates, including respondents’ and spouses’ self-rated health, order and length of current marriage at baseline, which were unrelated to the outcome measures and did not change the model outcomes, were excluded from the final adjusted model in the interest of parsimony. There was no substantive difference in major findings between the full and current models. All analyses were performed using STATA 14.0.

Results

Descriptive Statistics

Descriptive characteristics of the study sample are presented in Table 1. The mean age of the study sample was 73 years. Sixty-eight percent were women, and the majority of the sample was identified as non-Hispanic White (80%). About 28% of the sample rated their health as either fair or poor. In addition, only seven percent reported that they received functional help from the spouse at baseline, when 26% provided help to the spouse. At baseline, 59% of the sample had a spouse who rated his or her health as fair or poor. The average length of marriage for the study sample was 44 years and 71% were in a first marriage at baseline. About 77% of the sample were living alone in widowhood.

The study sample showed differences with regard to race/ethnicity, household income, self-rated health, spousal help provision, as well as the marriage characteristics when it was stratified by MQ types (see Table 1). Especially, respondents identified as having aversive MQ were distinctively discernable from others: they were not only the youngest (M = 70.96, SD = 8.30), but also had the lowest proportion of non-Hispanic Whites (68%) and the lowest level of household income, the shortest length of marriage (41 years), the lowest proportion of respondents in a first marriage (67%), and the highest proportion of respondents with fair or poor health (41%) at baseline. Respondents identified as having ambivalent MQ were more likely to be female, were highly educated with the highest level of household income, had spent the longest time with the spouse (M = 49.46 years, SD = 14.91) with the highest proportion of respondents in a first marriage (88%). They had the lowest proportion of reporting fair or poor self-rated health (23%), having a spouse who reported his or her health as fair or poor (54%) and providing help to spouse (22%).

Table 1.

Descriptive Characteristics of the Study Sample by Preloss Marital Quality Groups at Baseline (N = 546).

	Full Sample		Neutral		Aversive		Supportive		Ambivalent		χ² or F
	M/% (SD)		M/% (SD)		M/% (SD)		M/% (SD)		M/% (SD)		χ² or F
Respondent^a
Age (51–95)	73.22	8.74	74.12	8.10	70.96	8.30	73.40	8.51	73.51	8.59	2.29	†
Female	68.1%		71.8%		74.7%		64.2%		76.7%		5.95
Non-Hispanic White	79.7%		73.6%		68.0%		84.0%		83.7%		12.83	**
Education in years	12.40	3.05	12.05	3.12	11.75	3.26	12.61	3.02	12.88	2.49	2.52
Household income (logged)	10.61	0.89	10.50	0.72	10.18	1.35	10.73	0.78	10.76	0.85	8.85	***
Fair or poor self-rated health ^b	28.2%		35.5%		41.3%		23.3%		23.3%		13.58	**
Received help from spouse	7.3%		4.5%		5.3%		9.1%		4.7%		3.65
Provided help to spouse ^c	26.0%		38.1%		25.7%		22.4%		22.0%		10.43	*
Length of current marriage ^d	44.03	17.25	47.10	15.36	40.73	17.25	43.04	17.90	49.46	14.91	3.81	*
First marriage ^e	71.4%		78.9%		66.7%		67.6%		88.4*		12.12	**
Living alone at follow-up	76.6%		79.1%		65.3%		77.7%		81.4%		6.44	†
Spouse
Fair or poor self-rated health ^{b, f}	59.1%		66.4%		68.6%		55.1%		53.7%		7.47	†
N	546	(100.0%)	110	(20.1%)	75	(13.7%)	318	(58.2%)	43	(7.9%)

Note. ^†p < .10. *p < .05. **p < .01. ***p < .001.

^a all baseline unless noted otherwise.

^b 1 = poor or fair, 0 = good, very good, or excellent.

^c n = 520.

^d n = 531.

^e n = 545.

^f n = 523.

Overall, respondents showed more compromised health status in widowhood compared to prior to the loss (see Table 2). A series of paired t-tests indicated that the mean numbers of both depressive symptoms and chronic health conditions significantly increased in widowhood compared to baseline. The average number of depressive symptoms at baseline was 1.43, which subsequently increased to 2.32 at follow-up (p < .001). The average number of chronic conditions also increased from 1.97 at baseline to 2.16 in widowhood (p < .001). However, such pattern was not observed in a consistent manner when the sample was stratified by MQ types. Supportive and ambivalent MQ types were associated with a significant increase in the average number of depressive symptoms, whereas such changes were not observed among aversive and neutral MQ types. All MQ types were associated with a significant increase in the average number of chronic conditions.

Table 2.

Comparison of Depressive Symptoms and Chronic Conditions at Baseline and Follow-Up by Preloss Marital Quality Groups (N = 546).

	Baseline (A)		Follow-up (B)		Paired t		Change (B-A)
	M (SD)		M (SD)		Paired t		M (SD)
Depressive symptoms (0–8)
Full Sample	1.43	(1.91)	2.32	(2.35)	8.50	***	0.88	(2.43)
Neutral	1.77	(1.92)	2.24	(2.21)	1.98	†	0.46	(2.46)
Aversive	2.49	(2.54)	2.65	(2.64)	0.47		0.16	(2.93)
Supportive	1.10	(1.67)	2.24	(2.32)	9.12	***	1.14	(2.23)
Ambivalent	1.21	(1.57)	2.53	(2.34)	3.54	**	1.33	(2.46)
f-test	13.10	*** ^{a, b, c}	0.81				5.08	** ^d
Chronic conditions (0–7)
Full Sample	1.97	(1.20)	2.16	(1.27)	9.98	***	0.19	(0.46)
Neutral	2.15	(1.11)	2.32	(1.10)	4.34	***	0.16	(0.40)
Aversive	2.01	(1.08)	2.31	(1.28)	4.32	***	0.29	(0.59)
Supportive	1.92	(1.26)	2.10	(1.32)	7.18	***	0.18	(0.45)
Ambivalent	1.77	(1.15)	1.95	(1.19)	2.71	**	0.19	(0.45)
f-test	1.49		1.52				1.45

Note. ^†p < .10. *p < .05. **p < .01. ***p < .001.

Post hoc Scheffe test was applied to examine the specific difference between each of the four preloss marital quality groups: ^a Neutral > Supportive (p < .05).

^b Aversive > Supportive (p < .001).

^c Aversive > Ambivalent (p < .01).

^d Supportive > Aversive (p < .05).

Psychological distress (measured with the number of depressive symptoms) and physical health (measured with the number of chronic conditions) status at baseline, as well as the change patterns between baseline and follow-up, also showed significant differences when observed by MQ types. At baseline, respondents with aversive MQ showed the highest level of depressive symptoms (M = 2.49), followed by neutral (M = 1.77), ambivalent (M = 1.21), and supportive (M = 1.10) MQ types (p < .001). At follow-up, no difference across four MQ types was found for the number of depressive symptoms. Ambivalent and supportive MQ types were associated with the highest net-increase in depressive symptoms in widowhood (M = 1.33; M = 1.14), whereas aversive MQ type showed the least amount change (M = 0.16).

The number of chronic conditions at baseline and follow-up was not statistically different across the four MQ types (f = 1.49, p = .22; f = 1.52, p = .21), although respondents with ambivalent MQ types reported the lowest (M = 1.77) number of chronic conditions at baseline, followed by supportive (M = 1.92), Aversive (M = 2.01), and neutral (M = 2.15). Aversive MQ type was associated with the highest net-increase in chronic conditions in widowhood (M = 0.29), but the difference across four MQ types was not statistically significant (f = 1.45, p = .21).

Results from Regression Models

Results from the OLS regression models indicated a substantial difference with regard to changes in health depending on the preloss MQ (Table 3). Models estimating changes in depressive symptoms (Panel 1) showed that supportive and ambivalent MQ types were associated with a greater increase in depressive symptoms compared to aversive MQ type. That is, the net increase in depressive symptoms in widowhood for persons with ambivalent and supportive MQ was 1.17 (p < .05) and 0.98 (p < .01) greater than their aversive MQ counterpart in the unadjusted model, respectively. Such gaps between the groups became more pronounced after covariates were taken into consideration; in the adjusted model, the net increase in depressive symptoms among ambivalent and supportive MQ types was 1.23 (p < .01) and 1.05 (p < .01) greater than their aversive group counterpart, respectively. We plotted the changes in depressive symptoms among respondents based on the results from the adjusted models: ambivalent and supportive MQ types were associated with an increase in the number of depressive symptoms, whereas changes in the number of depressive symptoms were not statistically significant from zero for neutral and aversive MQ types (Figure 1).

Models predicting changes in chronic conditions (Table 3, Panel 2) also showed some support for the differential associations between preloss MQ and physical health in widowhood. Although the differences in changes in chronic conditions across MQ types were less pronounced compared to changes in depressive symptoms, aversive MQ respondents showed a significantly greater increase in chronic conditions compared to their supportive MQ counterpart in the unadjusted model (p < .05). This relationship remained largely unchanged in the adjusted model (p < .05) (see Figure 1).

Table 3.

Effects of Preloss Marital Quality on Change in Depressive Symptoms and Chronic Conditions in Widowhood (N = 546).

	Panel 1.Depressive Symptoms								Panel 2.Chronic Conditions
	Unadjusted				Adjusted				Unadjusted				Adjusted
	B		95%	CI	B		95%	CI	IRR		95%		CI	IRR	95%	CI
Marital Quality^a
Neutral	0.30		−0.40	1.01	0.36		−0.36	1.08	0.56	†	0.30	1.04	0.53	†	0.28	1.00
Supportive	0.98	**	0.37	1.58	1.05	**	0.42	1.68	0.61	*	0.37	1.00	0.60	*	0.36	1.00
Ambivalent	1.17	*	0.26	2.07	1.23	**	0.31	2.15	0.63		0.28	1.42	0.61		0.27	1.38
Respondent ^b
Age					−0.02	†	−0.05	0.00					1.02		0.99	1.04
Female					−0.02		−0.46	0.42					1.44		0.92	2.27
Non-Hispanic White					−0.03		−0.58	0.52					1.33		0.76	2.32
Education in years					−0.01		−0.09	0.06					0.96		0.90	1.03
Household income (logged)					−0.01		−0.27	0.26					0.97		0.77	1.23
Received help from spouse					−0.31		−1.10	0.47					1.46		0.78	2.76
Living alone at follow-up					0.10		−0.39	0.60					1.00		0.63	1.58
Constant	0.16		−0.38	0.70	1.75		−1.64	5.14	0.29	***	0.19	0.45	0.06		0.00	1.27

Notes. ^†p < .10; *p < .05; **p < .01; ***p < .001; ^a reference: Aversive; ^b all baseline unless noted otherwise. IRR = Incidence Rate Ratio.

Discussion

The primary aim of this study was to examine how preloss marital quality was linked with changes in depressive symptoms and chronic conditions among persons who transitioned into widowhood. Based on the concept of ambivalence, we operationalized marital quality as four distinct subtypes. As expected, levels of depressive symptoms prior to widowhood, as well as chronic symptoms to a limited degree, varied by the quality of marital relationships. Further, the patterns of change for each health indicator also showed substantial differences depending on older widow(er)s’ preloss marital quality.

Descriptive analyses of the study sample indicated that, the majority of the sample was identified as the supportive group (58%). Respondents in this group were more likely to be healthy, non-Hispanic White, and men with a higher level of education attainment and household income. This group was more likely to have a healthy spouse (e.g., lower numbers of depressive symptoms and chronic conditions) and to receive spousal help, but less likely to provide help to a spouse. The smallest group in the study sample was the ambivalent group (8%) that can be characterized as being healthy, non-Hispanic White, and female with a higher level of education and income. They reported a lower level of spousal help exchange patterns, partly attributable to relatively good health status of the respondents and their spouses in the ambivalent group. Compared to those in other groups, respondents in this group were also more likely to be in a first marriage with the longest history of marital partnership.

Next, respondents with aversive MQ (20%) were the youngest and more likely to have racial/ethnic minority background and the lowest household income. Respondents and their spouse in this group reported poorer health. This group reported the shortest length of current marriage and the greatest proportion of respondents were remarried, compared to other groups. Lastly, the neutral group (14%) was the oldest with a relatively low level of income. Respondents and their spouse in this group reported poorer health, showing the lowest rate of receiving spousal help, but the highest rate of providing help. The descriptive characteristic of each group suggests that older couples’ marital quality are closely related to the differences in socioeconomic status, spousal care exchange patterns, and health (Choi & Marks, 2008; Grundy & Holt, 2001; Martire et al., 2003).

Prior to spousal loss, the supportive and ambivalent groups reported fewer number of depressive symptoms compared to the other two, neutral and aversive. The aversive group, in particular, showed the highest number of depressive symptoms, followed by the neutral group. Although we found no significant difference in baseline chronic conditions across these groups, this finding in psychological health is in line with the literature on marital relationships and health, which consistently indicates that marital strain is associated with poor health outcomes (Umberson et al., 2006), whereas positive marital assessments are associated with better health outcomes for married individuals (Abakoumkin et al., 2010; Carr et al., 2014).

The differences in the link between marital quality and health indicators became more pronounced when changes in health were examined in the regression models. First, results examining change scores in depressive symptoms indicated that individuals who had ambivalent and supportive marital quality showed an increase in distress levels as they transitioned into widowhood, whereas such changes were not observed among aversive and neutral MQ types. The regression models using different reference group confirmed this (Supplementary Table 1). This is consistent with earlier studies which suggest a greater level of postloss distress among those who had positive marital relations during marriage, compared to those who did not (Abakoumkin et al., 2010; Carr et al., 2000; Futterman et al., 1990). Findings suggest that the presence of high positivity may have a greater effect than the absence of negativity in explaining the differences in changes in psychological distress between older adults with ambivalent and neutral relationships. One possible explanation for the pronounced increase in the number of depressive symptoms in widowhood among those with ambivalent MQ is that spousal loss could also mean a loss of resource for intensive social engagement (Stroebe et al., 2007). Ambivalent MQ is often observed among those with high levels of marital engagement, companionship, or closeness involving both high positivity and negativity (i.e., high ambivalence), especially for those older couples in a long-term marriage. Therefore, the pronounced depressive symptoms observed in this group after loss may have been driven by disruptions in everyday routines and lifestyle developed over time with the deceased spouse.

Also noteworthy was the clear indication of a high level of psychological distress in older adults with aversive and neutral MQ prior to widowhood, but no subsequent change in psychological distress in widowhood. No sign of improvement in psychological health was somewhat puzzling because spousal loss is often considered to mark an end to a stressful life circumstance in cases where the context of the marital relationship was itself a source of distress. It is possible that increased psychological distress experienced in widowhood due to various disadvantages associated with aversive (and neutral) MQ may also have cancelled out the relief experienced due to the end of stressful life circumstances, in this case, a strained marriage. Relatedly, spousal loss can also mark an end to caregiving-related burdens and stressors, including suffering and pain of a severely ill spouse, which in turn can result in feelings of relief or improvements in psychological well-being (Bonanno et al., 2004; Pruchno et al., 2009; Schulz et al., 2003). Such possibilities were partly tested in an elaboration analysis, where we used spousal help provision as well as spouse’s disability status (e.g., number of ADL limitations) as additional covariates (Carr, 2004; Wickrama et al., 1997). Results showed that provision of spousal help and a higher level of spouse’s predeath disability were associated with a greater net decrease in depressive symptoms after loss; however, the association between marital quality types and psychological well-being remained largely unchanged.

Our findings further indicated that respondents experienced aggravated physical health regardless of the quality of marital relationships. However, the magnitude of the increase in the number of chronic conditions was more pronounced among persons reporting aversive marital relationships, compared to those with supportive relationships. Such deterioration in physical health may in part be attributable to the disadvantages associated with individuals classified as having aversive marital relationships in our sample. Notably, respondents reporting this type of strained relationship were the most likely to rate their subjective health as fair or poor at baseline, despite being the youngest of all groups. This suggests that the socioeconomic status-health gradient widely reported among older individuals may continue into widowhood (Grundy & Holt, 2001). Yet, regression models in our study showed no significant associations of these covariates with changes in health.

We therefore conclude that spousal loss does not necessarily take away the risk of experiencing greater physical health declines in widowhood for individuals with aversive marital relations, nor does it alleviate the psychological distress associated with widowhood. There may be factors other than marital quality that could account for the difference in changes in physical health across the marital quality groups. For instance, previous research identified physiological pathways (e.g., inflammation, ambulatory blood pressure, health regulation, health behaviors) by which marital quality may influence physical health over time (Birmingham et al., 2015; Donoho et al., 2013). Further studies are needed to confirm these possibilities, in part by comparing those reporting aversive marital quality who became widowed to their matched counterparts.

Limitations

Our study has several limitations. First, respondents differed in terms of the duration of widowhood experienced, but such differences could not be taken into consideration in the analyses because spouse’s exact date of death was not consistently available across respondents. Although we limited the study sample to persons who had experienced widowhood within a 2-year time-frame, we were not able to differentiate respondents based on the lengths of widowhood, which is known to be associated with psychological and physical health outcomes (Stroebe et al., 2007). In additional analysis with a smaller sample size (only including those with the information on the spouse’s date of death, n = 352), the duration of widowhood attenuated a negative impact of spousal loss in changes in depressive symptoms but no such influence was found in the model examining changes in physical health in widowhood (results available from the authors upon request).

Second, spousal support and caregiving-related information was not included in the main analyses due to the substantial missing data (5.7%) in the related variables in the HRS. Earlier studies have indicated that caregiving-related burden can play a role in health outcomes among widowed persons (Bonanno et al., 2004; Schulz et al., 2003). In additional analyses, we found that provision of spousal support was associated with a net decrease in depressive symptoms, but no association with changes in chronic conditions. Third, we observed no significant gender differences in our sample. However, gender differences in the consequences of widowhood merit additional attention in future studies. This is because widowhood is more common in women due to longer life expectancies and age difference between married couples (Carr & Bodnar-Deren, 2009), and also because marital quality may differentially affect men and women during marriage and in widowhood (Carr, 2004).

Finally, our results must be understood in the context of how marital quality was conceptualized and measured in this study. Although the concept of ambivalence is frequently employed in the research literature, there is no consensus on how best to measure ambivalence (Connidis, 2015; Suitor et al., 2011; Uchino et al., 2004). For instance, some scholars directly ask respondents whether they have ambivalent feelings toward specific persons (direct measure), whereas others combine positive and negative relationship quality measures (indirect measure). Further studies using different measurement strategies are needed to better understand the relationship quality of older couples and its association with late-life health and well-being.

Conclusion

This study contributes to the current literature on widowhood and its effects on health by elaborating on the role of the quality of preloss marital relationships. Our findings showed that, within the first 2 years of the loss, psychological adjustment to widowhood seems to be more challenging for older adults who had a high level of positivity in preloss marital quality (i.e., supportive and ambivalent) than for those did not (i.e., aversive and neutral), whereas those with aversive marital quality may face a greater risk of physical health vulnerability. Using four marital quality typologies that take into consideration positive and negative aspects of marital quality helped to capture the complexity of spousal relationships during marriage and their effect on psychological and physical health during the transition to widowhood in later life. Social support and intervention efforts directed to older individuals experiencing aversive marital relations may not only alleviate problems associated with the marriage, but also have far-reaching influences for psychological and physical well-being in widowhood.

Supplemental Material

Supplemental Material, sj-pdf-1-roa-10.1177_0164027521989083 - Psychological and Physical Health in Widowhood: Does Marital Quality Make a Difference?

Supplemental Material, sj-pdf-1-roa-10.1177_0164027521989083 for Psychological and Physical Health in Widowhood: Does Marital Quality Make a Difference? by Hyo Jung Lee, Sae Hwang Han and Kathrin Boerner in Research on Aging

Footnotes

Declaration of Conflicting Interests

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

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by grant, P30AG066614, awarded to the Center on Aging and Population Sciences at The University of Texas at Austin by the National Institute on Aging, and by grant, P2CHD042849, awarded to the Population Research Center at The University of Texas at Austin by the Eunice Kennedy Shriver National Institute of Child Health and Human Development. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

ORCID iD

Hyo Jung Lee

Supplemental Material

Supplemental material for this article is available online.

References

Abakoumkin

Stroebe

(2010). Does relationship quality moderate the impact of marital bereavement on depressive symptoms? Journal of Social and Clinical Psychology, 29(5), 510–526. https://doi.org/10.1521/jscp.2010.29.5.510

Andresen

E. M.

Malmgren

J. A.

Carter

W. B.

Patrick

D. L.

(1994). Screening for depression in well older adults: Evaluation of a short form of the ces-d. American Journal of Preventive Medicine, 10(2), 77–84.

Birmingham

W. C.

Uchino

B. N.

Smith

T. W.

Light

K. C.

Butner

(2015). It’s complicated: Marital ambivalence on ambulatory blood pressure and daily interpersonal functioning. Annals of Behavioral Medicine, 49(5), 743–753.

Bonanno

G. A.

Wortman

C. B.

Lehman

D. R.

Tweed

R. G.

Haring

Sonnega

Carr

Nesse

R. M.

(2002). Resilience to loss and chronic grief: A prospective study from preloss to 18-months postloss. Journal of Personality and Social Psychology, 83(5), 1150–1164.

Bonanno

G. A.

Wortman

C. B.

Nesse

R. M.

(2004). Prospective patterns of resilience and maladjustment during widowhood. Psychology and Aging, 19(2), 260–271.

Bowlby

(1980). Attachment and loss (Vol. 3). Basic books.

Carr

(2004). Gender, preloss marital dependence, and older adults’ adjustment to widowhood. Journal of Marriage and Family, 66(1), 220–235.

Carr

Bodnar-Deren

(2009). Gender, aging and widowhood. In Uhlenberg

(Ed.), International handbook of population aging (Vol. 1, pp. 705–728). Springer.

Carr

Boerner

(2009). Do spousal discrepancies in marital quality assessments affect psychological adjustment to widowhood? Journal of Marriage and Family, 71(3), 495–509.

10.

Carr

Freedman

V. A.

Cornman

J. C.

Schwarz

(2014). Happy marriage, happy life? Marital quality and subjective well-being in later life. Journal of Marriage and Family, 76(5), 930–948.

11.

Carr

House

J. S.

Kessler

R. C.

Nesse

R. M.

Sonnega

Wortman

(2000). Marital quality and psychological adjustment to widowhood among older adults a longitudinal analysis. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 55(4), S197–S207.

12.

Carr

Utz

(2001). Late-life widowhood in the United States: New directions in research and theory. Ageing International, 27(1), 65–88. https://doi.org/10.1007/s12126-001-1016-3

13.

Choi

Marks

N. F.

(2006). Transition to caregiving, marital disagreement, and psychological well-being: A prospective U.S. National study. Journal of Family Issues, 27(12), 1701–1722. https://doi.org/10.1177/0192513x06291523

14.

Choi

Marks

N. F.

(2008). Marital conflict, depressive symptoms, and functional impairment. Journal of Marriage and Family, 70(2), 377–390.

15.

Choi

K. H.

Vasunilashorn

(2014). Widowhood, age heterogamy, and health: The role of selection, marital quality, and health behaviors. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 69B(1), 123–134.

16.

Choi

N. G.

J.-H.

(2011). Relationship between spouse/partner support and depressive symptoms in older adults: Gender difference. Aging & Mental Health, 15(3), 307–317.

17.

Clarke

Fisher

House

Smith

Weir

(2007). Guide to content of the hrs psychosocial leave-behind participant lifestyle questionnaires: 2004 & 2006. https://hrs.isr.umich.edu/publications/biblio/5723

18.

Connidis

I. A.

(2015). Exploring ambivalence in family ties: Progress and prospects. Journal of Marriage and Family, 77(1), 77–95. https://doi.org/10.1111/jomf.12150

19.

Donoho

C. J.

Crimmins

E. M.

Seeman

T. E.

(2013). Marital quality, gender, and markers of inflammation in the midus cohort. Journal of Marriage and Family, 75(1), 127–141.

20.

Fayers

P. M.

Sprangers

M. A. G.

(2002). Understanding self-rated health. The Lancet, 359(9302), 187–188. https://doi.org/10.1016/S0140-6736(02)07466-4

21.

Fingerman

K. L.

Hay

E. L.

Birditt

K. S.

(2004). The best of ties, the worst of ties: Close, problematic, and ambivalent social relationships. Journal of Marriage and Family, 66(3), 792–808. https://doi.org/10.2307/3600228

22.

Fitzmaurice

(2001). A conundrum in the analysis of change. Nutrition, 17(4), 360–361.

23.

Freedman

V. A.

Cornman

J. C.

Carr

(2014). Is spousal caregiving associated with enhanced well-being? New evidence from the panel study of income dynamics. The Journals of Gerontology: Series B, 69(6), 861–869. https://doi.org/10.1093/geronb/gbu004

24.

Futterman

Gallagher

Thompson

L. W.

Lovett

Gilewski

(1990). Retrospective assessment of marital adjustment and depression during the first two years of spousal bereavement. Psychology and Aging, 5, 277–283.

25.

Grundy

Holt

(2001). The socioeconomic status of older adults: How should we measure it in studies of health inequalities? Journal of Epidemiology and Community Health, 55(12), 895–904. https://doi.org/10.1136/jech.55.12.895

26.

Itzhar-Nabarro

Smoski

M. J.

(2012). A review of theoretical and empirical perspectives on marital satisfaction and bereavement outcomes: Implications for working with older adults. Clinical Gerontologist, 35(3), 257–269.

27.

Lillard

L. A.

Waite

L. J.

(1995). Til death do us part: Marital disruption and mortality. American Journal of Sociology, 1131–1156.

28.

Lopata

H. Z.

(1981). Widowhood and husband sanctification. Journal of Marriage and Family, 43(2), 439–450. https://doi.org/10.2307/351394

29.

Lüscher

Pillemer

(1998). Intergenerational ambivalence: A new approach to the study of parent-child relations in later life. Journal of Marriage and Family, 60(2), 413–425.

30.

Mancini

A. D.

Robinaugh

Shear

Bonanno

G. A.

(2009). Does attachment avoidance help people cope with loss? The moderating effects of relationship quality. Journal of Clinical Psychology, 65(10), 1127–1136. https://doi.org/10.1002/jclp.20601

31.

Martire

L. M.

Schulz

Wrosch

Newsom

J. T.

(2003). Perceptions and implications of received spousal care: Evidence from the caregiver health effects study. Psychology and Aging, 18(3), 593–601. https://doi.org/10.1037/0882-7974.18.3.593

32.

Mikulincer

Shaver

P. R.

(2008). An attachment perspective on bereavement. In Stroebe

M. S.

Hansson

R. O.

Schut

Stroebe

(Eds.), Handbook of bereavement research and practice: Advances in theory and intervention (Vol. xiv, pp. 87–112). American Psychological Association.

33.

Ott

C. H.

Lueger

R. J.

Kelber

S. T.

Prigerson

H. G.

(2007). Spousal bereavement in older adults: Common, resilient, and chronic grief with defining characteristics. The Journal of Nervous and Mental Disease, 195(4), 332–341.

34.

Parkes

C. M.

(1988). Bereavement as a psychosocial transition: Processes of adaptation to change. Journal of Social Issues, 44(3), 53–65.

35.

Parkes

C. M.

Weiss

R. S.

(1983). Recovery from bereavement. Basic Books.

36.

Pearlin

L. I.

Menaghan

E. G.

Lieberman

M. A.

Mullan

J. T.

(1981). The stress process. Journal of Health and Social Behavior, 22(4), 337–356. https://doi.org/10.2307/2136676

37.

Prigerson

H. G.

Maciejewski

P. K.

Rosenheck

R. A.

(2000). Preliminary explorations of the harmful interactive effects of widowhood and marital harmony on health, health service use, and health care costs. The Gerontologist, 40(3), 349–357.

38.

Pruchno

R. A.

Cartwright

F. P.

Wilson-Genderson

(2009). Effects of marital closeness on the transition from caregiving to widowhood. Aging and Mental Health, 13(6), 808–817.

39.

Rendall

M. S.

Weden

M. M.

Favreault

M. M.

Waldron

(2011). The protective effect of marriage for survival: A review and update. Demography, 48(2), 481–506.

40.

Sandberg

J. G.

Harper

J. M.

Hill

Miller

R. B.

Yorgason

J. B.

Day

R. D.

(2013). “What happens at home does not necessarily stay at home”: The relationship of observed negative couple interaction with physical health, mental health, and work satisfaction. Journal of Marriage and Family, 75(4), 808–821. https://doi.org/10.1111/jomf.12039

41.

Sasson

Umberson

D. J.

(2014). Widowhood and depression: New light on gender differences, selection, and psychological adjustment. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 69B(1), 135–145.

42.

Schaan

(2013). Widowhood and depression among older Europeans—The role of gender, caregiving, marital quality, and regional context. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 68(3), 431–442. https://doi.org/10.1093/geronb/gbt015

43.

Schulz

Mendelsohn

A. B.

Haley

W. E.

Mahoney

Allen

R. S.

Zhang

Thompson

Belle

S. H.

(2003). End-of-life care and the effects of bereavement on family caregivers of persons with dementia. New England Journal of Medicine, 349(20), 1936–1942.

44.

Servais

M. A.

(2010). Overview of hrs public data files for cross-sectional and longitudinal analysis. https://hrs.isr.umich.edu/publications/biblio/5835

45.

Stroebe

Schut

Stroebe

(2007). Health outcomes of bereavement. The Lancet, 370(9603), 1960–1973. https://doi.org/10.1016/S0140-6736(07)61816-9

46.

Suitor

J. J.

Gilligan

Pillemer

(2011). Conceptualizing and measuring intergenerational ambivalence in later life. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 66(6), 769–781. https://doi.org/10.1093/geronb/gbr108

47.

Thoits

P. A.

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

48.

Thompson

M. M.

Zanna

M. P.

Griffin

D. W.

(1995). Let’s not be indifferent about (attitudinal) ambivalence. In Petty

R. E.

Krosnick

J. A.

(Eds.), Attitude strength: Antecedents and consequences (pp. 361–386). Erlbaum.

49.

Uchino

B. N.

Holt-Lunstad

Smith

T. W.

Bloor

(2004). Heterogeneity in social networks: A comparison of different models linking relationships to psychological outcomes. Journal of Social and Clinical Psychology, 23(2), 123–139.

50.

Umberson

Williams

(2005). Marital quality, health, and aging: Gender equity? Journals of Gerontology B: Psychological and Social Sciences, 60B(Special Issue II), 109–112.

51.

Umberson

Williams

Powers

D. A.

Liu

Needham

(2006). You make me sick: Marital quality and health over the life course. Journal of Health and Social Behavior, 47(1), 1–16.

52.

Walker

R. B.

Luszcz

M. A.

(2009). The health and relationship dynamics of late-life couples: A systematic review of the literature. Ageing and Society, 29(03), 455–480.

53.

Whisman

M. A.

Uebelacker

L. A.

(2006). Impairment and distress associated with relationship discord in a national sample of married or cohabiting adults. Journal of Family Psychology, 20(3), 369–377. https://doi.org/10.1037/0893-3200.20.3.369

54.

Wickrama

K. A. S.

Lorenz

F. O.

Conger

R. D.

Elder

G. H.

(1997). Marital quality and physical illness: A latent growth curve analysis. Journal of Marriage and Family, 59(1), 143–155. https://doi.org/10.2307/353668

55.

Wilcox

Evenson

K. R.

Aragaki

Wassertheil-Smoller

Mouton

C. P.

Loevinger

B. L.

(2003). The effects of widowhood on physical and mental health, health behaviors, and health outcomes: The women’s health initiative. Health Psychology, 22(5), 513–522.

56.

Williams

(2004). The transition to widowhood and the social regulation of health: Consequences for health and health risk behavior. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 59(6), S343–S349.

57.

Windsor

T. D.

Butterworth

(2010). Supportive, aversive, ambivalent, and indifferent partner evaluations in midlife and young-old adulthood. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 65(3), 287–295.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.15 MB