Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Dec 1.
Published in final edited form as: Pers Relatsh. 2022 Sep 9;29(4):933–955. doi: 10.1111/pere.12443

Loneliness and HbA1c among older Irish couples: Retirement as life stage context

Jeffrey E Stokes 1, Adrita Barooah 1
PMCID: PMC9799165  NIHMSID: NIHMS1831411  PMID: 36588975

Abstract

Loneliness is an important determinant of health and mortality among the aging population, including for cardiometabolic health. Yet research has largely focused on individual experiences of loneliness, rather than taking intimate relationships into account. However, recent studies have highlighted that loneliness of a partner may have implications for one’s own health and well-being as well. Indeed, this is particularly true among older couples, as life events and shifting time horizons (e.g., retirement, reduced social networks) can lead to greater prominence and salience of one’s closest relationships. This study uses dyadic structural equation modeling (SEM) to estimate actor-partner interdependence models (APIM) examining associations between loneliness and HbA1c levels among 1331 older married couples from The Irish Longitudinal Study on Aging (TILDA, 2009–2011). Furthermore, we test whether any such actor or partner effects vary by employment status. Results indicated that one’s own loneliness was not significantly linked with HbA1c, irrespective of employment status. However, loneliness of a dyadic partner was significantly associated with elevated HbA1c among retired persons only. These findings underscore that relationship and life course context are crucial when considering the dyadic implications of loneliness for health among the older population.

Keywords: dyadic data analysis, health, interdependence, life-events/lifespan, loneliness, marriage

1 |. INTRODUCTION

Loneliness reflects the disparity between individuals’ desired and perceived quality and quantity of social relationships (de Jong-Gierveld, 1987). Loneliness is also distinct from solitude or isolation, which reflect the absence of social relationships or contact; that is, not every isolated person is lonely, nor is every lonely person isolated (e.g., Dykstra et al., 2005; Perissinotto et al., 2012; Stokes, 2017a). Indeed, evidence suggests that loneliness is driven more by the perceived quality than by the sheer number of one’s intimate and social relationships (Ayalon et al., 2013; de Jong-Gierveld, 1987; Moorman, 2016; Stokes, 2017a, 2017b). Although marriage protects against loneliness overall, married persons experience loneliness as well, particularly when the perceived quality of the marital relationship is poor (Perissinotto et al., 2012; Stokes, 2017a, 2017b). That is, loneliness is relational; it is not merely influenced by the characteristics of one’s relationships, but it can affect those relationships and their characteristics as well, particularly among married older couples (Ayalon et al., 2013; Cacioppo et al., 2009; Moorman, 2016; Stokes, 2017a, 2017b).

Additionally, loneliness increases during later life, though the extent varies by national context (Luhmann & Hawkley, 2016; Yang & Victor, 2011). In previous studies concerning older Irish couples, loneliness increased somewhat over time, especially among wives (e.g., Stokes, 2017b). Moreover, though, the role and centrality of the marital relationship changes in later life, as well, making the relational influences on – and consequences of – loneliness of particular interest among aging couples. Socioemotional selectivity theory (SST; Carstensen et al., 1999) posits that as individuals age, and particularly as they move through normative life stages such as marrying, having children, and leaving the workforce, their time horizons, social goals, and thus their social behaviors change as well. In late life and after retirement, older adults’ final horizon is their own mortality. To maximize well-being during this last stage of life, older adults focus their time and energy on a smaller number of their closest, most emotionally meaningful and fulfilling relationships, including and especially with their spousal partner (Carstensen et al., 1999). Building upon prior research linking the experience of loneliness with older adults’ own (e.g., Donovan et al., 2017; Holt-Lunstad et al., 2015; Luo et al., 2012) and their partner’s (e.g., Stokes & Barooah, 2021) physical health, the present study aims to not only situate loneliness among the older population within a dyadic, relational context, but will also examine the extent to which retirement status – above and beyond age itself – may shape the potential health implications of dyadic loneliness for older adults’ health.

1.1 |. LONELINESS AND HEALTH WITHIN RELATIONSHIPS

Beyond being an unpleasant emotional state or experience, loneliness is also a known contributor to health declines across various aspects of mental and physical health, including greater depression and depressive symptoms, increased functional limitations, impaired cognition, and increased risk of mortality (Cacioppo et al., 2006; Donovan et al., 2017; Holt-Lunstad et al., 2015; Luo et al., 2012). As noted previously, loneliness is not limited to those who are socially isolated, and is often driven by perceived deficits in the quality of one’s social and intimate relationships instead (e.g., Dykstra et al., 2005; Perissinotto et al., 2012). The health effects of loneliness are likewise distinct from those of social isolation, and loneliness itself is not uncommon even among married older adults (Holt-Lunstad et al., 2015; Perissinotto et al., 2012).

Indeed, research has increasingly highlighted that loneliness is not merely an individual experience, but instead is affected by – and, in turn, influences – relationship, family, and social network dynamics (e.g., Cacioppo et al., 2009; Segrin et al., 2012; Stokes, 2017b). For instance, loneliness displays genetic heritability within families, yet beyond that is also affected by sociodemographic characteristics, as well as by family, friend, and spousal support (Distel et al., 2010; Moorman, 2016; Segrin et al., 2012). Older spouses’ “shared” loneliness may also partly reflect assortative mating, as individuals with similar backgrounds, sociodemographic characteristics, and personality traits are likely to partner with one another (Distel et al., 2010). In addition to the potential self-selection of lonelier or less lonely individuals into relationships with similar partners, research has repeatedly identified aspects of relationships that can provoke the experience of loneliness among partners (e.g., Burke & Segrin, 2014; Distel et al., 2010; Moorman, 2016; Stokes, 2017a, 2017b). Among the most consistent findings is that a lack of perceived support from social partners, including a spouse, can lead to increased loneliness (e.g., Ayalon et al., 2013; Moorman, 2016; Segrin et al., 2012; Stokes, 2017a). Moreover, perceived strain from a spouse has been associated both cross-sectionally and longitudinally with increased loneliness among older couples (e.g., Ayalon et al., 2013; Moorman, 2016; Stokes, 2017a, 2017b).

Loneliness is not merely an outcome of relationship stress, however. It is itself a stressor, with potential impacts on health and well-being for both partners in a relationship. In addition to links with individuals’ own depressive symptoms, well-being, and health, loneliness has also been linked with dyadic partners’ loneliness, both cross-sectionally and longitudinally (e.g., Ayalon et al., 2013; Moorman, 2016; Stokes, 2017a, 2017b). That is, having a lonely partner is a risk factor for experiencing loneliness oneself (Stokes, 2017b). Loneliness is associated not only with inferior health behaviors, but also with poorer interpersonal communication, with lonely persons often acting more anxious, negative, and hostile to their social partners (Berscheid & Reis, 1998; Cacioppo et al., 2006, 2009). This can lead to further deficits in already imperfect marital relationships, and provoke greater loneliness among spousal partners (Stokes, 2017b). For these same reasons, loneliness in either partner in a relationship may be a “stress generating” experience (e.g., Hammen, 2006). That is, having a lonely partner can itself be quite stressful, and this stress can lead to both inferior health behaviors, and to physiological responses to stress, which then impair physical health over time (Hammen, 2006; Stokes & Barooah, 2021). In accordance with the stress generation hypothesis, we anticipate that the experience of loneliness – in either partner within a couple – may lead to worse health outcomes for both partners.

1.2 |. RETIREMENT AND RELATIONSHIPS: THE PROCESS OF AGING AND RE-ENGAGING

The marital relationship, while often a “constant” in adults’ lives, does not remain constant throughout the life course. Rather, relationships and their importance fluctuate over time and across the age range, and in particular following major life stage events and transitions (Bookwala, 2012). In particular, an SST approach to relationships emphasizes the importance of age and future time perspectives for individuals’ and couples’ social goals and behaviors (Carstensen et al., 1999; Lang & Carstensen, 2002). As individuals enter into later life, they typically leave the work force, reduce their number of social partners, and focus their time and attention on the closest and most meaningful relationships in their life, and in particular their marital relationship (Carr et al., 2014; Carstensen et al., 1999; Mancini & Bonanno, 2006; Stokes, 2017a, 2017b; Umberson et al., 2006). Indeed, retirement is a major life event in this process, as it represents not only the cessation of numerous social relationships outside the home (i.e., workplace friendships and interactions), but can also lead to a large increase in the amount of time shared together by older spouses (Comi et al., 2022; Kulik, 2001; Wrzus et al., 2013). Importantly, the process of SST and older adults’ future time perspectives are not simply a function of age. Instead, we argue that they are likewise a function of life stage transitions, and especially of retirement.

Employment status is not typically considered a characteristic of one’s intimate relationships, as it refers to the individual’s labor force participation, yet it can shape those relationships in numerous ways. For instance, employment outside the home serves as an opportunity to build friendships and relationships with a wider social network, to interact on a regular basis with others who are often from different age groups or social backgrounds, and to find a sense of self-worth and identity that is removed from the home, family, or marriage (Börsch-Supan & Schuth, 2014; Rumens, 2016; Teuscher, 2010; van Solinge & Henkens, 2008).

More pertinently, retirement – as opposed to temporary unemployment, or longer term experience as self-employed or homemaker – reflects a decision to remove oneself permanently from the work force and the social involvement it entails, and generally occurs as part of the broader life stage process of socioemotional selectivity (Carstensen et al., 1999). The SST framework views individuals’ social preferences and behaviors as a reflection of their immediate and longer term goals, which vary based on life stage and future time horizons. At younger stages of the life course, it may be more advantageous to have a large network of “weak ties”, such that one can draw on those links for career advancement opportunities, or for entrée into other social groups, or even as a way to meet new social – and intimate – partners (Carstensen, 2006; Carstensen et al., 1999; Lang & Carstensen, 2002). At later stages of the life course, however, individuals’ goals and future time perspectives shift; rather than looking ahead to potential career changes or advancement, or to new relationships and experiences, older adults view their remaining time as more limited and shift their energy and attention to the closest and most emotionally rewarding relationships in their lives, particularly their family and spousal relationships (Carstensen et al., 2003; Lang & Carstensen, 2002).

As a result of this process of socioemotional selectivity, the marital relationship in particular may take on greater significance – and exert greater influence – later in life (Story et al., 2007). Yet this is not simply a function of age; rather, it is a function of life stage. Perceptions and decisions about the future depend less on chronological age than on individuals’ time horizons, that is, how individuals view their future in terms of time, goals, and expectations (Carstensen, 2006; Lang & Carstensen, 2002; Löckenhoff, 2012; Löckenhoff & Carstensen, 2004). Chronological age has the clearest impact on perceived future time horizons, yet significant life events such as retirement, family births and deaths, and health shocks can also alter perceived time horizons and goals (Fung & Carstensen, 2006; Löckenhoff, 2012). Retirement in particular represents a final removal from the labor force – and its social connections and interactions – in exchange for leisure and other activities, including spending time with loved ones such as a spouse (Butrica & Schaner, 2005; Comi et al., 2022).

The present study focuses on the unique case of retirement in Ireland. In recent decades, Ireland has seen a shift away from “welfare” and towards “workfare”, that is, towards less generous social welfare programs supported by the government (Vis, 2007). This move toward austerity has also contributed to the gender pension gap in Ireland (Ní Léime et al., 2015). Additionally, the Irish government has set incremental increases to the retirement age, in order to discourage early retirement and entrance into the pension system (e.g., Barrett & Mosca, 2013). Taken alongside the severe economic recession in 2008, when Ireland’s housing bubble crashed and the national banking system broke down (Whelan, 2014), these changes speak to the changing nature of retirement in Ireland: A life course transition that is further delayed for many, due to lost savings and pension system changes, but which can occur earlier – whether voluntarily or involuntarily – based on factors such as financial capacity, as well as health or health problems (e.g., Barrett & Mosca, 2013; Mosca & Barrett, 2014). Indeed, those who retire involuntarily, due to poor health, report worse mental health than those who retire by choice (Mosca & Barrett, 2014), which may have further implications for the health and well-being of their dyadic partners (Hammen, 2006; Stokes, 2017a; Stokes & Barooah, 2021).

Despite these recent trends and changes, however, among the young-old (ages 65–74) in Ireland, the majority of both men (82%) and women (52%) describe themselves as retired, with a larger proportion of women (36%) reporting that they are “looking after home or family” (Barrett et al., 2011). Moreover, a recent study found that among adults aged 50 and older in Ireland who transitioned into retirement between 2009 and 2013, the vast majority (79%) did so voluntarily, with approximately 9% reporting poor health as a reason for retirement (Mosca & Barrett, 2014). Among the comparatively older and longer retired sample used in the present study, approximately 14% of husbands and 21% of wives who were retired listed ill health among the reasons. These cases may be unique, insofar as the selection into retirement and the changes to future time horizons and intimate relationships may be thrust upon older participants in poor health, rather than freely chosen, with potential implications for both well-being and health. Moreover, the examination of retirement status and loneliness within older couples allows for an examination of the ways in which factors such as age, ill health, employment status, and well-being are intricately intertwined, not only for individual older adults themselves (e.g., Mosca & Barrett, 2014), but for their dyadic partners as well (e.g., Stokes & Barooah, 2021).

1.3 |. HBA1C AS A BIOLOGICAL OUTCOME OF LONELINESS

Research concerning the health implications of loneliness has utilized a number of different measures of health, including depression and depressive symptoms, impaired sleep, cognitive and functional decline, and even mortality (Cacioppo et al., 2006; Donovan et al., 2017; Hawkley et al., 2010; Holt-Lunstad et al., 2015; Luo et al., 2012; O’Luanaigh & Lawlor, 2008; Perissinotto et al., 2012). Studies focusing on the potential pathways or mechanisms of loneliness’ general effects on health have largely focused on cardiometabolic health (e.g., Hawkley et al., 2010; Hawkley & Capitanio, 2015; O’Luanaigh et al., 2012; Valtorta et al., 2016; Xia & Li, 2018). This has included analysis of loneliness and HDL, LDL, and total cholesterol; metabolic syndrome; systolic blood pressure; and HbA1c, among others (e.g., Das, 2019; Hawkley & Cacioppo, 2010; O’Luanaigh et al., 2012; Stokes & Barooah, 2021; Whisman, 2010). This is due to loneliness’ role as a psychosocial stressor, which may lead to heightened stress reactivity or abnormal stress responses that can be reflected by impaired cardiometabolic health (see, e.g., Das, 2019; O’Luanaigh et al., 2012).

HbA1c itself is a measure of glycemic control and an indicator of diabetes; it is also a risk factor for diabetes-related disorders, including cardiovascular disease, and higher levels of HbA1c have been associated with increased risk of cardiovascular disease among persons both with and without diabetes (Santos-Oliveira et al., 2011). Moreover, although studies concerning loneliness and cardiometabolic health in general have produced somewhat mixed results across different biological markers and across different populations (see, e.g., Stokes & Barooah, 2021 for discussion of this), studies examining HbA1c in particular have identified significant associations of loneliness with elevated HbA1c among older adults in Ireland (O’Luanaigh et al., 2012) as well as with older adults and their spouses’ HbA1c levels in the U.S., although these latter associations appear contingent upon relational context (Stokes & Barooah, 2021). Taken together, this emerging literature suggests that (a) loneliness can have cardiometabolic implications for older persons, including older married adults; (b) HbA1c is a marker of cardiometabolic health that appears sensitive to the experience of loneliness as a stressor; and (c) the cardiometabolic implications of loneliness may materialize only under certain relational circumstances.

1.4 |. MARRIAGE AND AGING IN IRELAND

Research to date concerning loneliness among married older adults in Ireland remains rather scant. Yet Ireland offers an intriguing setting for examining the implications of loneliness for dyadic partners, particularly in comparison with U.S. and Western European samples. While loneliness varies across national contexts, these differences are mostly due to marital, economic, and health characteristics of the respective countries (Fokkema et al., 2012). Ireland, however, has a very unique cultural history, particularly as concerns the institution of marriage (Kennedy, 1973). For instance, Ireland has the highest proportion of adults in Western Europe who have never married or who married later in life (Kamiya et al., 2013, 2014; Kamiya & Sofroniou, 2011). Moreover, divorce was only legalized in Ireland in 1996, making it particularly infrequent in comparison with peer countries (Kamiya et al., 2013, 2014; Kamiya & Sofroniou, 2011).

In addition, the current cohort of older adults in Ireland lived through what is known as the Irish “marriage boom” of the 1960s and 1970s, when provisions of the Irish Marriage Bar – that is, the requirement that women leave paid employment within certain sectors upon marriage – were progressively eliminated (see Mosca & Wright, 2018) and marriage rates and trends moved into greater alignment with countries such as the United States (Kamiya & Sofroniou, 2011). Relationships and relational influences among older married couples in Ireland, therefore, may differ from those in other comparable countries.

1.5 |. DYADIC PROBLEMS REQUIRE DYADIC SOLUTIONS

Loneliness is a relational experience, which is sensitive to relationship dynamics such as support and strain, but also influences the relationship and a spousal partner’s health and wellbeing, as well (e.g., Cacioppo et al., 2009; Segrin et al., 2012; Stokes, 2017b; Stokes & Barooah, 2021). In order to better understand the context and circumstances under which loneliness has consequences for both partners within marriages, dyadic rather than individual data and analyses are required. Dyadic data – that is, data collected from both partners within a relationship – allow for direct examination of the ways in which each partner’s characteristics, experiences, and emotions may affect both partners’ health and well-being. Dyadic research concerning loneliness and health specifically remains fairly limited, yet the existing literature – and the related literature concerning anxiety and depressive symptoms among older couples – offers general support for two main theoretical frameworks: First, that loneliness appears “contagious” among older couples, such that having a lonely partner is a risk factor for experiencing loneliness oneself (e.g., Ayalon et al., 2013; Stokes, 2017a; Stokes, 2017b; Stokes, 2017c; Thomeer et al., 2013). Second, that just as the experience of loneliness is a stressor with consequences for mental and physical health (e.g., Cacioppo et al., 2006; Holt-Lunstad et al., 2015; O’Luanaigh et al., 2012), so too is the experience of having a lonely partner a potential stressor, in keeping with the stress generation hypothesis (e.g., Berscheid & Reis, 1998; Cacioppo et al., 2009; Hammen, 2006; Stokes & Barooah, 2021).

Based upon prior research regarding loneliness and cardiometabolic health (e.g., Holt-Lunstad et al., 2015; O’Luanaigh et al., 2012), loneliness within older couples (e.g., Ayalon et al., 2013; Moorman, 2016; Stokes, 2017a, 2017b), and contextual relationship factors that may shape the influence of loneliness – and a partner’s loneliness – on health outcomes (e.g., Stokes & Barooah, 2021), and utilizing the theoretical frameworks of socioemotional selectivity theory (Carstensen et al., 1999), future time perspectives (e.g., Lang & Carstensen, 2002), and the stress generation hypothesis (Hammen, 2006), the present study uses nationally representative data from older married couples in Ireland to determine whether (a) own and/or dyadic partner’s loneliness predict HbA1c levels among older Irish spouses, and (b) employment status serves as a contextual factor that shapes the influence of (dyadic) loneliness on cardiometabolic health among older couples.

1.6 |. STUDY AIMS

The present analysis uses nationally representative data from 1331 opposite-sex couples drawn from the first wave of The Irish Longitudinal Study on Aging (TILDA). This study builds upon recent research concerning (1) loneliness and cardiometabolic health among older adults in Ireland (e.g., O’Luanaigh et al., 2012); (2) dyadic loneliness among older married couples in Ireland (e.g., Stokes, 2017a, 2017b); and (3) dyadic loneliness and cardiometabolic health among older married couples in the United States (e.g., Stokes & Barooah, 2021), and has two primary aims:

Aim 1:

To determine whether one’s own loneliness and/or a spouse’s loneliness are associated with elevated HbA1c among older married adults in Ireland.

Hypothesis 1. One’s own loneliness will be associated with increased levels of HbA1c.

Hypothesis 2. Partner’s loneliness will be associated with increased levels of HbA1c.

Aim 2:

To determine whether employment status impacts the association(s) of either or both spouses’ loneliness on HbA1c levels.

Hypothesis 3. Loneliness will be more strongly associated with HbA1c levels among the retired.

Hypothesis 4. Partner’s loneliness will be more strongly associated with HbA1c levels among the retired.

2 |. METHOD

2.1 |. Data & sample

Data for this study came from the initial (2009–2011) wave of The Irish Longitudinal Study on Aging (TILDA). TILDA recruited a nationally representative sample of 8504 adults aged 50 or over and their spouses/partners resident in Ireland at baseline (Barrett et al., 2011; Kenny et al., 2010). The study design for TILDA is based on those of other large national studies of aging, such as the Health and Retirement Study (HRS) and the English Longitudinal Study on Aging (ELSA) (Kamiya et al., 2014; Kenny et al., 2010). In the baseline wave of TILDA, 62% of eligible persons participated, and 83% of those interviewees also completed the supplemental questionnaire that included questions concerning loneliness (Kamiya et al., 2014). Additionally, 72% of interviewees provided a blood sample at baseline for collection of biological marker data (Cronin et al., 2013). There were a total of 1855 opposite-sex married couples where both spouses completed the interview and questionnaire. Of these, a total of 1331 couples (72%) also saw both spouses complete the blood sample draw, and have valid data for HbA1c. Because variables of interest for this study were drawn from the questionnaire and the blood sample data, the analytic sample was limited to these 1331 opposite-sex older married couples.

2.2 |. Measures

2.2.1 |. HbA1c

Glycosylated hemoglobin (HbA1c) measures glycemic control over the previous 8–12 weeks, and was analyzed on 1 ml Buffy coat aliquot samples of participants’ bloods using reversed-phase action exchange chromatography with an ADAMS A1c HA-8180 V analyzer (TILDA, 2019). To correct for significant positive skew, HbA1c was transformed using the inverse of the square root. Scores were then reversed for directionality, and standardized for ease of interpretation.

2.2.2 |. Loneliness

Loneliness was measured using the 3-item UCLA loneliness scale (Hughes et al., 2004). The three questions pertaining to loneliness were “How often do you feel isolated from others?”, “How often do you feel left out?”, and “How often do you feel you lack companionship?” Response categories ranged from 1 (Hardly ever or never) to 3 (Often). Husbands and wives evaluated their loneliness independently of one another. Loneliness was constructed as a mean-score scale, and was set to missing for participants who answered fewer than half of the three items. Cronbach’s alpha for the scale indicated good reliability (α = 0.80 for husbands, α = 0.81 for wives). The loneliness scale was transformed using the inverse of the square, to address significant positive skew. The scale was then reversed, such that higher scores indicated greater loneliness, and was standardized for ease of interpretation.

2.2.3 |. Employment and retirement status

Employment status was self-reported as of the time of interview, with response options for currently employed, retired (reference group), and other employment status. The most common employment status for both husbands (48%) and wives (40%) was currently employed, with 40% of husbands and 24% of wives reporting a retired status.

2.2.4 |. Age

Age was measured as a continuous variable, in years, and ranged in this sample from 49 to 80 years old. Age was an important factor to include not only because of its associations with both loneliness and employment/retirement status, but because it is a central component of SST. Its inclusion here allowed us to examine whether employment and retirement status itself served as a moderator, or whether any such effects were instead a function of age.

2.2.5 |. Marital support and strain

Participants were asked to answer a series of questions concerning their perceived support and strain from their spouse (Walen & Lachman, 2000). Sample items include “How much does [your spouse] really understand the way you feel about things?”(support) and “How much does [your spouse] get on your nerves?” (strain). Marital support and strain were measured separately for husbands and wives. Marital support was generated as a 3-item mean-score scale, displayed good reliability (α = 0.76 for husbands, α = 0.82 for wives), and was transformed using the inverse of the reversed square to correct significant negative skew. Marital support was then standardized for ease of interpretation. Marital strain was generated as a 4-item mean-score scale, also displayed good reliability (α = 0.75 for husbands, α = 0.77 for wives), and was transformed using the inverse to address significant positive skew. Marital strain was then reversed and standardized, for ease of interpretation.

2.2.6 |. Additional covariates

A variety of control measures were also included in the present analysis, to ensure the validity of results. Controls were included for income, living situation, number of children, region, education, and both own and partner’s depressive symptoms (see Burholt & Scharf, 2014; Dykstra et al., 2005; Pinquart & Sörensen, 2001; Thomeer et al., 2013; Victor et al., 2000). Moreover, because entrance into retirement may have different impacts on mental or physical health and well-being depending on the reasons and voluntariness of the decision to retire (e.g., Mosca & Barrett, 2014), we included a dichotomous indicator of whether retired participants reported “own ill health” as a reason for retirement. Participants who were not retired were coded as “0” on this indicator, in keeping with dummy variable adjustment (Allison, 2002). Results of interest were consistent whether these covariates were included or excluded from the models, underscoring the robustness of the results presented.

2.3 |. Analytic strategy and missing data

Dyadic structural equation modeling (SEM) was utilized to address our research aims. Models estimated both actor and partner effects of loneliness on both spouses’ HbA1c levels, in accordance with the actor-partner interdependence model (APIM; Kenny & Cook, 1999). SEM using APIM also explicitly accounts for the covariance of couples’ data not only on the independent variables, but also on the dependent variables. Figure 1 displays a conceptual diagram of the APIM estimated in this study.

FIGURE 1.

FIGURE 1

Open in a new tab

Conceptual model

Analysis began with a fully unconstrained APIM model, with all coefficients estimated independently for husbands and wives. Gender differences were then examined using Wald tests of coefficient equivalence. None of the coefficients differed significantly by gender; therefore, coefficients were constrained to equivalence for husbands and wives. Interaction terms were then tested between (a) own loneliness and employment status categories, and (b) partner’s loneliness and employment status categories. Again, coefficients for these interaction terms were first estimated freely for husbands and wives; Wald tests revealed no significant differences in any of the interaction term coefficients by gender, thus the interaction terms were constrained to equivalence for husbands and wives. Additionally, in order to determine whether results concerning employment status were merely a function of age, interaction terms of both (a) own loneliness, and (b) partner’s loneliness with age were examined. No significant moderation by age was detected; therefore these interactions are not shown (results available upon request). The final model included all predictors of interest, covariates, and significant interaction terms that were identified throughout this process.

Missing data were handled using full information maximum likelihood (FIML) methods. Overall, 1063 (80%) of the 1331 couples that comprised the sample had both partners provide complete data on all measures included in the analysis. Taken separately, 1220 (92%) of husbands and 1137 (85%) of wives provided complete data on all measures in the analysis. For both men and women, the variable with the greatest amount of missing data was income, which was missing for 5% of husbands and 12% of wives. Missing data diagnostics were performed, with no clear patterns of missingness identified. Therefore, FIML was used to protect against the potential for bias from listwise deletion. Data management was performed using Stata/SE Version 17. Final SEM analyses were performed using Mplus Version 8.4.

3 |. RESULTS

Descriptive statistics for all measures included in the analysis are reported in Table 1. Overall, both husbands (33.68 mmol/mol, or 5.2%) and wives (32.67 mmol/mol, or 5.1%) displayed average HbA1c levels in the healthy range, approximately 20 to 42 millimoles per mole in persons without diabetes (Health Service Executive, 2010). However, husbands’ and wives’ HbA1c levels were significantly different from one another (p < 0.001). Husbands were also significantly older than wives on average (62 vs. 60 years, p < 0.001). Likewise, in keeping with prior studies of older Irish couples in particular (e.g., Stokes, 2017a, 2017b), husbands reported significantly lower loneliness (p < 0.05) and higher marital support (p < 0.001) than wives, though there was no significant gender difference concerning marital strain. Despite these gender differences, average levels of loneliness and marital strain were fairly low for both men and women, while levels of marital support were fairly high overall. In terms of employment status, a plurality of both husbands (47.86%) and wives (39.74%) reported being currently employed, though this gender difference was also significant (p < 0.001). More strikingly, 40.42% of husbands reported being retired, compared with only 23.67% of wives (p < 0.001); in contrast, 36.59% of wives reported an “other” employment status, compared with only 11.72% of husbands, reflecting the continued gender differences in labor force participation among the current cohort of aging couples in Ireland, despite the liberalization of employment regulations surrounding married women that led to the Irish “marriage boom” in the late 1960s and 1970s (Kennedy, 1973). Among covariates, husbands were disproportionately represented in both the lowest and highest education categories (p < 0.001 for the lowest and highest categories in particular), and reported significantly fewer depressive symptoms than their wives (p < 0.001). There were no gender differences concerning income or the likelihood of having retired due to illness.

TABLE 1.

Descriptive statistics, the Irish longitudinal study on aging, 2011 (N = 1331 dyads)

Variables Husbands (N = 1331) Wives (N = 1331) p
Measures of interest M (SD) or % M (SD) or %
HbA1c (mmol/mol)a 33.68 (.17) 32.67 (.15) ***
Lonelinessa 1.25 (.41) 1.28 (.44) *
Marital supporta 3.74 (.47) 3.58 (.61) ***
Marital straina 1.65 (.5×6) 1.68 (.63) -
Age 62.09 (.24) 59.68 (.23) ***
Employment status
 Retired 40.42% 23.67% ***
 Employed 47.86% 39.74% ***
 Other employment status 11.72% 36.59% ***
Individual-level covariates
Education
 Some primary (not complete) 2.85% 1.13% ***
 Primary or equivalent 22.16% 16.00% ***
 Intermediate/junior/group certificate or equivalent 24.19% 26.00% -
 Leaving certificate or equivalent 17.66% 21.19% *
 Diploma/certificate 13.60% 21.79% ***
 Primary degree 11.87% 9.09% *
 Postgraduate/higher degree 7.66% 4.81% ***
Incomea 41,023 (31,154) 42,488 (33,981) -
Depressive symptomsa 1.66 (.19) 1.72 (.23) ***
Retired due to ill healthb 5.48% 4.88% -
Couple-level covariates
Number of children 2.85 (1.16) 2.85 (1.16) -
Living situation
 Lives with spouse only 55.60% 55.60% -
 Lives with spouse and others 44.40% 44.40% -
Region
 Dublin City or County 24.65% 24.65% -
 Other City or County 24.64% 24.64% -
 Rural 50.71% 50.71% -
Open in a new tab

Note:

*

p < .05

***

p < .001.

a

Raw statistics presented; variable transformed for analysis.

b

Percentages reported refer to the proportion of retired persons who listed ill health as a reason for retiring; participants who were employed or held an “other” employment status were coded as “0” for analysis, as per dummy variable adjustment.

Table 2 presents the results from dyadic structural equation models (SEM) concerning dyadic loneliness, employment status, and older husbands’ and wives’ HbA1c levels. Model 1 examined the main effects of own and partner’s loneliness as well as employment status, accounting for all other measures and covariates. Among measures of interest, only age (B = 0.03, p < 0.001) was significantly associated with husbands’ and wives’ HbA1c levels. Among covariates, only retiring due to ill health (B = 0.21, p < 0.05) displayed significant results. Model 2 included all main effects and covariates, and added interaction terms between spouses’ own loneliness and employment status to Model 1. Again, among measures of interest only age (B = 0.02, p < 0.001) was significantly associated with older spouses’ HbA1c, and neither of the interaction terms was significant for either husbands or wives. No other significant effects were altered from Model 1. Model 3 included all main effects and covariates, and added interaction terms between a partner’s loneliness and employment status to Model 1. In this case, husbands’ and wives’ own loneliness remained non-significant, yet a partner’s loneliness became a significant predictor of both spouses’ HbA1c (B = 0.10, p < 0.01). Furthermore, interaction terms between partner’s loneliness and being currently employed (B = −0.14, p < 0.01) and having an “other” employment status (B = −0.13, p < 0.05) were both significant. This indicates that having a lonely partner was associated with significant higher levels of HbA1c, but only among those who were retired. Among both the currently employed and those with an “other” employment status, this association remained non-significant. Figure 2 illustrates these significant interaction results. Age (B = 0.02, p < 0.001) remained a positive and significant predictor of husbands’ and wives’ HbA1c levels, as did retiring due to ill health (B = 0.20, p < 0.01). A fully saturated model (not shown; available upon request) that included the interactions terms tested in Model 2 and Model 3 simultaneously produced results that were consistent with those reported in Model 3; models estimated without covariates also produced the same significant results of interest as those presented here.

TABLE 2.

Results of dyadic structural equation models predicting older Irish Husbands’ and Wives’ HbA1c, 2011 (N = 1331 dyads)

Model 1
 Model 2
 Model 3
Husbands
(N = 1331)		 Wives
(N = 1331)		 Husbands
(N = 1331)		 Wives
(N = 1331)		 Husbands
(N = 1331)		 Wives
(N = 1331)
Measures of interest B (SE) B (SE) B (SE) B (SE) B (SE) B (SE)
Loneliness (own)a 0.01 (0.02) 0.01 (0.02) 0.03 (0.04) 0.03 (0.04) 0.01 (0.02) 0.01 (0.02)
Marital support (own)a 0.00 (0.02) 0.00 (0.02) 0.01 (0.02) 0.01 (0.02) 0.01 (0.02) 0.01 (0.02)
Marital strain (own)a 0.00 (0.02) 0.00 (0.02) 0.00 (0.02) 0.00 (0.02) 0.00 (0.02) 0.00 (0.02)
Partner’s lonelinessa 0.00 (0.02) 0.00 (0.02) 0.00 (0.02) 0.00 (0.02) 0.10** (0.04) 0.10** (0.04)
Partner’s marital supporta 0.01 (0.02) 0.01 (0.02) 0.01 (0.02) 0.01 (0.02) 0.01 (0.02) 0.01 (0.02)
Partner’s marital straina 0.01 (0.02) 0.01 (0.02) 0.01 (0.02) 0.01 (0.02) 0.01 (0.02) 0.01 (0.02)
Age 0.03*** (0.00) 0.03*** (0.00) 0.02*** (0.00) 0.02*** (0.00) 0.02*** (0.00) 0.02*** (0.00)
Employedb 0.06 (0.06) 0.06 (0.06) 0.06 (0.06) 0.06 (0.06) 0.05 (0.06) 0.05 (0.06)
Other employment statusb 0.03 (0.06) 0.03 (0.06) 0.03 (0.06) 0.03 (0.06) 0.01 (0.06) 0.01 (0.06)
Interaction terms
Loneliness (own)a × Employedb - - −0.03 (0.05) −0.03 (0.05) - -
Loneliness (own)a × Other Employment statusb - - −0.03 (0.05) −0.03 (0.05) - -
Partner’s lonelinessa × Employedb - - - - −0.14** (0.05) −0.14** (0.05)
Partner’s lonelinessa × Other Employment statusb - - - - −0.13* (0.05) −0.13* (0.05)
Covariates
Incomea −0.03 (0.02) −0.03 (0.02) −0.03 (0.02) −0.03 (0.02) −0.03 (0.02) −0.03 (0.02)
Primary or equivalentc −0.01 (0.14) −0.01 (0.14) −0.01 (0.14) −0.01 (0.14) 0.01 (0.14) 0.01 (0.14)
Intermediate or equivalentc −0.00 (0.14) −0.00 (0.14) 0.00 (0.14) 0.00 (0.14) 0.01 (0.14) 0.01 (0.14)
Leaving certificate or equivalentc −0.06 (0.15) −0.06 (0.15) −0.06 (0.15) −0.06 (0.15) −0.06 (0.15) −0.06 (0.15)
Diploma/certificatec −0.13 (0.15) −0.13 (0.15) −0.13 (0.15) −0.13 (0.15) −0.13 (0.15) −0.13 (0.15)
Primary degreec −0.12 (0.15) −0.12 (0.15) −0.12 (0.15) −0.12 (0.15) −0.11 (0.15) −0.11 (0.15)
Postgraduate/higher degreec −0.22 (0.17) −0.22 (0.17) −0.22 (0.16) −0.22 (0.16) −0.22 (0.16) −0.22 (0.16)
Depressive symptoms (own)a −0.02 (0.02) −0.02 (0.02) −0.02 (0.02) −0.02 (0.02) −0.03 (0.02) −0.03 (0.02)
Partner’s depressive symptomsa 0.03 (0.02) 0.03 (0.02) 0.03 (0.02) 0.03 (0.02) 0.03 (0.02) 0.03 (0.02)
Retired due to ill healthd 0.21* (0.09) 0.21* (0.09) 0.21* (0.09) 0.21* (0.09) 0.20* (0.09) 0.20* (0.09)
Number of children −0.01 (0.02) −0.01 (0.02) −0.01 (0.02) −0.01 (0.02) −0.01 (0.02) −0.01 (0.02)
Lives with spouse and otherse −0.03 (0.05) −0.03 (0.05) −0.03 (0.05) −0.03 (0.05) −0.03 (0.05) −0.03 (0.05)
Other City or Countyf 0.04 (0.06) 0.04 (0.06) 0.04 (0.06) 0.04 (0.06) 0.03 (0.06) 0.03 (0.06)
Ruralf 0.03 (0.05) 0.03 (0.05) 0.03 (0.05) 0.03 (0.05) 0.02 (0.05) 0.02 (0.05)
Model fit
CFI 0.99 0.98 0.98
RMSEA 0.01 0.01 0.01
Open in a new tab

Note:

*

p < 0.05

**

p < 0.01

***

p < 0.001.

a

Transformed variable.

b

Reference is retired.

c

Reference is some primary education.

d

Reference is did not retire due to ill health. Values set to “0” for N/A, as per dummy variable adjustment.

e

Reference is lives with spouse only.

f

Reference is Dublin City or County.

FIGURE 2.

FIGURE 2

Open in a new tab

Dyadic partner’s loneliness and HbA1c among older married adults in Ireland

4 |. DISCUSSION

Previous research has established loneliness as a serious public health concern, especially among the aging population (e.g., Cacioppo et al., 2006; Donovan et al., 2017; Holt-Lunstad et al., 2015). In particular, numerous recent studies have identified cardiometabolic health as a potential pathway for the harmful health effects of loneliness (e.g., Hawkley et al., 2010; Hawkley & Capitanio, 2015; Valtorta et al., 2016; Xia & Li, 2018), including HbA1c specifically (O’Luanaigh et al., 2012; Stokes & Barooah, 2021), although some research has called these findings into question (Das, 2019). A common limitation in most of these studies is a focus on loneliness among individuals, rather than among partners situated within relationships. Moreover, the extent to which loneliness affects health for oneself or one’s partner may depend on additional contextual factors, as well (Stokes & Barooah, 2021). The present study analyzed dyadic data drawn from a nationally representative sample of older couples in the Republic of Ireland, in order to determine (a) whether loneliness of one partner was associated with HbA1c levels of either partner, and (b) whether any such associations varied by employment/retirement status. We also examined whether marital support or strain affected dyadic HbA1c levels, as well as whether effects of employment and retirement status were in fact a function of age.

Findings revealed that – contrary to expectations – loneliness was not associated with older married adults’ own HbA1c levels, irrespective of employment/retirement status. However, loneliness of a partner was associated with elevated HbA1c among retired persons, but not among the employed or those reporting an “other” employment status. Moreover, although marital support and strain are important factors in determining older spouses’ loneliness itself, they did not have any independent association with husbands’ or wives’ HbA1c levels. Lastly, sensitivity analyses examining interactions between dyadic loneliness measures and age confirmed that the results of interest concerning employment and retirement status were not merely a function of age, but rather due to employment status itself. These results further highlight the importance of examining loneliness and health within a relational context, and of thoroughly considering the circumstances under which loneliness may be harmful for health – even if it is the health of others, and not oneself. Moreover, findings offer support for the stress generation hypothesis (Hammen, 2006) and further illuminate the process of socioemotional selectivity and the importance of life stage transitions and future time horizons for older adults’ relationships and well-being (Carstensen et al., 1999; Lang & Carstensen, 2002). We discuss the implications of our study for theory and future research below.

4.1 |. Loneliness and HbA1c

As noted, numerous prior studies have examined links between loneliness and measures of cardiometabolic health. These include studies concerning increased blood pressure (e.g., Hawkley et al., 2006; Hawkley et al., 2010), coronary heart disease, and stroke risk (e.g., Valtorta et al., 2016), and HbA1c (e.g., O’Luanaigh et al., 2012). The last of these studies analyzed an older adult sample from Ireland, as well. Despite a substantial literature on loneliness and health more broadly, however, the literature specific to measures of cardiometabolic health remains more limited (Das, 2019).

Little research has examined loneliness and health among married older adults specifically, in part because the married report both lower loneliness and better health than their unmarried counterparts on average (e.g., Perissinotto et al., 2012; Stokes & Barooah, 2021). One recent study examining loneliness and HbA1c among older married couples in the United States, however, found that loneliness was associated with elevated HbA1c levels, but only among those reporting poor marital support (Stokes & Barooah, 2021). Based on the existing prior research, we expected loneliness to be associated with elevated HbA1c levels among older married Irish adults, at least among retired persons, if not among the sample as a whole. In contrast, our study indicated no clear link between own loneliness and HbA1c, irrespective of employment status. Differences in our findings compared with those from O’Luanaigh et al. (2012) may be due to the latter’s use of a non-probability sample of older adults located in and around Dublin, whereas data for this study were drawn from a national probability sample of older adults living across the Republic of Ireland (see, e.g., Das, 2019). Moreover, these differences in findings may also be due to the focus in this study on married older adults, resulting in a sample that is substantially younger (60.89 vs. 75.45 years old, on average) and healthier (HbA1c of 5.19% on average vs. 5.98%) than that from O’Luanaigh et al. (2012), and most likely less lonely as well, though measures of loneliness differed across these two studies. Thus, it is possible that loneliness is more strongly associated with HbA1c and other markers of cardiometabolic health at older ages, higher levels of loneliness, or among the not married than it is among the comparatively younger, healthier, and partnered sample analyzed here. However, evidence provided by the present analysis does not offer support for our expectation that loneliness would be associated with elevated HbA1c, which contributes additional knowledge to this growing field (Das, 2019).

4.2 |. Dyadic loneliness and partners’ health

Although research to date on loneliness has predominantly examined individuals (e.g., Das, 2019; Donovan et al., 2017; Holt-Lunstad et al., 2015; O’Luanaigh et al., 2012), recent studies have increasingly situated loneliness within a dyadic and relational context (e.g., Ayalon et al., 2013; Moorman, 2016; Stokes, 2017a, 2017b). Generally, these studies indicate that loneliness is influenced by both partners’ perceptions of marital quality (e.g., Moorman, 2016; Stokes, 2017a, 2017b), and that loneliness can be “contagious” within older couples (e.g., Ayalon et al., 2013; Stokes, 2017a, 2017b). One recent study has also noted that – in the setting of lacking marital support – loneliness of one partner can be detrimental for the health of both partners over time (Stokes & Barooah, 2021). Overall, this body of literature suggests that loneliness may be detrimental for health not merely through individual pathways, but through relationship mechanisms as well. Furthermore, loneliness itself is an outcome of relationship factors and dynamics.

Results of the present study bolster this approach, and offer evidence supporting both the stress generation hypothesis (e.g., Hammen, 2006) and socioemotional selectivity theory (SST; Carstensen et al., 1999). Specifically, the stress generation hypothesis notes that emotional experiences of one partner (such as depression or loneliness) can influence that person’s behavior and interactions with others, leading to stressful and toxic interactions with close partners (see Cacioppo et al., 2009; Hammen, 2006; Stokes & Barooah, 2021). In our analyses, we did not find evidence indicating that loneliness among dyadic partners was associated with elevated HbA1c overall. However, the interaction analysis revealed that having a lonely partner was associated with elevated HbA1c among those who were retired.

This particular finding highlights the intersection of stress generation and SST processes among the aging population. As adults age and their future time horizons change, so too do their social goals and behaviors (Carstensen et al., 1999). In general, this process is associated with smaller – but higher quality – social networks among older adults, as they trim negative or straining ties from their lives, and focus more time and energy on their closest and most meaningful relationships (Carstensen et al., 1999). Conversely, however, the process of SST – coupled in particular with the transition into retirement, which removes persons from social settings and ties fostered in the workplace – also leads to these fewer remaining relationships becoming more salient and potentially more impactful for one’s well-being, for better or for worse (e.g., Stokes, 2017b). In the present case, partner’s loneliness was associated with elevated HbA1c only among the retired, eve accounting for age; moreover, supplemental analyses (not shown) testing interactions by age rather than employment status confirmed that this effect was not due to age, but to retirement status itself. This underscores the importance of examining life stage events and major transitions when studying the life course, as age itself may be an imperfect proxy for older persons’ life circumstances and future time horizons.

4.3 |. Gender

It is worth noting that, while no effects in this study differed by gender, there are certainly gendered implications of our findings. First, men were substantially more represented in the retired group than women (40.42% vs. 23.67%, p < 0.001), whereas women were more likely to have an “other” employment status (36.59% vs. 11.72%, p < 0.001). This suggests that older married men are the most likely to have transitions from the workplace into retirement, and are thus most likely to be harmed by the presence of a lonely partner. Second, although both husbands and wives reported fairly low levels of loneliness overall, wives’ loneliness was significantly higher than husbands’ (1.28 vs. 1.25, p < 0.05), further exposing retired husbands to harmful health effects. Lastly, our findings suggest that the transition from the labor force into retirement matters for dyadic health effects, as those reporting an “other” employment status do not show the same susceptibility to their spouses’ loneliness. It is possible that this is due in part to the gender composition of these employment status groups, though it is also possible that those who have been outside of the formal labor force for extended periods of the life course have established other compensatory social networks that remain in later life, whereas the retired are newly facing the loss of prior social relationships.

4.4 |. Limitations

The present study has a number of limitations worth noting. First, data used in this analysis were cross-sectional, preventing any examination of change over time (i.e., HbA1c preceding and following the transition into retirement) and limiting any assessment of causality. Moreover, the duration and reasons for retirement could not be fully examined either, beyond an indicator of whether retirement was due to ill health. Although TILDA itself is longitudinal, and the health assessment was repeated at Wave 3 (2014–2015), HbA1c was measured only at Wave 1. Other biological markers that were included in both TILDA health assessments (e.g., HDL) have not been linked with loneliness either individually or dyadically in previous studies (e.g., O’Luanaigh et al., 2012; Stokes & Barooah, 2021). Future research should examine longitudinal associations of dyadic loneliness and multiple biological markers of cardiometabolic health, across different national settings. Lastly, although TILDA is nationally representative of the population aged 50 and older in the Republic of Ireland, there is potential for self-selection in our sample due to the use of blood-based biomarkers, as only 72% of TILDA participants completed the health assessment at Wave 1. Moreover, because the analytic sample is restricted to married couples, these participants are likely younger, healthier, and less lonely than the overall TILDA sample; thus, we may be unable to detect associations that manifest only at later ages, higher levels of loneliness, or among those in poor health.

5 |. CONCLUSION

Despite these limitations, the present study makes a number of contributions to the literature concerning loneliness and health, particularly within a dyadic context. First, contrary to both prior research and our expectations, loneliness was not associated with elevated HbA1c among older married adults in Ireland (O’Luanaigh et al., 2012). This was true both overall, and among retired persons specifically. This suggests that associations between loneliness and HbA1c may be contingent upon factors not included or available in this study, for example among the oldest-old or the very lonely (e.g., Das, 2019; O’Luanaigh et al., 2012). Furthermore, this result was in contrast to recent findings concerning loneliness among married older adults in the United States, suggesting that cultural or national setting may be an important contextual factor as well (Stokes & Barooah, 2021). Notably, both loneliness and HbA1c were substantially higher in the U.S. sample than in the TILDA sample analyzed here (Stokes & Barooah, 2021).

Secondly, however, this study revealed that loneliness of a dyadic partner was significantly associated with elevated HbA1c among retired persons, but not among those currently employed or with an “other” employment status. This underscores the importance of relationship circumstances for determining dyadic effects and the influence one spouse can have on the other. Moreover, these results imply the intersection of stress generation and socioemotional selectivity, such that the loneliness of an intimate partner becomes a meaningful and impactful stressor with health implications only once adults exit the labor force, engage in the progressive trimming of social networks (whether chosen or unchosen), and their marital relationship takes on greater prominence and salience in their daily lives. These insights will aid policy makers and practitioners in effectively identifying the populations most at risk of experiencing loneliness, and of experiencing poor health due to (their partner’s) loneliness. Finally, these findings will be of use for future researchers, as they underscore the importance of incorporating relationship and life course context when examining loneliness and health, as the harmful influence of loneliness – for oneself or for others – may only manifest under particular circumstances, rather than on a population average level (e.g., Das, 2019; Stokes & Barooah, 2021).

Statement of Relevance:

Loneliness is not just an individual experience, it has relational implications as well. Indeed, research shows that loneliness is “contagious” among older couples, such that having a lonely partner is a risk factor for experiencing loneliness oneself. Recently, researchers have begun addressing whether having a lonely partner is also a risk factor for poorer health among the aging population. This study provides evidence that this is in fact the case, particularly among retired older persons.

FUNDING INFORMATION

This research was supported by a grant from the National Institute on Aging (R03AG064283).

Footnotes

CONFLICT OF INTEREST

The authors declare that there is no conflict of interest.

DATA AVAILABILITY STATEMENT

Researchers interested in using TILDA data may access the data for free from the following sites: Irish Social Science Data Archive (ISSDA) at University College Dublin http://www.ucd.ie/issda/data/tilda/. Interuniversity Consortium for Political and Social Research (ICPSR) at the University of Michigan http://www.icpsr.umich.edu/icpsrweb/ICPSR/studies/34315.

REFERENCES

  1. Allison P (2002). Missing data: Quantitative applications in the social sciences SAGE Publications Inc. 10.4135/9781412985079 [DOI]
  2. Ayalon L, Shiovitz-Ezra S, & Palgi Y (2013). Associations of loneliness in older married men and women. Aging and Mental Health, 17(1), 33–39. 10.1080/13607863.2012.702725 [DOI] [PubMed] [Google Scholar]
  3. Barrett A, Burke H, Cronin H, Hickey A, Kamiya Y, Kenny RA, Layte R, Maty S, McGee H, Morgan K, Mosca I, Normand C, O’Regan C, O’Sullivan V, Savva G, Sofroniou N, Timonen V, & Whelan B (2011). Fifty plus in Ireland 2011: First results from the Irish longitudinal study on ageing (TILDA) 10.25419/RCSI.10770389.V1 [DOI]
  4. Barrett A, & Mosca I (2013). The psychic costs of migration: Evidence from Irish return migrants. Journal of Population Economics, 26(2), 483–506. 10.1007/S00148-012-0438-4 [DOI] [Google Scholar]
  5. Berscheid E, & Reis HT (1998). Attraction and close relationships. In Gilbert DT, Fiske ST, & Lindzey G (Eds.), The handbook of social psychology (4th ed., pp. 193–281). McGraw-Hill. https://psycnet.apa.org/record/1998-07091-022 [Google Scholar]
  6. Bookwala J (2012). Marriage and other partnered relationships in middle and late adulthood. In Blieszner R & Bedford VH (Eds.), Handbook of families and aging (2nd ed., pp. 91–123). Praeger. https://ldr.lafayette.edu/concern/publications/3197xm48m [Google Scholar]
  7. Börsch-Supan A, & Schuth M (2014). Early retirement, mental health, and social networks. In Wise DA (Ed.), Discoveries in the economics of aging (pp. 225–250). University of Chicago Press. [Google Scholar]
  8. Burholt V, & Scharf T (2014). Poor health and loneliness in later life: The role of depressive symptoms, social resources, and rural environments. The Journals of Gerontology. Series B, Psychological Sciences and Social Sciences, 69(2), 311–324. 10.1093/geronb/gbt121 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Burke TJ, & Segrin C (2014). Bonded or stuck? Effects of personal and constraint commitment on loneliness and stress. Personality and Individual Differences, 64, 101–106. 10.1016/J.PAID.2014.02.027 [DOI] [Google Scholar]
  10. Butrica BA, & Schaner SG (2005). Satisfaction and engagement in retirement http://www.urban.org
  11. Cacioppo JT, Fowler JH, & Christakis NA (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. 10.1161/CIRCULATIONAHA.110.956839 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Cacioppo JT, Hughes ME, Waite LJ, Hawkley LC, & Thisted RA (2006). Loneliness as a specific risk factor for depressive symptoms: Cross-sectional and longitudinal analyses. Psychology and Aging, 21(1), 140–151. 10.1037/0882-7974.21.1.140 [DOI] [PubMed] [Google Scholar]
  13. Carr D, Freedman VA, Cornman JC, & Schwarz N (2014). Happy marriage, happy life? Marital quality and subjective well-being in later life. Journal of Marriage and Family, 76(5), 930–948. 10.1111/jomf.12133 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Carstensen LL (2006). The influence of a sense of time on human development. Science, 312(5782), 1913–1915. 10.1126/SCIENCE.1127488 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Carstensen LL, Fung HH, & Charles ST (2003). Socioemotional selectivity theory and the regulation of emotion in the second half of life. Motivation and Emotion, 97(1–4), 103–123. [Google Scholar]
  16. Carstensen LL, Isaacowitz DM, & Charles ST (1999). Taking time seriously: A theory of socioemotional selectivity. American Psychologist, 54(3), 165–181. 10.1037/0003-066X.54.3.165 [DOI] [PubMed] [Google Scholar]
  17. Cronin H, O’Regan C, Finucane C, Kearney P, & Kenny RA (2013). Health andaging: development of the Irish Longitudinal Study on Ageing health assessment. Journal of the American Geriatrics Society, 61, S269–S278. 10.1111/jgs.12197 [DOI] [PubMed] [Google Scholar]
  18. Comi SL, Cottini E, & Lucifora C (2022). The effect of retirement on social relationships. In German Economic Review De Gruyter Open Ltd. 10.1515/GER-2020-0109/DOWNLOADASSET/SUPPL/GER-2020-0109SUPP.PDF [DOI]
  19. Das A (2019). Loneliness does (not) have cardiometabolic effects: A longitudinal study of older adults in two countries. Social Science and Medicine, 223, 104–112. 10.1016/j.socscimed.2018.10.021 [DOI] [PubMed] [Google Scholar]
  20. de Jong-Gierveld J (1987). Developing and testing a model of loneliness. Journal of Personality and Social Psychology, 53(1), 119–128. 10.1037/0022-3514.53.1.119 [DOI] [PubMed] [Google Scholar]
  21. Distel MA, Rebollo-Mesa I, Abdellaoui A, Derom CA, Willemsen G, Cacioppo JT, & Boomsma DI (2010). Familial resemblance for loneliness. Behavior Genetics, 40(4), 480–494. 10.1007/S10519-010-9341-5/TABLES/8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Donovan NJ, Wu Q, Rentz DM, Sperling RA, Marshall GA, & Glymour MM (2017). Loneliness, depression and cognitive function in older U.S. adults. International Journal of Geriatric Psychiatry, 32(5), 564–573. 10.1002/gps.4495 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Dykstra PA, Van Tilburg TG, & Gierveld JDJ (2005). Changes in older adult loneliness: Results from a seven-year longitudinal study. Research on Aging, 27(6), 725–747. 10.1177/0164027505279712 [DOI] [Google Scholar]
  24. Fokkema T, De Jong Gierveld J, & Dykstra PA (2012). Cross-national differences in older adult loneliness. Journal of Psychology: Interdisciplinary and Applied, 146(1–2), 201–228. 10.1080/00223980.2011.631612 [DOI] [PubMed] [Google Scholar]
  25. Fung HH, & Carstensen LL (2006). Goals change when Life’s fragility is primed: Lessons learned from older adults, the September 11 attacks and Sars. Social Cognition, 24(3), 248–278. 10.1521/SOCO.2006.24.3.248 [DOI] [Google Scholar]
  26. Hammen C (2006). Stress generation in depression: Reflections on origins, research, and future directions. Journal of Clinical Psychology, 62(9), 1065–1082. 10.1002/JCLP.20293 [DOI] [PubMed] [Google Scholar]
  27. Hawkley LC, Masi CM, Berry JD, & Cacioppo JT (2006). Loneliness is a unique predictor of age-related differences insystolic blood pressure. Psychology and Aging, 21(1), 152–164. 10.1037/0882-7974.21.1.152 [DOI] [PubMed] [Google Scholar]
  28. Hawkley LC, & Cacioppo JT (2010). Lonelinessmatters: A theoretical and empirical review of consequences and mechanisms. Annals of Behavioral Medicine, 40(2), 218–227. 10.1007/s12160-010-9210-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Hawkley LC, & Capitanio JP (2015). Perceived social isolation, evolutionary fitness and health outcomes: A lifespan approach. Philosophical Transactions of the Royal Society, B: Biological Sciences, 370(1669), 20140114. 10.1098/rstb.2014.0114 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Hawkley LC, Thisted RA, Masi CM, & Cacioppo JT (2010). Loneliness predicts increased blood pressure: Five-year cross-lagged analyses in middle-aged and older adults. Psychology and Aging, 25(1), 132–141. 10.1037/a0017805 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Health Service Executive. (2010). Information for people with diabetes: International changes to HbA1c levels https://www.diabetes.ie/wp-content/uploads/2014/11/hba1c.pdf
  32. Holt-Lunstad J, Smith TB, Baker M, Harris T, & Stephenson D (2015). Loneliness and social isolation as risk factors for mortality: A meta-analytic review. Perspectives on Psychological Science, 10(2), 227–237. 10.1177/1745691614568352 [DOI] [PubMed] [Google Scholar]
  33. Hughes ME, Waite LJ, Hawkley LC, & Cacioppo JT (2004). A short scale for measuring loneliness in large surveys: Results from two population-based studies. Research on Aging, 26(6), 655–672. 10.1177/0164027504268574 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Kamiya Y, Doyle M, Henretta JC, & Timonen V (2013). Depressive symptoms among older adults: The impact of early and later life circumstances and marital status. Aging and Mental Health, 17(3), 349–357. 10.1080/13607863.2012.747078 [DOI] [PubMed] [Google Scholar]
  35. Kamiya Y, Doyle M, Henretta JC, & Timonen V (2014). Early-life circumstances and later-life loneliness in Ireland. The Gerontologist, 54(5), 773–783. 10.1093/GERONT/GNT097 [DOI] [PubMed] [Google Scholar]
  36. Kamiya Y, & Sofroniou N (2011). Socio-demographic characteristics of older people in Ireland. In Barrett A, Savva G, Timonen V, & Kenny RA (Eds.), Fifty Plus in Ireland 2011: First Results from The Irish Longitudinal Study on Ageing (pp. 23–36). The Irish Longitudinal Study on Ageing. 10.38018/TildaRe.2011-00.c2 [DOI] [Google Scholar]
  37. Kennedy RE (1973). The Irish; emigration, marriage, and fertility University of California Press. [Google Scholar]
  38. Kenny DA, & Cook W (1999). Partner effects in relationship research: Conceptual issues, analytic difficulties, and illustrations. Personal Relationships, 6(4), 433–448. 10.1111/j.1475-6811.1999.tb00202.x [DOI] [Google Scholar]
  39. Kenny RA, Whelan BJ, Cronin H, Kamiya Y, Kearney P, O’Regan C, & Ziegel M (2010). The Design of the Irish Longitudinal Study on Ageing
  40. Kulik L (2001). The impact of men’s and women’s retirement on marital relations: A comparative analysis. Journal of Women & Aging, 13(2), 21–37. 10.1300/J074V13N02_03 [DOI] [PubMed] [Google Scholar]
  41. Lang FR, & Carstensen LL (2002). Time counts: Future time perspective, goals, and social relationships. Psychology and Aging, 17(1), 125–139. 10.1037/0882-7974.17.1.125 [DOI] [PubMed] [Google Scholar]
  42. Löckenhoff CE (2012). Understanding retirement: The promise of life-span developmental frameworks. European Journal of Ageing, 9(3), 227–231. 10.1007/S10433-012-0241-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Löckenhoff CE, & Carstensen LL (2004). Socioemotionol selectivity theory, aging, and health: The increasingly delicate balance between regulating emotions and making tough choices. Journal of Personality, 72(6), 1395–1424. 10.1111/j.1467-6494.2004.00301.x [DOI] [PubMed] [Google Scholar]
  44. Luhmann M, & Hawkley LC (2016). Age differences in loneliness from late adolescence to oldest old age. Developmental Psychology, 52(6), 959. 10.1037/DEV0000117 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Luo Y, Hawkley LC, Waite LJ, & Cacioppo JT (2012). Loneliness, health, and mortality in old age: A national longitudinal study. Social Science and Medicine, 74(6), 907–914. 10.1016/j.socscimed.2011.11.028 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Mancini AD, & Bonanno GA (2006). Marital closeness, functional disability, and adjustment in late life. Psychology and Aging, 21(3), 600–610. 10.1037/0882-7974.21.3.600 [DOI] [PubMed] [Google Scholar]
  47. Moorman SM (2016). Dyadic perspectives on marital quality and loneliness in later life. Journal of Social and Personal Relationships, 33(5), 600–618. 10.1177/0265407515584504 [DOI] [Google Scholar]
  48. Mosca I, & Barrett A (2014). The Impact of Voluntary and Involuntary Retirement on Mental Health: Evidence from Older Irish Adults (IZA DP No. 8723) [PubMed]
  49. Mosca I, & Wright RE (2018). Effect of retirement on cognition: Evidence from the Irish marriage Bar. Demography, 55(4), 1317–1341. 10.1007/S13524-018-0682-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Ní Léime A, Duvvury N, & Callan A (2015). Pension provision, gender, ageing and work in Ireland. In Walsh K, Carney GM, & Ní Léime Á (Eds.), Ageing through austerity: Critical perspectives from Ireland (pp. 63–78). Bristol University Press, Policy Press. [Google Scholar]
  51. O’Luanaigh C, O’Connell H, Chin AV, Hamilton F, Coen R, Walsh C, Walsh JB, Coakley D, Molloy A, Scott J, Cunningham CJ, & Lawlor BA (2012). Loneliness and vascular biomarkers: The Dublin healthy ageing study. International Journal of Geriatric Psychiatry, 27(1), 83–88. 10.1002/gps.2695 [DOI] [PubMed] [Google Scholar]
  52. O’Luanaigh C, & Lawlor BA (2008). Loneliness and the health of older people. International Journal of Geriatric Psychiatry, 23, 1213–1221. 10.1002/gps.2054 [DOI] [PubMed] [Google Scholar]
  53. Perissinotto CM, Stijacic Cenzer I, & Covinsky KE (2012). Loneliness in older persons: A predictor of functional decline and death. Archives of Internal Medicine, 172(14), 1078–1083. 10.1001/archinternmed.2012.1993 [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Pinquart M, & Sörensen S (2001). Influences on loneliness in older adults: A meta-analysis. Basic and Applied Social Psychology, 23(4), 245–266. 10.1207/S15324834BASP2304_2 [DOI] [Google Scholar]
  55. Rumens N (2016). Researching workplace friendships: Drawing insights from the sociology of friendship. Journal of Social and Personal Relationships J, 34(8), 1149–1167. 10.1177/0265407516670276 [DOI] [Google Scholar]
  56. Santos-Oliveira R, Purdy C, Da Silva MP, Dos Anjos Carneiro-Leão AM, Machado M, & Einarson TR (2011). Haemoglobin A1c levels and subsequent cardiovascular disease in persons without diabetes: A meta-analysis of prospective cohorts. Diabetologia, 54(6), 1327–1334. 10.1007/S00125-011-2078-8 [DOI] [PubMed] [Google Scholar]
  57. Segrin C, Burke TJ, & Dunivan M (2012). Loneliness and poor health within families. Journal of Social and Personal Relationships, 29(5), 597–611. 10.1177/0265407512443434 [DOI] [Google Scholar]
  58. Stokes JE (2017a). Marital quality and loneliness in later life: A dyadic analysis of older married couples in Ireland. Journal of Social and Personal Relationships, 34(1), 114–135. 10.1177/0265407515626309 [DOI] [Google Scholar]
  59. Stokes JE (2017b). Two-wave dyadic analysis of marital quality and loneliness in later life: Results from the Irish longitudinal study on ageing. Research on Aging, 39(5), 635–656. 10.1177/0164027515624224 [DOI] [PubMed] [Google Scholar]
  60. Stokes JE (2017c). Do “his” and “her” marriages influence one another? Contagion in personal assessments of marital quality among older spouses over a four-year span. Social Psychology Quarterly, 80(2), 204–215. 10.1177/0190272517701701 [DOI] [Google Scholar]
  61. Stokes JE, & Barooah A (2021). Dyadic loneliness and changes to HbA1c among older US couples: The role of marital support as stress buffer. Journal of Aging and Health, 33(9), 698–708. 10.1177/08982643211006498 [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Story TN, Berg CA, Smith TW, Beveridge R, Henry NJM, & Pearce G (2007). Age, marital satisfaction, and optimism as predictors of positive sentiment override in middle-aged and older married couples. Psychology and Aging, 22(4), 719–727. 10.1037/0882-7974.22.4.719 [DOI] [PubMed] [Google Scholar]
  63. Teuscher U (2010). Change and persistence of personal identities after the transition to retirement. International Journal of Aging & Human Development, 70(1), 89–106. 10.2190/AG.70.1.D [DOI] [PubMed] [Google Scholar]
  64. Thomeer MB, Umberson D, & Pudrovska T (2013). Marital processes around depression: A gendered and relational perspective. Society and Mental Health, 3(3), 151–169. 10.1177/2156869313487224 [DOI] [PMC free article] [PubMed] [Google Scholar]
  65. TILDA. (2019). The Irish Longitudinal study on Ageing (TILDA) Wave 1, 2009–2011. [dataset]. Version 1.9. Irish Social Science Data Archive. SN:0053–01 www.ucd.ie/issda/data/tilda/wave1
  66. Umberson D, Williams K, Powers DA, Liu H, & Needham B (2006). You make me sick: Marital quality and health over the life course. Journal of Health and Social Behavior, 47(1), 1–16. 10.1177/002214650604700101 [DOI] [PMC free article] [PubMed] [Google Scholar]
  67. Valtorta NK, Kanaan M, Gilbody S, Ronzi S, & Hanratty B (2016). Loneliness and social isolation as risk factors for coronary heart disease and stroke: Systematic review and meta-analysis of longitudinal observational studies. Heart, 102, 1009–1016. 10.1136/heartjnl-2015-308790 [DOI] [PMC free article] [PubMed] [Google Scholar]
  68. van Solinge H, & Henkens K (2008). Adjustment to and satisfaction with retirement: Two of a kind? Psychology and Aging, 23(2), 422–434. 10.1037/0882-7974.23.2.422 [DOI] [PubMed] [Google Scholar]
  69. Victor C, Scambler S, Bond J, & Bowling A (2000). Being alone in later life: Loneliness, social isolation and living alone. Reviews in Clinical Gerontology, 10(4), 407–417. 10.1017/S0959259800104101 [DOI] [Google Scholar]
  70. Vis B (2007). States of welfare or states of workfare? Welfare state restructuring in 16 capitalist democracies, 1985–2002. Policy & Politics, 35(1), 105–122. [Google Scholar]
  71. Walen HR, & Lachman ME (2000). Social support and strain from partner, family, andfriends: Costs and benefits for men and women in adulthood. Journal of Social and PersonalRelationships, 17, 5–30. 10.1177/0265407500171001 [DOI] [Google Scholar]
  72. Whelan K (2014). Ireland’s economic crisis: The good, the bad and the ugly. Journal of Macroeconomics, 39, 424–440. 10.1016/j.jmacro.2013.08.008 [DOI] [Google Scholar]
  73. Whisman MA (2010). Loneliness and the metabolic syndrome in a population-based sample of middle-aged and older adults. Health Psychology, 29(5), 550–554. 10.1037/A0020760 [DOI] [PubMed] [Google Scholar]
  74. Wrzus C, Hänel M, Wagner J, & Neyer FJ (2013). Social network changes and life events across the life span: A meta-analysis. Psychological Bulletin, 139(1), 53–80. 10.1037/A0028601 [DOI] [PubMed] [Google Scholar]
  75. Xia N, & Li H (2018). Loneliness, social isolation, and cardiovascular health. Antioxidants & Redox Signaling, 28(9), 837–851. 10.1089/ars.2017.7312 [DOI] [PMC free article] [PubMed] [Google Scholar]
  76. Yang K, & Victor C (2011). Age and loneliness in 25 European nations. Ageing and Society, 31(8), 1368–1388. 10.1017/S0144686X1000139X [DOI] [Google Scholar]

Associated Data

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

Data Availability Statement

Researchers interested in using TILDA data may access the data for free from the following sites: Irish Social Science Data Archive (ISSDA) at University College Dublin http://www.ucd.ie/issda/data/tilda/. Interuniversity Consortium for Political and Social Research (ICPSR) at the University of Michigan http://www.icpsr.umich.edu/icpsrweb/ICPSR/studies/34315.

RESOURCES