Development of Loneliness and Social Isolation After Spousal Loss: A Systematic Review of Longitudinal Studies on Widowhood

Abstract

Background:

Spousal loss is a stressful life event that is associated with loneliness and social isolation, both of which affect mental and physical health. The primary objective of this paper was to synthesize longitudinal studies that investigated loneliness and social isolation in widowhood.

Methods:

A systematic search of the literature was conducted using three electronic databases. 26 longitudinal studies published through June 2024 were included for further analysis. Participant characteristics, study design, and key findings were extracted.

Results:

Most studies were from the United States or Europe, included more widows than widowers, and assessed loneliness in older adults aged > 60 years. Loneliness peaked directly following spousal death, but findings were inconsistent regarding the lasting effects of widowhood. Heterogeneity in the longitudinal trajectories of loneliness was noted, with studies showing linear increases, decreases, or curvilinear relationships over time. Several factors modified the relationship between widowhood and loneliness, including volunteerism, military experience, income, and age. Widowers consistently reported greater loneliness and worse social isolation when compared to widows. Few studies investigated social isolation specifically, but those that did found that social isolation may decrease in widowhood.

Conclusions:

As the world grapples with a social pandemic of loneliness and social isolation, widowed adults may be uniquely affected. Few studies investigated the longitudinal trajectory of loneliness and especially social isolation in widowhood, and those that did found heterogenous results. Future work is needed to understand why some widowed adults are uniquely affected by feelings of loneliness and social isolation while others are not, and whether potentially modifiable factors that moderate or mediate this relationship could be leveraged by psychosocial interventions.

INTRODUCTION

Loneliness and social isolation are commonly experienced by older adults and have been linked to adverse mental and physical health outcomes. Approximately one-third of older adults experience loneliness later in life and half are at risk of social isolation.¹ Loneliness is a subjective feeling of perceived deficiencies in meaningful social relationships or companionship, while social isolation is an objective condition of a lack of social connections. Loneliness can be further categorized as social loneliness, the absence of a desired social network, or emotional loneliness, the absence of a close emotional attachment.² Prior research has connected loneliness and social isolation to increased risk of cardiovascular disease, frailty, and mortality as well as cognitive and functional decline.^3–5 Social isolation and loneliness are also major predictors of psychiatric disorders like anxiety and depression.^6–8 While research has consistently linked loneliness and social isolation to poorer physical and mental health, less is known about how loneliness and social isolation may develop after significant and traumatic life events (e.g., loss of a spouse and transition to widowhood).⁹

Chronic, unremitting, stress has been posited as a potential mediator between the relationships of loneliness, social isolation, and adverse health conditions in older adulthood.^10–12 Death of a spouse or romantic partner is one of the most significant stressors experienced by older adults.¹³ Widowhood has been associated with a variety of mental and physical health consequences and is a known risk factor for mortality.^14,15 Widowed individuals often experience depressive symptoms, loneliness, and lower life satisfaction.^14,16,17 Further, widowhood is associated with overall worse health status and higher rates of comorbidities.^14,18,19 Social health, the ability to interact and form meaningful relationships with others, is also negatively impacted by widowhood. In addition to increased loneliness, widow(er)s are at higher risk for social isolation and decreased emotional support, often due to shifting social networks post-loss.^20,21

Recent research has increasingly focused on potential protective factors that may mitigate the adverse psychosocial effects of widowhood. Several studies have found that increased social support and social participation may act as buffers against negative emotional and physical health outcomes linked to widowhood.^21–23 Prior literature has also found significant variability in the progression of loneliness and social isolation after spousal loss.^24,25 In general, loneliness tends to gradually decline as time passes post-loss,^26,27 but social support does not consistently moderate this relationship²⁷ and long-term patterns of loneliness tend to differ based on sociodemographic factors. Further, researchers have attempted to identify key social determinants of health (SDoH) that may moderate longitudinal trajectories of loneliness and social isolation in widow(er)s (e.g., military status, engagement in volunteerism, economic resources), but it is unclear which contextual factors are most influential.^23,28–30

One major methodological limitation consistently found in the widowhood and loneliness/social isolation literature is the over-reliance on cross-sectional rather than longitudinal study designs. By utilizing longitudinal data, investigators can better characterize the trajectory of loneliness and social isolation before, at the time of, and after spousal loss, and the moderating factors that influence this trajectory. Seeing this, the present systematic review aims to synthesize longitudinal studies on loneliness and social isolation in widowhood.

METHODS

To synthesize studies investigating loneliness and social isolation in widowed adults, a systematic review was conducted and reported using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. Due to the methodological diversity and heterogeneity between studies, a meta-analysis was not performed. This review was not pre-registered and a protocol was not published.

Literature Search

Three electronic databases (PubMed, PsycINFO, and Web of Science) were originally searched in June 2022 and an updated search was conducted in June 2024. Titles and abstracts were searched using the terms: (1) “widow*” OR “spous*” AND (2) “lonel*” OR “social isolat*” OR “social connect*” OR “social disconnect*” in all databases. Search terms were left broad to capture a large section of the literature.

Inclusion & Exclusion Criteria

To be included, studies had to be longitudinal, quantitative, and published or translated in English in a peer-reviewed journal between January 2000 and June 2024. Each study was required to enroll participants who experienced the death of their spouse/partner and measured loneliness and/or social isolation over time. Conference abstracts, dissertations/theses, book chapters, reviews, and interventions were excluded. Figure 1 displays the PRISMA flow diagram for the selection of studies for this review. The literature search yielded 2,346 results, of which 1,016 duplicate articles were removed. In accordance with PRISMA guidelines, 1,330 articles were screened by title/abstract after consulting the inclusion and exclusion criteria listed above. After evaluating 178 full-texts to confirm eligibility, 152 studies were excluded. 76 studies were not longitudinal, 19 did not assess widow(er)s, 15 did not specify the relationship between widow(er)s and loneliness and/or social isolation, 13 were qualitative, 13 did not assess loneliness and/or social isolation, 11 combined widow(er)s with those who were divorced or separated, and 5 were intervention studies. A total of 26 studies were included in this review.

Figure 1.

PRISMA flow diagram

Data Extraction

For each article, the following data were extracted: 1) study aim, 2) study design, 3) sample characteristics, 4) scale used to measure loneliness and/or social isolation, and 5) pertinent results. When relevant, statistical coefficients, p-values, and effect sizes were reported. Although multiple studies investigated other bereavement outcomes (e.g., depression, stress, grief, anxiety) or other risk factors for loneliness in the absence of widowhood (e.g., self-rated health, educational level, divorce, retirement), analysis was restricted to loneliness or social isolation outcomes in widowed adults. Two authors (KN, MAP) confirmed extracted data for this systematic review for increased validity.

Study Quality

Study quality was assessed with the Newcastle–Ottawa Scale for longitudinal cohort studies.³¹ The scale evaluates studies on selection of study groups, comparability of groups, and ascertainment of outcome, assigning a total score of up to 9 points. Study quality cutoffs are as follows: rating 0–2 (poor quality), 3–5 (fair quality), 6–9 (good/high quality).³¹ Quality assessment was completed by two reviewers (KN, MAP), and a consensus was obtained for each included study.

RESULTS

Characteristics of Longitudinal Cohort Studies

Table 1 summarizes the main findings of the 26 longitudinal studies. See the Supplemental Material for more detailed findings. Most studies were conducted in Europe (50%) or the United States (23.1%). Two studies took place in China, with another two in Australia, and one each in Thailand and South Africa. One study looked at data from cross-national studies representing 20 countries from around the world. The sample size ranged from 99 to 89,423 participants with the majority (80.7%) including participants aged 60–80 years. One study did not specify participants’ age. Majority of participants identified as female in most studies (84.6%). One study focused exclusively on widows (women)²³ while another study only included widowers (men).³² Participants were followed for 14 months to 28 years, including articles that incorporated a pre-widowhood period. One-fifth (19.2%) of the included studies did not incorporate a pre-widowhood period. See the Supplemental files for individual study details.

Table 1:

Key Characteristics of Included Studies

Study	N = widow(er)s	N = comparison group	Pre-Widowhood Observation Duration	Post-Widowhood Observation Durationa	Key Outcome Measures	Modifiers of Loneliness and/or Social Isolation
Robinson-Whelen 2001	49	94	1 year	2–3 years	3-item NYU Loneliness Scale	Caregiving
Van Baarsen 2001	99		Up to 7 months	2.8 years	11-item De Jong Loneliness Scale	None
Jylha 2004	787*		n/a	Up to 20 years	Single item question	None
Aartsen 2011	202		< 28 years	Up to 28 years**	Single item question	None
Caserta 2013	325		n/a	15 months	13-item UCLA Loneliness Scale	Death expectedness
Dahlberg 2015	587*		< 7 years	Up to 7 years**	Single item question	None
Brittain 2017	436	314	n/a	3 years**	Single item question	None
Carr 2018a	592	5057	< 4 years	Up to 4 years	3-item survey	Volunteerism
Carr 2018b	2148*		< 4 years	Up to 4 years	3-item survey	Military experience, death exposure
Szabo 2020	686		12 years	12 years	11-item De Jong Loneliness Scale	None
von Soest 2020	5555*		n/a	5 years**	3-item De Jong Loneliness Scale	None
Buecker 2021	164	164	5 years	5 years	6-item De Jong Loneliness Scale	Age
Eckhard 2021	4595		Not specified	Not specified, but > 8 years	3-item survey	Gender
King 2021	428		2–4 years	Up to 2 years	3-item UCLA Loneliness Scale	Partner veteran status
Nyqvist 2021	816	3260	< 6 years	Up to 6 years**	Single item question	None
Yang 2021	258	2446	< 3 years	Up to 3 years	Single item question	Gender, education
Freak-Poli 2022	749	7885	3 years	3 years	3-item survey	Income
Kamp 2022	310*		n/a	6–20 months	3-item UCLA Loneliness Scale	None
Sun 2022	17,429*		< 4 years	Up to 4 years	Face-to-face interview	Age, gender
Nicolaisen 2022	2315*		< 15 years	Up to 15 years	Single item question	Age
Pengpid 2023	2863*		< 5 years	Up to 5 years**	Single item question	None
Sheftel 2023	89,423*		Not specified	Not specified	Single item question	None
Jiao 2024	14933	7844	< 16 years	16 years**	Single item question	Age, gender
Kapelle 2024	764	7388	2–3 years	Up to 19 years	Single item question	Gender
Kotwal 2024	2098*		2 years	2 years	Single item question	Partner cognitive impairment
Pengpid 2024	3707*		< 7 years	Up to 7 years**	Single item question	None

^aDuration since incident widowhood

^*Not separated out by control group versus comparison group

^**Included participants who were widowed at baseline with an unspecified duration of widowhood

65.4% of the studies had married and/or non-widowed participants as their control group, while 23.1% had no control group. All studies identified loneliness and/or social isolation as an outcome of interest, whether primary or secondary. Various measures of loneliness and social isolation were used, including a single item measure of loneliness and/or social isolation (50%), versions of the DeJong Loneliness Scale (15.4%), a multi-item measure of loneliness and/or isolation (15.4%), and 3-item versions of the UCLA Loneliness Scale (7.7%). Isolated studies used a 13-item version of the UCLA Loneliness scale, the NYU Loneliness Scale, and face-to-face interviews about a participant’s core network. While 25 out of the 26 studies investigated loneliness, only three studies looked at social isolation.

Quality Assessment

Quality of included longitudinal studies was assessed, and most were found to have limitations. Quality scores ranged from 4 to 8 with a mean of 6.08. Of the 26 studies, 18 (69%) were high quality and 8 (31%) were fair quality. See full assessments in Table 2.

Table 2:

Quality Assessment of Included Longitudinal Studies

	Selection				Comparability	Outcome			Quality Score
Study	Representativeness of exposed cohort	Selection of non-exposed cohort	Ascertainment of Exposure	Outcome of interest not present at start of study	Comparability of cohorts on basis of design or analysis	Assessment of outcome	Was follow-up long enough for outcomes to occur	Adequacy of follow-up of cohorts
Robinson-Whelen 2001	*	*	*			*	*	*	6
Van Baarsen 2001	*		*			*	*	*	5
Jylha 2004	*	*	*				*	*	5
Aartsen 2011	*	*	*	*			*		5
Caserta 2013	*		*		*	*	*	*	6
Dahlberg 2015	*	*	*				*	*	5
Brittain 2017	*	*	*				*	*	5
Carr 2018a	*	*	*		**	*	*	*	8
Carr 2018b	*	*	*		**	*	*	*	8
Szabo 2020	*		*			*	*	*	5
von Soest 2020	*	*	*			*	*	*	6
Buecker 2021	*	*	*		*	*	*	*	7
Eckhard 2021	*		*		*		*		4
King 2021	*	*	*		**	*	*	*	8
Nyqvist 2021	*	*	*		**		*	*	7
Yang 2021	*	*	*	*	*		*	*	7
Freak-Poli 2022	*	*	*		*		*	*	6
Kamp 2022	*		*		*	*	*	*	6
Sun 2022	*	*	*		**		*	*	7
Nicolaisen 2022	*	*	*		*		*	*	6
Pengpid 2023	*	*	*		**		*		6
Sheftel 2023	*	*	*	*	*		*	*	7
Jiao 2024	*	*	*		*		*	*	6
Kapelle 2024	*	*	*		*		*	*	6
Kotwal 2024	*		*		*		*	*	5
Pengpid 2024	*	*	*		*		*	*	6

Longitudinal Cohort Study Findings and Themes

Loneliness following widowhood:

Multiple studies consistently showed that widowhood was associated with higher levels of loneliness directly following the loss of one’s spouse/partner.^26,32–44 However, studies yielded inconsistent findings on the lasting effect of widowhood on loneliness. Seven studies found a strong relationship between widowhood and unremitting loneliness over time,^{34,37,39–41,45,46} another five found that loneliness diminished after a few years following widowhood,^{36,42,47–49} and one study found no change in loneliness over time.⁵⁰

Heterogeneity in the trajectory of loneliness:

Two studies identified various trajectories of loneliness following bereavement.^24,25 Szabo et al.²⁴ identified five trajectories of emotional loneliness, including recovery (32.3%, low pre-bereavement emotional loneliness with a recovery to baseline after peaking after loss), resilient (13.5%, low pre- and post-bereavement), prolonged (17%, low pre-bereavement with a prolonged increase post-bereavement), increased and prolonged (28.3%, increasing loneliness starting from pre-bereavement), and chronically high (8.9%, high pre- and post-bereavement). Van Baarsen et al.²⁵ identified four trajectories of emotional loneliness, including ascending-descending (28.3%, similar to recovery), low stable (33.3%, similar to resilient), ascending (30.3%, similar to prolonged or increasing and prolonged), and descending. For social loneliness, Szabo et al.²⁴ identified four trajectories, including very low (43.3%), low (27.5%), increasing (20.2%), and chronically high (9%). Van Baarsen et al.²⁵ similarly identified a low-stable trajectory (69.7%), suggesting that most widow(er)s in both studies followed trajectories with low social loneliness levels throughout bereavement. However, van Baarsen et al.²⁵ also identified descending (9.1%), ascending-descending (13.1%), and variable (8.1%) trajectories, of which there are no comparable trajectories from Szabo et al.’s study.

Moderators:

Caregiving:

Former caregivers who were widowed and current married caregivers experienced similar levels of loneliness, suggesting that caregiving itself may contribute to increased loneliness.⁵¹

Partner cognitive impairment:

Prior to widowhood, spouses of individuals with cognitive impairment or dementia reported greater loneliness compared to spouses of individuals with non-impaired cognition.⁵² However, in the post-bereavement period, spouses of partners across all cognitive statuses experienced similarly high loneliness.⁵²

Anticipatory grief:

The circumstances surrounding the spouses’ death, such as whether it was expected or due to cancer, also influenced loneliness.⁴⁸ Deaths that were anticipated and unrelated to cancer were associated with lower levels of loneliness among widowed adults when compared to unexpected deaths and deaths due to cancer.

Volunteering:

Volunteering >100 hours per year among widow(er)s was associated with similar levels of loneliness as those who were continuously married and engaged in similar volunteer work.³⁵

Military:

Military experience with exposure to death among widowed male Veterans was associated with lower levels of loneliness compared to civilian widowers, and this effect was not solely explained by social engagement.³² Additionally, widows of Veterans experienced lower levels of loneliness compared to widows of civilians, but this was primarily explained by perceived friend social support.²³

Income:

Higher income was associated with a smaller increase in loneliness and a decline in social isolation for widowers, but no significant effects were observed for widows.³⁹

Age:

One study by Buecker et al. found no significant age-normative effects regarding widowhood and loneliness.⁴⁵ Regardless of age at spousal loss, widowhood had a similar effect on the development of loneliness in this event group. A different study by Jiao et al. found that widowhood’s effect on loneliness, both immediately following bereavement as well as longitudinally, decreases with age.²⁶ Similarly, a study by Sun and Schafer looking at core networks in participants aged 50 years or older found that younger individuals were less likely to replenish their core network following widowhood compared to older individuals.⁵³ In contrast to the former study by Buecker et al., these latter two studies demonstrate the larger effect that widowhood may have on loneliness at earlier ages.

Gender:

Dahlberg et al.³⁴ found that incident widowhood significantly predicted loneliness at follow-up for both men and women, but the relationship was stronger in men (odds ratio [OR] = 18.11, 95% confidence interval [CI] = 5.25–62.39) compared to women (OR = 3.35, 95% CI = 1.38–8.15). This was also demonstrated by Pengpid et al.’s⁴⁰ study: men (OR = 1.92, 95% CI = 1.61–2.29) and women (OR = 1.45, 95% CI = 1.27–1.66), and by Pengpid and Pelzer⁴¹ (men: OR = 2.64, 95% CI = 2.15–3.24, women: OR = 1.47, 95% CI = 1.26–1.71). Similarly, Freak-Poli et al.³⁹ found that in the first year following spousal death, women experienced a two-fold increase in loneliness, while men experienced a three-fold increase compared to the two years prior to bereavement. Following widowhood, Kapelle and Monden⁴² demonstrated that men’s loneliness levels remained elevated after up to 19 years while women’s loneliness continued to decline. Jiao et al.²⁶ found that in at younger ages, widowhood had a stronger effect on loneliness in men, but the gender gap narrowed with age. Lastly, Eckhard⁵⁴ found that both men and women experienced an increased risk of social disconnectedness, including living alone, even after 8 years post-loss. However, meetings with friends and relatives and involvement in clubs and organizations increased for women, but not for men. Sun and Schafer⁵³ discovered similar results, showing that more women (90%) than men (less than 60%) added new connections to their previously partner-exclusive core network following partner death.

Social isolation and disconnectedness:

Living alone increased following partner loss⁵⁴ while meetings with friends and relatives and involvement in clubs or organizations actually increased for women. Freak-Poli et al.³⁹ also a decrease in social isolation following widowhood, and this reduction in social isolation may last after three years. Sun and Schafer⁵³ reported that partner loss due to death was associated with a higher probability of adding new connections to their core network in the years following widowhood.

DISCUSSION

This review summarized the existing literature exploring loneliness and social isolation in widowhood through examination of prospective and retrospective cohort studies. Although the included longitudinal studies differed in method and focus, results revealed several overarching themes that may be useful in informing clinical practice. Overall, studies concluded that widowhood was a predictor of loneliness post-loss, but the duration and severity of this relationship remains heterogenous. Studies identified various trajectories of loneliness after spousal loss, with some experiencing increasing or decreasing loneliness and others remaining stable, but no clear pattern could be discerned from the included studies. Regarding social isolation, most of the included longitudinal studies reported a decrease in widowhood, which was attributed to strengthening one’s social network post-loss and engaging in social clubs and organizations.

One of the strongest moderators of loneliness and social isolation was gender. Studies found that loneliness was greater among widowers than widows.^34,39,55 While widowers experienced less change overall within their social network, widows were more likely to increase social engagement after spousal loss, which may explain the discrepancy in loneliness magnitude between genders.⁵⁴ Eckhard et al⁵⁴ and Sun and Schafer⁵³ suggested that a gender disparity in social engagement may widen post-loss, with women strengthening social connections and joining social clubs as widowed adults more often than men. Other major moderators included the context of the spouse’s health pre-loss and the anticipation of death. Widowed adults who provided caregiving duties to their spouse experienced greater loneliness, and this was further complicated by whether the spouse had a cognitively deteriorating condition, cancer, or if the death was anticipated or not. These findings underscore the importance of ascertaining the social context surrounding loss among widowed adults in a clinical setting, and providing psychosocial resources unique to the widowed adult’s specific situation to reduce the burden of loneliness, social isolation, and their associated mental health and physical health sequelae.

It is important to note that methodological discrepancies across studies, such as differences in follow-up duration, likely contributed to the varied findings regarding loneliness trajectory. The included studies varied in the length of follow-up of loneliness and social isolation, but also in the length of time since the death of their spouse. Future studies should include a pre-widowhood period where loneliness and/or social isolation were measured before the death of the spouse. This quasi-experimental design would allow for investigators to determine if a change-point exists (i.e., whether the trajectory or “slope” of loneliness and/or social isolation differs in the pre-widowhood and post-widowhood periods). The age at which the participant experienced the death of their spouse should also be considered. These individual differences (i.e., how long loneliness and/or social isolation were measured, length of time since the death, age at spousal/partner death, gender, race and ethnicity, etc.) are difficult to tease apart in a single study, but may contribute significantly to the heterogeneity seen in the development of loneliness and/or social isolation post-loss. Despite issues with comparability, longitudinal studies are essential for understanding the long-term effects of widowhood on loneliness and social isolation.

Several longitudinal studies examined SDoH that may moderate the association between widowhood and loneliness or social isolation such as military status, economic resources, and exposure to death.^{23,32,34,39,54} Military history with exposure to death was associated with less loneliness among Veteran widowers compared to civilian widowers.³² Additionally, widows of Veterans experienced lower levels of loneliness post-loss, but this was primarily explained by greater social support from friends²³. Further investigation is needed to understand the possible psychosocial mechanisms behind these SDoH. One possible explanation may be resiliency, a psychosocial trait defined as the capacity for adaptation in the face of adversity, that has the capacity to be enhanced through targeted interventions⁵⁶. Prior studies have found that higher resilience is related to better widowhood outcomes (e.g., likelihood of re-partnering, reduced depression)^57–59 and that resilience may buffer loneliness in older adults.^60–63 Future clinical interventions targeting loneliness and social isolation in widow(er)s should consider resilience training.

In recent years, there has been growing support for using positive psychiatry interventions to address loneliness, depression, and stress in older adults.^1,60,63,64 In contrast to typical psychosocial treatments like cognitive behavioral therapy (CBT) which target deficits or maladaptive behaviors and thinking patterns, positive psychiatry interventions primarily focus on improving positive psychosocial traits to increase overall wellbeing and, in turn, decrease negative mental health outcomes. Psychosocial traits like emotional regulation, resilience, and self-compassion have been identified as possible protective factors against loneliness, and interventions enhancing these characteristics have shown promising results.^60,63,65,66 However, these types of interventions have not yet been modified for the unique stressors of widowed adults, nor the gender-specific experiences and needs of widows versus widowers. Future interventions should consider adapting positive psychiatry interventions to include grief-specific support and coping strategies.

Study Quality

Of the 26 included studies, 69% were assessed as high quality and 31% as fair quality. The lower quality studies shared two common limitations which lowered their scores on the Newcastle-Ottawa Scale.³¹ First, studies did not confirm the absence of the outcome of interest (loneliness and/or social isolation) at baseline. Given the prevalence of loneliness among older adults, this requirement may be unrealistic and likely inconsequential to the overall findings of this review. Second, studies used non-validated scales to measure the outcome (most frequently, a single-item measure of loneliness). However, while validated scales (e.g., UCLA Loneliness Scale) are the gold standard, single-item measures of loneliness are found to closely correlate with longer loneliness scales.⁶⁷

Although the experience of spousal loss is a universal phenomenon, the cultural, economic, and social resources that one has at their disposal to buffer against loneliness and social isolation are bound by various social identities (e.g., gender identity, race, ethnicity, religion, sexual preference). Therefore, practitioners seeking to treat loneliness and social isolation in widowed adults must carefully consider these multifaceted constructs among other SDoH that impact a patient’s risk of developing adverse mental health or physical health sequelae. Finally, more work is needed to understand how the DSM-5-TR diagnostic criteria for prolonged grief disorder (PGD)⁶⁸ would map onto our findings. Conceivably, those with the most extreme and unremitting loneliness may be more likely to meet criteria for PGD than those who experienced normative declines in loneliness with the passage of time.

Limitations of the Present Review

Few studies differentiated between subtypes of loneliness. Covariates also varied between studies (e.g., age, sociodemographic factors, health status, social engagement). Few studies investigated social isolation and focused almost exclusively on loneliness. While these two constructs are distinct and may operate differently post-loss, change in loneliness should be considered when measuring change in social isolation and vice versa. Due to heterogeneity in the study design and statistical analyses used by each article, meta-analyses were not feasible.

CONCLUSION

We examined the state of the literature on loneliness and social isolation in widowhood through the lens of longitudinal cohort studies. Widowhood appears to increase loneliness in older adults, but the duration and strength of this relationship is heterogenous. Conversely, this review suggests social isolation may decrease after spousal loss, but a consensus on the most accurate way to assess social isolation in the context of widowhood is needed. Widowers appeared to be more affected than widows post-loss for loneliness, and widowers were less likely than widows to form new connections to alleviate social isolation. Further research is necessary to identify possible protective factors of loneliness and social isolation in widowhood as well as to develop interventions specific to the experiences and unique needs of aging widows and widowers.

Supplementary Material

Supinfo

Supplementary Table 1. Data Extracted from Longitudinal Studies on Loneliness and/or Social Isolation in Widowhood

Key points:

• Heterogeneity between studies was detected regarding longitudinal trajectories of loneliness in widowed adults.

• Social isolation may decrease following spousal loss, although few studies specifically investigated social isolation in the context of widowhood.

• Some modifiers of the relationship between widowhood and loneliness or social isolation included military status, prior exposure to death, gender, age, economic resources, and volunteerism.

Why does this paper matter?

This study indicates the need for further research on the longitudinal impact of widowhood on loneliness and especially social isolation in older adults. Future studies should also investigate factors that moderate the relationship between loneliness and/or social isolation in widowhood, including which factors predominantly affect this relationship.