Initially, 679 publications were retrieved, including books and book chapters. Duplicate publications among the three databases were removed, which left 572 potentially relevant publications. After reviewing the title and abstracts for relevance based on the inclusion criteria, the number of publications was reduced to 92. These publications were then read in their entirety, and 18 were eliminated. After comprehensively re-reading the remaining 74 publications, 4 were removed because they did not yield data relevant to the concept of social isolation; thus, 70 publications were retained for final review (Fig. 1).
Results will be organized by reviewing the negative health aspects of social isolation guided by the ubiquitous conceptual framework developed by Berkman, Glass, Brissette, and Seeman, (2000), which includes health behavioral, psychological, and physiological categories. Known negative health aspects of social isolation that did not fit into the Berkman et al. framework are presented subsequently. Next, variables associated with social isolation are summarized according to categories outlined by Howat, Iredell, Grenade, Nedwetzky, and Collins (2004). The categories include: (1) physical; (2) psychological; (3) economic; (4) work/family changes; and (5) environmental. A demographic category was also included as an overview of negative health exchanges. Finally, an overview of assessment of social isolation is presented followed by a brief review of interventions.
Known Negative Health Consequences of Social Isolation
There is a wealth of evidence regarding the negative health outcomes predicted by social isolation. According to the Berkman et al. (2000) framework, limited social networks impact health downstream through three pathways: (1) health behavioral, (2) psychological, and (3) physiological. In this review article, these pathways were used to organize negative health outcomes related to limited social networks, or social isolation. There were certain negative health outcomes of social isolation that did not fit into the Berkman et al. framework, including all-cause mortality, falls, re-hospitalization, and institutionalization; these outcomes are described separately. Most topics related to social isolation fit into one of the three categories of the Berkman et al. framework, but those that did not will be placed in the Other Outcomes category.
Social isolation impacts the health and behavioral habits of older adults. An older adult's social network can impact health positively through encouragement to adhere to medical treatment or to refrain from negative or risky behaviors (Berkman et al., 2000). Without the positive influence of social network members, older adults who are socially isolated are at risk for many negative behaviors such as heavy drinking (Hanson, 1994), smoking, and being sedentary (Eng et al., 2002). Older adults who are socially isolated also have an increased nutritional risk (Locher et al., 2005).
Social isolation has been demonstrated to impact the psychological and cognitive well-being of older adults. Those who have poor social connections and do not participate in social activities are at an increased risk of cognitive decline (Beland, Zunzunegui, Alvarado, Otero, & Del Ser, 2005). Less socially connected men are at a significantly increased risk of death from suicide, as well as from other causes (Eng et al., 2002). Conversely, older adults who have an extensive social network are more protected against dementia (Fratiglioni et al., 2004; Wang, Karp, Winblad, & Fratiglioni, 2002).
The physiological effects of social isolation in the geriatric population are well documented. It is striking how much evidence exists on social isolation as a predictor of mortality from coronary heart disease/stroke (Boden-Albala, Litwak, Elkind, Rundek, & Sacco, 2005). Other physiological afflictions resulting from social isolation, such as contracting common colds (Cohen, Doyle, Skoner, Rabin, & Gwaltney, 1997), have also been documented.
There are many aspects of an older adult's physical and psychosocial life affected by social isolation that do not fit into the three categories of the Berkman et al. framework. These include all-cause mortality, falls, re-hospitalization, and institutionalization. Socially isolated individuals are at an increased risk for all-cause mortality (Berkman, 1984; Eng et al., 2002), which is defined as "the annual death rate or mortality rate from all causes" (Gordis, 2009, p. 62). Conversely, it has been suggested that social networks with greater numbers of friends are protective against mortality (Giles, Glonek, Luszcz, & Andrews, 2005). Social isolation among older adults has also been associated with an increased number of falls (Faulkner et al., 2003). Those older individuals who are socially isolated are also four to five times more likely to be re-hospitalized within one year of original admittance (Mistry et al., 2001). In addition, social isolation is a major predictor of institutionalization (Brock & O'Sullivan 1985); conversely, larger social networks are associated with a lower risk of institutionalization (Colantonio, Kasl, Osfeld, & Berkman, 1993). The overwhelming majority of data collected as part of this literature review focuses on outcomes related to social isolation, rather than risk factors for or predictors of this condition. A graphic representation of how these variables fit together is shown in Fig. 2.
Potential Risk Factors of Social Isolation: Associated Variables
Numerous variables have been associated with social isolation. The individual-level variables found to be potential risk factors for social isolation in cross-sectional studies are summarized below in categories outlined by Howat et al. (2004). The categories include: (1) physical; (2) psychological; (3) economic; (4) work/family changes; and (5) environmental. In addition to these categories, pertinent demographic variables were added as a category of risk factors for social isolation.
It is possible that individuals with a poor body image attributable to being overweight may decrease or cease interactions with their social networks to the point where they could be at risk for social isolation. For example, individuals who are overweight may be self-conscious or embarrassed and, therefore, less likely to engage their social networks. Alternatively, many social engagements may center around food, so individuals who are focused on maintaining a healthy weight may avoid interactions with those who are overweight. It has been demonstrated that weight gain in one person is associated with weight gain in friends, siblings, and spouses (Christakis & Fowler, 2007).
In addition to being overweight, having a higher number of health problems has been associated with social isolation (Havens, Hall, Sylvestre, & Jivan, 2004). Havens et al. (2004) found that older adults who have four or more chronic illnesses were at risk for social isolation (χ 2 = 16.46, p < .05). Moreover, Havens and Hall (2001) found that older adults with four or more chronic illnesses were 1.7 times more likely to be socially isolated than those who had fewer than four chronic illnesses. An examination of co-morbidities is necessary, as chronic illness may be a significant risk factor for social isolation in older adults. Knowing the frequency and number of co-morbidities will aid in the assessment of social isolation by adding important information regarding disease severity and burden.
Other health issues such as sensory loss, including impaired vision and hearing, have been associated with communication disruptions in older adults (Heine, Erber, Osborn, & Browning, 2002), which over time could potentially lead to social isolation. Older adults with poor vision may also have limited opportunities for social interactions with others (Jang et al., 2003).
Additionally, older adults with untreated hearing impairment are less likely to be involved in social activity (The National Council on the Aging, 1999). Those with the most severe hearing loss are 24 % less likely to be socially integrated than those without a hearing impairment (The National Council on the Aging, 1999). In an older study examining lack of treatment of hearing loss in relation to social isolation, hearing loss was associated with a decrease in effective social functioning (Weinstein & Ventry, 1982), which was defined as the basic ways of interacting with others, such as speaking or listening. In a more recent study of hearing impairment in old age, individuals with a hearing impairment had a dose–response decrease in social functioning compared to those with no hearing impairment (p < .01) (Strawbridge, Wallhagen, Shema, & Kaplan, 2000).
Urinary incontinence has also been found to limit social activities. In a study of urinary incontinence among older adults, incontinence resulting in large volumes of urine loss was 5.6 times more likely to negatively influence their lifestyle (including social activities) than incontinence resulting in small volumes of urine loss (OR = 5.61, 95 % CI [2.76, 11.42) (Gavira Iglesias et al., 2000). In a study examining potential social and emotional limitations of urinary incontinence, older adults who were incontinent were more likely to be socially isolated (χ 2 = 31.06, p < .001) (Fultz & Herzog, 2001). A literature review found that urinary incontinence had a significant negative impact on social engagement among community-dwelling older adults (Wyman, Harkins, & Fantl, 1990).
In a study examining sleep complaints within the Established Populations for Epidemiologic Studies of the Elderly (EPESE) cohort, 23–34 % of adults had symptoms of insomnia (Foley et al., 1995). If insomnia persists over time, older adults may cease to meet new social network members or even lose those they already have, resulting in social isolation. Aspects of sleep quality and daytime alertness may be impacted by psychosocial variables such as social support (Driscoll et al., 2008) and may impact older adults' ability to engage others socially.
Having fewer social network members has been associated with heavy drinking and having a drinking problem in elderly men (Hanson, 1994). Those individuals who have limited social networks are 2.5 times more likely to be heavy drinkers (OR = 2.5, 95 % CI [1.6, 3.8]) (Hanson, 1994). Although excessive alcohol consumption is a complicated problem, it is plausible that those who drink excessively may lose social network members because of the problems associated with binge drinking and alcoholism.
Although no studies examining social isolation and smoking were found, social networks have been suggested as an important aspect of smoking cessation programs (Hanson, Isacsson, Janzon, & Lindell, 1990). Individuals who have robust social networks are three times more likely to succeed in quitting smoking for good (OR = 3.1, 95 % CI [1.9, 5.4]) (Hanson et al., 1990). Just as there is a relationship between having a supportive social network and successful smoking cessation, there may be a relationship between smokers and their social desirability. With current bans on smoking in public places and related attitudes, there may be increased social isolation among smokers due to their potential exclusion from some social networks.
Social engagement may influence the functional disability process by preventing decline or facilitating recovery (Mendes de Leon et al., 1999). Consequently, functional disability may impact social networks by preventing older adults from seeking engagement with other social network members. Network size and social interaction are significantly associated with functional disability risk (Mendes de Leon, Gold, Glass, Kaplan, & George, 2001). Older adults who are more socially engaged report less functional disability (OR = 0.84, 95 % CI [0.75, 0.95]) (Mendes de Leon, Glass, & Berkman, 2003), and those who are a strong part of a social network have been found to have reduced risk of functional disability (β = −0.009, p < .01) (Mendes de Leon et al., 1999).
There is a close relationship between the concepts of social isolation and depression; however, that relationship is not well understood. Given what is known about depression and social isolation, it makes sense that there is a relationship between the two; however, the nature of that relationship is not clear. Blazer (2005) states that frustration well in stating that depression has a "clear, but not obvious relationship" with social isolation (p. 497). There is no consensus in the research that is available regarding the relationship between depression and social isolation. Some researchers have found strong correlations between social isolation and depression (Dorfman et al., 1995), whereas others have not (Schoevers et al., 2000). In a qualitative study examining social workers' views of depression, almost all of those interviewed perceived depression as a key cause of social isolation in older adults (McCrae et al., 2005). In a quantitative study, Iliffe et al. (2007) found that older adults in a depressed state were at significant risk for social isolation.
There are many studies touting the link between increased levels of social support and increased quality of life (Gallicchio, Hoffman, & Helzlsouer, 2007). However, there are fewer studies examining the relationship between social isolation and quality of life. It has been suggested that being socially isolated impacts quality of life (Lim & Zebrack, 2006). In one study, it was found that the social network structures (including network size and reliance on formal and informal social ties) were associated with quality of life in long-term cancer survivors (Lim & Zebrack, 2006). In another study, women who were socially isolated prior to their cancer diagnosis were more adversely impacted by breast cancer (Michael, Berkman, Colditz, Holmes, & Kawachi, 2002). The authors concluded that prediagnosis level of social integration is an important factor in future health-related quality of life among breast cancer survivors (Michael et al., 2002).
Religious engagement also appears to be an important aspect of the social networks of older persons. In the social network index, church membership was part of the scoring structure used to determine if an individual was socially isolated or not (Berkman & Syme, 1979). Little research has been published in the literature regarding the number of people that an individual knows in their church and its relationship to social networks or social isolation. Most available research pertains to the number of times an individual attended church. An older study found that religious background protected against mortality among older adults with poor health (Zuckerman, Kasl, & Ostfeld, 1984). Those older adults who were not religious were 2.32 times more likely to die (Zuckerman et al., 1984). Conversely, those frequently attending religious services have been found to have lower mortality rates than those with infrequent attendance (Strawbridge, Cohen, Shema, & Kaplan, 1997). In that study, a Cox proportional hazards model showed that those who were frequent attendees of religious services had a lower mortality rate than those who were infrequent attendees (relative hazard = 0.64, 95 % CI [0.53, 0.77]) (Strawbridge et al., 1997).
There are also a number of studies that discuss the relationship between social isolation and cognitive decline in older persons (Bassuk, Glass, & Berkman, 1999; Havens et al., 2004). A decrease in social engagement may have a negative effect on older adults' cognition (Zunzunegui, Alvarado, Del Ser, & Otero, 2003), and not participating in leisure activities has been found to be an antecedent to lower cognition (Wang et al., 2002). Lack of a robust social network is also a significant precursor to cognitive decline (Wang et al., 2002). Conversely, an active and socially integrated lifestyle has been found to protect against dementia (Fratiglioni et al., 2004).
One study examining a cohort of workers showed that social isolation may predict early retirement (Elovainio et al., 2003). Specifically, women with small social networks were approximately five times more likely to retire early than women with large social networks (OR = 5.1, 95 % CI [2.8, 9.2]). Social networks and retirement are clearly related. There is reason to believe that retirement may also lead to a decrease in social networks and social network contact, thus leading to social isolation. Retirement may be a stressful event for older adults, who are used to working and depending on the support of social network members at work.
Changes in Work and Family Roles
It has been demonstrated that the loss of a relative, friend, or close neighbor may lead to an increase in social isolation among older people (Wenger & Burholt, 2004). A stressful negative life event, such as the death of a close friend or relative, may prevent older adults from engaging their social networks, therefore placing them at an increased risk for social isolation.
Older adults who live in neighborhoods where safety is a concern may be at an increased risk of becoming socially isolated (Ross & Jang, 2000). There is a significant association between neighborhood disorder and social ties (Ross & Jang, 2000). Communities with high levels of social disorder are characterized as having poor safety, high levels of vandalism, and increased incivilities (Lewis & Salem, 1986).
A large number of older adults in the United States, about 30 % of the elderly population (or 10.9 million people), lives alone (Fowles & Greenberg, 2009). Living alone has been found to be a risk factor for a decrease in social networks or an increase of social isolation (Berkman, 2000; Havens et al., 2004; Howat et al., 2004; Iliffe et al., 2007; Lubben & Gironda, 2003). A study of older adults living alone in China found that they did not need any help solving problems and they learned to be flexible in order to reduce the need of being dependent on others (Tsai & Tsai, 2007). Thus, older adults who do not depend on others for resources may be more likely to get their social needs met, which is in direct contrast with the majority of research articles found in the literature.
Aging has been found to be a potential risk factor for social isolation (Iliffe et al., 2007). There are few prospective studies examining the relationship between aging and social factors such as social isolation; however, available studies indicate that it is necessary to measure aging in older adults (Mendes de Leon et al., 2003).
It has been suggested that race impacts the characteristics of social networks including size, frequency of contact, and composition (Ajrouch, Antonucci, & Janevic, 2001; Peek & O'Neill, 2001). A number of significant differences in social network characteristics can be attributed to race. These include a higher family composition, as well as more support received and less support given, among African Americans as compared to among Whites (Peek & O'Neill, 2001). Also, African Americans have more frequent contact with smaller networks made up of primarily family members (Ajrouch et al., 2001). In terms of health outcomes, those who are most socially connected are three times more likely to undergo colorectal screening (OR = 3.2, 95 % CI [1.7, 6.2]), with this association being stronger in African Americans than in Whites (Kinney, Bloor, Martin, & Sandler, 2005).
Social isolation has been shown to correlate with changes in salary and socioeconomic status. Economic constraints (Iliffe et al., 2007) and low income (Bassuk et al., 1999) are both factors associated with social isolation along with inadequate personal resources (Ackley & Ladwig, 2010). Socioeconomic status has been suggested as a potential risk factor for social isolation (Havens & Hall, 2001; Iliffe et al., 2007).
Level of education has also been suggested as a risk factor for social isolation (Iliffe et al., 2007). One study found that older adults with less than 12 years of education were 1.6 times more likely to become socially isolated than those with 12 or more years of education (OR = 1.6, 95 % CI [1.3, 1.9]) (Bassuk et al., 1999).
There is a significant amount of literature on the benefits of marriage to an individual's health. A spouse is seen as a special type of social network member who is invaluable to an individual's overall well-being. The loss of an intimate partner or spouse may be a strong factor leading to social isolation (Chipperfield & Havens, 2001) and decreased life expectancy in men (House, Landis, & Umberson, 1988). An association between being unmarried and negative outcomes related to social isolation has been repeatedly found (Boden-Albala et al., 2005; Fratiglioni, Wang, Ericsson, Maytan, & Winblad, 2000).
Summary of Variables Associated With Social Isolation
Numerous variables have been found to be associated with social isolation. Additionally, most research has focused on social isolation as an independent variable that leads to negative health outcomes, but not much has examined social isolation itself as an outcome (Smith & Hirdes, 2009). As a result, there is limited information regarding predictors of social isolation (see below). The literature that is available and has been discussed above provides an overview of the variables associated with social isolation.
Variables That Predict Social Isolation
Only a few studies have specifically examined predictors of social isolation (Havens & Hall, 2001; Havens et al., 2004; Howat et al., 2004; Iliffe et al., 2007; Luggen & Rini, 1995; Smith & Hirdes, 2009; Wenger & Burholt, 2004). Table 1 provides a summary of the specific risk factors of social isolation based on findings from these studies.
Negative Social Relationships
Some researchers have found that certain social relationships, including marriage, can actually have a negative impact on health (Antonucci, Akiyama, & Lansford, 1998; De Vogli, Chandola, & Marmot, 2007; Pagel, Erdly, & Backer, 1987; Seeman, 2000). These negative social exchanges are harmful and may lead to negative health outcomes rather than positive benefits from having close-knit social network members. Therefore, simply knowing if an individual has numerous social relationships does not provide enough information to accurately determine whether these relationships are positive and beneficial. Having negative relationships with social network members can be a source of additional demands, responsibilities, conflicts, embarrassment, and disappointment (Seeman, 2000). Thus, measuring the quality of relationships (negative vs. positive) in addition to the quantity of social network members is important to provide insight into potential negative social exchanges.
Social Isolation Assessment
There are several instruments that can be used to assess for social isolation in older adults, one of which is the Lubben Social Network Scale (LSNS). The LSNS has several versions including a revised version (LSNS-R), an extended 18-item version (LSNS-18), and an abbreviated 6-item version (LSNS-6). The LSNS-6 can be used as an extremely quick measure to screen for social isolation in the clinical setting. The LSNS-6 measures the following three aspects of social networks: (1) emotional, (2) tangible, and (3) actual network size. The LSNS-6 assesses these three aspects of social networks relating to the individuals family members and friends to determine if an older adult warrants further social isolation assessment (Lubben & Gironda, 2003). This instrument has been shown to be successful in the clinical setting in identifying those who are at risk for social isolation (Lubben & Gironda, 2000; Lubben et al., 2006). In psychometric testing, the LSNS-6 was found to have a Cronbach's alpha of 0.78 (Lubben & Gironda, 2003).
Review of Interventions on Social Isolation
Although interventions for social isolation are not refined yet, some suggestions can be gleaned from the literature. For example, convening groups of 7–8 members who meet regularly (weekly) is one way to increase number of friends and increase psychological well-being (Routasalo, Tilvis, Kautiainen, & Pitkala, 2009). However, obtaining new friends does not necessarily impact social isolation (Routasalo et al., 2009). A systematic review of interventions of social isolation found that educational and social activity group interventions targeting specific groups can lessen social isolation in older adults (Cattan, White, Bond, & Learmouth, 2005). Therefore, referring older adults to social activity/senior centers in their local area where these types of programs may be ongoing could be helpful. Because the living situation of older adults in long-term care is substantially different from that of those living in the community, interventions for this elderly population may need to be different.
J Prim Prev. 2012;33(2-3):137-152. © 2012 Springer
Springer Science+Business Media