Human beings are social animals and our biological, psychological, and social systems evolved to thrive in collaborative networks of people. In many societies, social networks are likely to thin as people age, leading in many cases to isolation and loneliness. In this article we review the evidence that social isolation affects health and mortality, whether or not the isolation is accompanied by subjective loneliness. Some studies suggest that the impact of isolation and loneliness on health and mortality are of the same order of magnitude as such risk factors as high blood pressure, obesity, and smoking. We also review what is known about the mechanisms underlying the effects of isolation and loneliness on health. Cardiovascular, inflammatory, hormonal, sleep-related, and emotional factors are all relevant. Finally, we look at the preliminary evidence that interventions to address social isolation and loneliness may improve health outcomes. Throughout all the research referenced in this review are cautions that it can be difficult to isolate cause and effect in these studies, since people with pre-existing health conditions may be prone to social isolation, and many chronic health conditions make socialization more challenging. We must also remember that not all who are isolated are lonely and not all who are lonely are isolated. Being in unhealthy relationships can be more stressful than being alone. Nevertheless, we conclude that effortsto address social isolation in older adults, including those relying on remote technologies, are likely to be cost effective for health care systems, and are, at the very least, humane approaches to a very common form of distress in older adults.
Health Effects of Social Isolation and Loneliness
We are a social species. Our social networks (families, tribes, communities, etc.) enabled us to survive and thrive. Our survival was served by the evolutionary development of behaviors and physiologic mechanisms (neural, hormonal, cellular, genetic) that support social interactions (Cacioppo et al., 2011). But as with all human traits, there is variation in our social behaviors and needs. The fact is, most of us are psychologically and biologically “programmed” to need social networks. It is logical that social isolation may impose stress on our minds and bodies that has a significant impact on health.
Since social isolation and loneliness are common in older adults, much attention has been paid to clarifying their adverse effects on health in old age. However, it is surprisingly difficult to study these effects and to distinguish the effects of social isolation and loneliness on health when pre-existing health conditions, such as immobility and depression, can themselves both contribute to ill health as well as increase isolation and loneliness. It is also challenging to distinguish social isolation and loneliness from one another; not all who are isolated are lonely and not all who are lonely are alone. In this article, we review what is known on this topic.
Defining social isolation and loneliness
Not all people experience “aloneness” in the same way. Social scientists who study isolation and loneliness have attempted to define these terms in specific ways, since a person is considered socially isolated if they live alone, have less than monthly contact with friends or family, and don’t belong to a group (religious congregation, club, work or volunteer organization, etc.). Of course, some choose isolation as a preferred lifestyle. Others, likely far more in number, have isolation imposed on them through the death of loved ones, family and friends moving away, remote rural housing, recent moves to an unfamiliar city, impaired mobility, and other situations leading to depleted social networks and isolation. People in these situations may be more likely to experience loneliness and to feel isolated (perceived isolation). There are validated research instruments that quantify social isolation and loneliness primarily in terms of number and frequency of social contacts. However, defining isolation in quantitative terms may not always be valid. Research, as well as our own experience, tells us that the quality of our social interactions, more than the number of our relationships, determines loneliness.
Researchers have also approached these issues using qualitative methods. Cornwell and Waite (2009) use terms such as “social disconnectedness” and “perceived isolation” to define social isolation and loneliness using the ob-jective and subjective nature of these states. Social disconnectedness is defined as lack of contact with others. Perceived isolation is defined as the subjective experience of lack of companionship and support. Loneliness may be part of that, although people can still experience subjective isolation around others. The assumption is that social disconnectedness without perceived isolation (i.e. isolation without loneliness) would be more “ego syntonic” and less stressful than states of loneliness and depression, therefore having less impact on health. Research has not always supported this assumption (Cornwell & Waite, 2009). Social isolation, with or without loneliness, can have as large effect on mortality risk as smoking, obesity, sedentary lifestyle and high blood pressure (Cacioppo et al. 2011).
Associations of isolation and loneliness on health
Several indicators of social isolation have been associated with poor health. There is a vast literature on this topic that is beyond the scope of this article, but several studies can help us better understand the relationships of social networks, perceived isolation, health, and mortality. From a methodological perspective, these studies assume that health status contributes to one’s ability to be socially engaged. Therefore, health status can contribute to loneliness and isolation, thereby creating a “cause and effect” dilemma when attempting to define the relationships between loneliness, social isolation, health, and mortality. Investigators must control for baseline health status in the design of their studies and in the analysis of their data. Despite this, the effects of social isolation and loneliness on health are a strong enough force that they consistently emerge as unambiguous risk factors for ill health and mortality in the many studies that have examined these relationships through various methodologies, including longitudinal cohort studies and meta-analyses (quantitative analysis of the combined results of carefully selected studies).
An older, but large and well-designed prospective study over four years looked at total mortality in a group of men for whom social networks were known. Some 32,624 healthy men were followed and 511 deaths occurred. Socially isolated men (not married, fewer than six friends or relatives, no memberships in religious or social organizations) had a 90% increased risk of cardiovascular death and more than double the risk of death from an accident or suicide. They also had double the risk of non-fatal stroke. They had no increased risk from non-fatal MI in this study, raising the question of whether or not social isolation contributes to either the severity or survivability of cardiac events (Kawachi et al., 1996). These investigators did not look at loneliness versus social isolation as relative risk factors.
It is natural to assume that loneliness has a greater effect on health and some studies support that conclusion. Adverse effects on health from loneliness are seen at every stage of the lifecycle (Hawkley & Capitanio, 2014). But the elderly are at particular risk both for loneliness and the health consequences of loneliness. For example, in a questionnaire study involving a large number of older adults in Finland, 39% suffered loneliness at least some of the time; 5% often or always. Loneliness was statistically associated with several demographic variables, including rural living, older age, living alone or in residential care, widowhood, low level of education, and low income. Subjectively, the people in this study attributed their loneliness to illness, loss of spouse, and lack of friends. Poor health status and poor functional status were also associated with greater feelings of loneliness (Savikko et al., 2005). A study done by Cacioppo and Caciappo (2014) found loneliness to be associated with ill health to a greater degree than just social isolation. They examined two elements of social isolation independently (social disconnectedness and perceived isolation) on both physical and mental health. Stronger relationships were shown between loneliness and worse health, including cardiovascular disease, inflammation, and depression, than social isolation itself. Loneliness in older adults was shown to significantly increase risk of functional decline and death in a recent longitudinal cohort study of 1604 followed over six years. Some 43% of the cohort reported loneliness and they were at higher risk for both functional decline (ADLs, mobility) and death. The authors of this study found that loneliness was associated with these poor outcomes even after adjusting for baseline health status and depression, but did not compare those who were isolated to those who were lonely (Perissinotto et al., 2012).
On the other hand, many investigators have found social isolation itself to be a risk factor for ill health. In a meta-analysis of studies examining the magnitude of effect of social isolation and loneliness on mortality in which important baseline health variables were controlled in the analysis, Holt-Lunstad and colleagues (2015) found a 29% increased risk of mortality over time from social isolation and 26% increase in mortality risk from loneliness. Interestingly, they found a 32% increased risk from just living alone, independent of social isolation. That is, they found no correlation of objective versus subjective social isolation. This finding is counter-intuitive, in that we would think that the stress of loneliness would be a driving factor for ill health, yet “aloneness” seems to be at least as strong, if not a stronger influence on health. Steptoe et al. (2013) investigated whether the health impact of social isolation was “caused by loneliness” in 6500 men and women more than 52 years of age participating in the English Longitudinal Study of Aging. They quantified contact with family, friends, and community organizations and administered a loneliness questionnaire. They monitored mortality for an average of 7.25 years per subject. After adjusting for demographic variables, social isolation increased mortality whereas loneliness did not. Those with the highest social isolation (least social contact) had an even higher risk. It is very important to note that although there was an increased mortality risk in lonely people, they also had higher baseline mental and physical health problems that may have accounted for the increased risk over the period of observation. That is, loneliness in this study was association with high baseline levels of depression, arthritis, and mobility impairment than the social isolation without loneliness cohort. So, when baseline health variables were factored out, the loneliness cohort did not seem to have as high a mortality rate. In reality, both social isolation and loneliness are associated with increased mortality rates (Steptoe A et al. 2013).
Whether or not the impact of social isolation and perceived isolation (i.e. loneliness) on health are comparable remains unclear, but the evidence seems to be leaning towards the conclusion that both pose risks to health. In an effort to clarify the relative effect of loneliness and social isolation on cardiovascular mortality risk, Valtorta et al. (2016) conducted a meta-analysis of 11 cardiac and eight stroke studies. Poor social relationships in general (social isolation and loneliness) were associated with a 29% increase in risk of coronary heart disease and 32% increase in stroke risk. This increased risk is comparable to the risk of obesity and lack of physical activity and whether isolated people were lonely or not did not appear to make a difference.
Many potential mechanisms have been proposed to account for the relationships between social integration, perceived social support, and health outcomes. First of all, spending time with people who exhibit healthy habits may reinforce healthy behaviors, improve access to health-related information, better nutrition, more physical activity, transportation to health care providers, and even increase financial resources. Of course, peer relationships can easily lead to unhealthful behaviors or interpersonal stress as well, but in the literature pertaining to older adults, the health-promoting benefits of social relationships seems to outweigh the negative effects. (Cornwell & Waite, 2009) But changing health behaviors is likely not the only mechanism by which social contacts protect health and well-being.
Loneliness is known to be a major risk factor for depression, which itself accelerates functional decline and increases mortality rate. (Mehta et al., 2002) Even sub-clinical depression may increase risk of all-cause mortality. (Culjpers & Smit, 2002), so depression may have contributed to the increased mortality and cardiovascular diseases found in the loneliness cohorts of those studies cited previously. Depression may increase mortality and illness through several mechanisms. Depression can increase platelet aggregation through diminished serotonin function and thereby increase risk for myocardial infarction and stroke. There may also be increased heart rate variability (unstable autonomic nervous system) and increased release of adrenaline, both leading to increased risk of cardiac arrhythmia (Seymour & Benning, 2009). Whatever the mechanism, the effect of depression on mortality is significant in size. In a large cohort study (Cardiovascular Health Study), investigators found that depression increased mortality risk by 24% when they accounted for all important co-variables (Schultz et al., 2000).
Social isolation can have direct effects on cardiovascular disease risk factors. Perceived isolation and loneliness are associated with increased sympathetic nervous system activity, increased inflammation, and decreased sleep, all of which can accelerate brain and cardiovascular aging (Cacioppo, et al., 2011). Loneliness increases risk for dementia, likely through these mechanisms, however the absence of social interaction itself may also be a primary factor in that social stimulation can help maintain brain health (Cacioppo & Hawkley, 2009; Cacioppo et al., 2014). Grant and colleagues examined key metabolic risk factors for cardiovascular mortality, looking at blood pressure, lipids, and cortisol responses to stress. Using a measure of social integration (Close Persons Questionnaire), they found dysregulated blood pressure and cortisol responses to acute stress in people (238 middle-aged men and women) with few close friends. They also saw increased cholesterol in the socially isolated men, but not women. These physiologic changes increase risk of heart attacks and stroke. The authors note that these changes in cardiovascular risk factors in isolated individuals were independent of whether they expressed feelings of loneliness (Grant et al., 2009).
Finally, there is some evidence that loneliness can affect immune function, increasing susceptibility to infec-tion (Cohen S et al., 1997). Loneliness is also associated with disrupted sleep. Insomnia affects immune function, glucose regulation, cardiovascular risk, dementia risk, mood, and daytime function (Hawkley et al., 2010).
We do not yet know whether efforts to reduce isolation and loneliness can actually improve health. Despite this, Valtorta et al. (2015) note that the evidence linking social isolation in old age with poor health is strong enough that efforts to reduce cardiovascular disease need to consider social interventions aimed at reducing isolation (Valtorta NK et al., 2015). While this claim may be premature, there are studies that do suggest increasing social networks can improve health. In one such study, conducted over a 10-year period of follow-up, men (aged 42-77) with lower levels of “social integration” (by a standard social network index) were, as expected, found to be at greater risk of total mortality than those with more social connections. What was surprising in this study was that in a sub-analysis of the older men of the sample who showed increasing social network size over the 10-years of study, an increased number of close friends or increased attendance at religious services were both associated with a reduced risk of death. The effect size was robust. Those reporting having more friends over time, showed a reduction of 29% in mortality risk per year (Eng et al., 2002). This doesn’t prove causality; perhaps improvements in health for other reasons promoted behaviors that lead to more friends. Nevertheless, the finding is encouraging.
Although the stress of being a caregiver to a disabled family member is not the same kind of stress as social isolation, caregivers consistently describe the isolation of the caregiver’s role as one of the most stressful aspects of the caregiving role. Caregivers consistently report higher levels of stress than non-caregivers and chronic stress is associated with poorer health outcomes and higher rates of mortality. But caregivers overall have a lower mortality rate. The important factor is stress. Not all caregivers experience significant stress, and those that don’t may experience health benefits from the caregiving relationship. In fact, in one study, non-stressed caregivers had 43% lower rates of mortality relative to non-caregivers. In previous studies, caregivers experiencing significant emotional stress showed a 60% increase in mortality rate (Fredman L et al., 2010). These findings are relevant to considerations of interventions for social isolation. Non-stressed caregivers are more likely to experience positive emotions from the person they are providing care for and to gain strength from having a vital role to play in another person’s life. To be a caregiver and not feel some reciprocal caring from your partner is a special form of isolation that is particularly demoralizing, stressful, and unhealthy. Even small efforts to make isolated people feel appreciated and useful may reduce the stress of loneliness and thereby improve health.
Innovative ways to help depressed, isolated people may also have positive effects on health. In a twelve-month multi-modality, home-based intervention, randomized controlled trial for older adults with depression, those receiving a home-based (as opposed to usual, office-based) treatment had significantly better responses. The home-based treatment group were more likely to be in remission from depression, had greater quality of life improvements, and greater gains in functional well-being and emotional well-being (Ciechanowski et al., 2004).
Given the mobile nature of our society, social relationships frequently are maintained at a distance through telephone contact, email, and social media when physical contact is not practical. Interventions relying on technology to reduce isolation may be better than no intervention at all, but they are not the same as in-person visits. A large cohort study has recently revealed that different methods of contact are not equal in reducing feelings of loneliness and depression. These investigators found a higher risk of depression in those with less than once-a-month face to face contact with children, family, or friends. People with once or twice-a-week contact had the lowest rates of depression. However, older age, interpersonal conflict, and depression at baseline decreased the effect of physical contact. That is, if a person is prone to depression, is physically frail, or the relationship causes tension, a phone call may be as good (or better) than in-person contact (Teo et al., 2015).
There is an increasing amount of evidence that pets, especially dogs and cats, are associated with health ben-efits and reduced mortality. Research into whether animal companions can offset the deleterious effects of social isolation on health is needed.
Implications for Aging Life Care/care management:
Aging Life Care / care managers may be in a better position than any other member of the health care team both to recognize social isolation and to organize interventions. Based on current evidence, they can justify in-creased focus on social relationships in the multidisciplinary health care treatment plan and in their individual efforts to reduce isolation in their clients. An understanding that social isolation is a significant risk factor to health, of similar magnitude to obesity and diabetes, may be persuasive for some of their clients who are able to increase social contact with others, either in person or through social technologies.
We have reviewed studies examining the complex relationships of health, mortality and social isolation in old age. There is strong evidence that many older adults feel isolated, and that loneliness is associated with poor health and higher rates of mortality. There is also evidence that social isolation even without subjective loneliness increases risk. The effect of social isolation on health appears to be of a similar magnitude to other risks to health, such as high blood pressure, smoking and obesity. Whereas these health risk factors have stimulated major public health interventions in recent decades, efforts to reduce isolation and loneliness have not been made on a level of population-health. Some authors, however, warn that such large-scale efforts based on health risk may be premature. They say that increased risks to health from isolation and loneliness are actually “modest” in magnitude and that the strong associations found in many studies are due to failure to control for baseline health status (Corman et al., 2003). We also have to keep in mind that being in toxic relationships may be even more stressful and unhealthy than loneliness. Nevertheless, there is enough evidence to consider social isolation and loneliness among older adults a significant public health issue. There are also compelling hypotheses and some experimental data to explain the physiologic mechanisms by which social isolation drives disease. And perhaps more importantly, we are starting to see evidence that interventions to reduce loneliness may provide health benefits. I have not offered simple prescriptions to address isolation and loneliness. That is not the purpose of this review, which is meant to offer evidence that population-health authorities should take this issue as seriously as other known health risk factors. While we don’t have definitive evidence at the present time, it is very likely that social interventions provided at relatively modest costs will have very significant cost savings in public health. Much more research is needed for intervention trials, including those employing social media and telephone contacts. At the very least, such efforts provide a safe, humane approach to a common cause of suffering in older adults.