Decades of research worldwide have yielded reliable insights into factors associated with suicide in later life. Encompassing constructs in multiple domains, they can be usefully framed as the “5Ds” (Van Orden etal., Reference Van Orden, Conwell and Silva2019). The first is access to deadly means. Although the choice of methods used by older adults to take their own lives is influenced by cultural and environmental factors, evidence shows that lethality of intent and implementation are higher among older adults than those who die by suicide at younger ages (Conwell etal., Reference Conwell1998). In the U.S., 75% of older people who die by suicide used a firearm (Centers for Disease Control and Prevention). The remaining four Ds are depression, disease, disability, and disconnectedness. While other mental disorders are also present in older people who die by suicide, major and minor affective disorders have been found through psychological autopsy studies in up to 87% (Van Orden etal., Reference Van Orden, Conwell and Silva2019). Disease and disability refer to medical illnesses and their functional impact that, although so common as to be the rule in later life, are known to elevate risk approximately twofold. Finally, social disconnectedness is the lack of structural, functional, and emotional supports that people want and need from each other. Whether the deficit is expressed as the experience of loneliness or objectively measured paucity of social supports and contacts, disconnectedness too has been closely tied to a wide range of adverse health outcomes from compromise of immune function to complications of diabetes and cardiovascular health, cognitive decline, depression and suicide (Conwell etal., Reference Conwell2020; Holt-Lunstad etal., Reference Holt-Lunstad, Robles and Sbarra2017).
The 5Ds framework can be useful as an aide to clinicians in appreciating the multi-determined nature of suicide in later life. As well, its visual depiction as five intersecting domains underscores the dynamic interactions between determinants of risk that more closely reflect the complex assessments that skilled clinicians must perform. An older man with chronic pain (disease) who is unable to work (disability) may adapt with the support of friends and family. If he becomes separated from his support network (disconnected), however, and becomes depressed, his risk increases markedly beyond that due to any one or two factors alone. The presence of a firearm in the home under those conditions may be lethal.
In this issue of the journal, Wongpakaran and colleagues provide additional empirical support for the need to understand late-life suicide through examination of interactions between factors known to place older people at risk (Wongpakaran etal., Reference Wongpakaran, Oonarom, Wongpakaran and Kuntawong2020). Importantly, their work also suggests the potential need to add another dimension to the 5D framework – the role of lifespan development. In research on risk and protective factors for suicide in late life, longitudinal and lifespan developmental approaches are relatively rare, while much research focuses on cross-sectional measures and age-acquired risk factors rather than the impact of stability or fluctuations in risk across the lifespan. However, Wongpakaran and colleagues’ results support the notion that experiences earlier in life, in this case those that determine attachment style, may reverberate at the opposite end of the life course and contribute to making people more vulnerable to the “Ds” they encounter in their later years.
The authors conducted cross-sectional analyses of data collected from 191 older adults who had participated in studies of mood and anxiety disorders in mental health settings of 10 tertiary care facilities in Thailand. They assessed depressive symptoms, suicidal ideation in the preceding month, subjects’ perceived social support, and their level of attachment anxiety. In a series of mediation and moderation analyses, they found that attachment anxiety was significantly associated with suicidal ideation, but that the relationship was fully mediated by depression. Further, they observed a significantly stronger association between depression and suicidal ideation among those who were more socially disconnected (lower perceived social support). In other words, attachment anxiety, which is likely to be a trait reflecting the nature of one’s early-life attachment relationships, predisposes to depression and thereby suicide risk, and that that risk may be buffered (the impact of depression on suicidal ideation may be reduced) by intact social connectedness.
Suicide in later life is clearly a multi-determined behavior the complexity of which poses challenges for prevention (Conwell, Reference Conwell2014). At the same time, however, it suggests opportunities to intervene at numerous points in the causal web. This study, for example, reinforces the central role of depression as a driver of suicidal ideation in older adults, and therefore an important target for indicated preventive interventions to reduce suicide-related morbidity and mortality. Diagnosis and treatment of depression in later life, especially in primary care settings where depressed older people so commonly present, is both feasible and effective, as shown by numerous controlled trials of integrated depression care management interventions (Archer etal., Reference Archer2012). It also suggests that by detecting older people who are lonely or with fewer social supports and augmenting their social connectedness, we may be able to reduce the risk that the depressed older person will become suicidal and act on their wish to die. The addition of developmental factors such as attachment style to the framework has further implications, however.
Although replete with challenges, aging has benefits as well. Carstensen and colleagues, for example, have shown that as we age, we become more selective about our relationships, pruning our social networks to focus more on those that are more rewarding and serve to maximize emotional well-being (Carstensen etal., Reference Carstensen, Isaacowitz and Charles1999). Older adults then tend to be more satisfied with their lives than those in mid-life. As reviewed by Van Assche and colleagues (Van Assche etal., Reference Van Assche, Luyten, Bruffaerts, Persoons, Van De Ven and Vandenbulcke2013), attachment theory offers potential insights into how older adults develop and maintain such resilience in the face of aging-related challenges. Attachment style is established early in life based on experience of safe and trusted relationships with others. It persists through adulthood, however, expressed in modified forms. The relatively small body of research on attachment in later life indicates that older adults may have a smaller number of attachments that nonetheless, consistent with Carstensen’s Socioemotional Selectivity Theory (Carstensen etal., Reference Carstensen, Isaacowitz and Charles1999), meet their needs under most circumstances. As people transition to old age, there is an associated decrease in attachment anxiety, whereas persisting attachment anxiety has been related to low levels of well-being and high levels of distress. Social support on the other hand may moderate these relationships.
Related research has also focused on the impact of adverse childhood experiences (ACEs, e.g. child abuse and neglect) on a wide range of health outcomes showing, for example, that the association between ACEs and prevalence of symptomatic depression, anxiety, and personality disorders remained significant across the life course (Raposo etal., Reference Raposo, Mackenzie, Henriksen and Afifi2014). A relationship between ACEs and suicidal ideation and behavior is increasingly well established as well (Fuller-Thomson etal., Reference Fuller-Thomson, Baird, Dhrodia and Brennenstuhl2016). Additional research in older adult samples is needed to understand the relationships between ACEs and attachment disorders, their full impact on late-life mental health, and the mechanisms by which they have that effect.
As an addition to the 5D framework, lifespan developmental processes are likely to impact and interact with the other “Ds.” As discussed previously, earlier life experiences with attachment may affect an older adult’s structural, functional, and/or emotional social connectedness. In addition, disability and disease in late life may be highly influenced by lifelong patterns of health and behaviors (Zhang etal., Reference Zhang, Feldman and Du2019), while an individual’s adjustment to disease or disablement in late life may partially depend upon the manner in which these processes occurred (e.g. gradual vs. sudden onset). Depression and other mental health problems may impact older adults differentially depending on their developmental course (e.g. onset in earlier life vs. late-life onset) (Fiske etal., Reference Fiske, Wetherell and Gatz2009). Even access to deadly means may be somewhat influenced by developmental processes such as lifelong exposure to certain means (e.g. firearms). Thus, the study of suicide in late life is necessarily dependent upon an understanding of lifespan development, where risk in late life may be seen as a culmination not just of risk in the current life stage, but of lifelong processes.
Suicide rates rise with age across the life course for both men and women in most countries that report mortality statistics to the World Health Organization (Van Orden etal., Reference Van Orden, Conwell and Silva2019). Furthermore, older adults are the most rapidly growing segment of the population worldwide. It is likely, therefore, that the global burden of disease associated with suicidal behavior in later life will rise dramatically in coming decades. As efforts continue to understand the factors that place older people at risk for suicide, the evidence provided by Wongpakaran etal. (Reference Wongpakaran, Oonarom, Wongpakaran and Kuntawong2020) suggests the need to complement research on age-acquired factors with studies informed by lifespan developmental theory. Examining the influence of early-life experiences, both adverse events and the health promoting effects of interpersonal relationships and supports, as well as fluctuations and development in risk across the lifespan, will enhance our understanding of risk for suicide a lifetime later. That understanding may in turn lead to new approaches to upstream preventive interventions that increase resilience of adults to the age-related stressors they encounter in older adulthood, impacting not only risk as it presents at that time (e.g. late-life social disconnectedness) but also the developmental processes that may drive that risk (e.g. attachment disorders).
You have
Access
Decades of research worldwide have yielded reliable insights into factors associated with suicide in later life. Encompassing constructs in multiple domains, they can be usefully framed as the “5Ds” (Van Orden etal., Reference Van Orden, Conwell and Silva2019). The first is access to deadly means. Although the choice of methods used by older adults to take their own lives is influenced by cultural and environmental factors, evidence shows that lethality of intent and implementation are higher among older adults than those who die by suicide at younger ages (Conwell etal., Reference Conwell1998). In the U.S., 75% of older people who die by suicide used a firearm (Centers for Disease Control and Prevention). The remaining four Ds are depression, disease, disability, and disconnectedness. While other mental disorders are also present in older people who die by suicide, major and minor affective disorders have been found through psychological autopsy studies in up to 87% (Van Orden etal., Reference Van Orden, Conwell and Silva2019). Disease and disability refer to medical illnesses and their functional impact that, although so common as to be the rule in later life, are known to elevate risk approximately twofold. Finally, social disconnectedness is the lack of structural, functional, and emotional supports that people want and need from each other. Whether the deficit is expressed as the experience of loneliness or objectively measured paucity of social supports and contacts, disconnectedness too has been closely tied to a wide range of adverse health outcomes from compromise of immune function to complications of diabetes and cardiovascular health, cognitive decline, depression and suicide (Conwell etal., Reference Conwell2020; Holt-Lunstad etal., Reference Holt-Lunstad, Robles and Sbarra2017).
The 5Ds framework can be useful as an aide to clinicians in appreciating the multi-determined nature of suicide in later life. As well, its visual depiction as five intersecting domains underscores the dynamic interactions between determinants of risk that more closely reflect the complex assessments that skilled clinicians must perform. An older man with chronic pain (disease) who is unable to work (disability) may adapt with the support of friends and family. If he becomes separated from his support network (disconnected), however, and becomes depressed, his risk increases markedly beyond that due to any one or two factors alone. The presence of a firearm in the home under those conditions may be lethal.
In this issue of the journal, Wongpakaran and colleagues provide additional empirical support for the need to understand late-life suicide through examination of interactions between factors known to place older people at risk (Wongpakaran etal., Reference Wongpakaran, Oonarom, Wongpakaran and Kuntawong2020). Importantly, their work also suggests the potential need to add another dimension to the 5D framework – the role of lifespan development. In research on risk and protective factors for suicide in late life, longitudinal and lifespan developmental approaches are relatively rare, while much research focuses on cross-sectional measures and age-acquired risk factors rather than the impact of stability or fluctuations in risk across the lifespan. However, Wongpakaran and colleagues’ results support the notion that experiences earlier in life, in this case those that determine attachment style, may reverberate at the opposite end of the life course and contribute to making people more vulnerable to the “Ds” they encounter in their later years.
The authors conducted cross-sectional analyses of data collected from 191 older adults who had participated in studies of mood and anxiety disorders in mental health settings of 10 tertiary care facilities in Thailand. They assessed depressive symptoms, suicidal ideation in the preceding month, subjects’ perceived social support, and their level of attachment anxiety. In a series of mediation and moderation analyses, they found that attachment anxiety was significantly associated with suicidal ideation, but that the relationship was fully mediated by depression. Further, they observed a significantly stronger association between depression and suicidal ideation among those who were more socially disconnected (lower perceived social support). In other words, attachment anxiety, which is likely to be a trait reflecting the nature of one’s early-life attachment relationships, predisposes to depression and thereby suicide risk, and that that risk may be buffered (the impact of depression on suicidal ideation may be reduced) by intact social connectedness.
Suicide in later life is clearly a multi-determined behavior the complexity of which poses challenges for prevention (Conwell, Reference Conwell2014). At the same time, however, it suggests opportunities to intervene at numerous points in the causal web. This study, for example, reinforces the central role of depression as a driver of suicidal ideation in older adults, and therefore an important target for indicated preventive interventions to reduce suicide-related morbidity and mortality. Diagnosis and treatment of depression in later life, especially in primary care settings where depressed older people so commonly present, is both feasible and effective, as shown by numerous controlled trials of integrated depression care management interventions (Archer etal., Reference Archer2012). It also suggests that by detecting older people who are lonely or with fewer social supports and augmenting their social connectedness, we may be able to reduce the risk that the depressed older person will become suicidal and act on their wish to die. The addition of developmental factors such as attachment style to the framework has further implications, however.
Although replete with challenges, aging has benefits as well. Carstensen and colleagues, for example, have shown that as we age, we become more selective about our relationships, pruning our social networks to focus more on those that are more rewarding and serve to maximize emotional well-being (Carstensen etal., Reference Carstensen, Isaacowitz and Charles1999). Older adults then tend to be more satisfied with their lives than those in mid-life. As reviewed by Van Assche and colleagues (Van Assche etal., Reference Van Assche, Luyten, Bruffaerts, Persoons, Van De Ven and Vandenbulcke2013), attachment theory offers potential insights into how older adults develop and maintain such resilience in the face of aging-related challenges. Attachment style is established early in life based on experience of safe and trusted relationships with others. It persists through adulthood, however, expressed in modified forms. The relatively small body of research on attachment in later life indicates that older adults may have a smaller number of attachments that nonetheless, consistent with Carstensen’s Socioemotional Selectivity Theory (Carstensen etal., Reference Carstensen, Isaacowitz and Charles1999), meet their needs under most circumstances. As people transition to old age, there is an associated decrease in attachment anxiety, whereas persisting attachment anxiety has been related to low levels of well-being and high levels of distress. Social support on the other hand may moderate these relationships.
Related research has also focused on the impact of adverse childhood experiences (ACEs, e.g. child abuse and neglect) on a wide range of health outcomes showing, for example, that the association between ACEs and prevalence of symptomatic depression, anxiety, and personality disorders remained significant across the life course (Raposo etal., Reference Raposo, Mackenzie, Henriksen and Afifi2014). A relationship between ACEs and suicidal ideation and behavior is increasingly well established as well (Fuller-Thomson etal., Reference Fuller-Thomson, Baird, Dhrodia and Brennenstuhl2016). Additional research in older adult samples is needed to understand the relationships between ACEs and attachment disorders, their full impact on late-life mental health, and the mechanisms by which they have that effect.
As an addition to the 5D framework, lifespan developmental processes are likely to impact and interact with the other “Ds.” As discussed previously, earlier life experiences with attachment may affect an older adult’s structural, functional, and/or emotional social connectedness. In addition, disability and disease in late life may be highly influenced by lifelong patterns of health and behaviors (Zhang etal., Reference Zhang, Feldman and Du2019), while an individual’s adjustment to disease or disablement in late life may partially depend upon the manner in which these processes occurred (e.g. gradual vs. sudden onset). Depression and other mental health problems may impact older adults differentially depending on their developmental course (e.g. onset in earlier life vs. late-life onset) (Fiske etal., Reference Fiske, Wetherell and Gatz2009). Even access to deadly means may be somewhat influenced by developmental processes such as lifelong exposure to certain means (e.g. firearms). Thus, the study of suicide in late life is necessarily dependent upon an understanding of lifespan development, where risk in late life may be seen as a culmination not just of risk in the current life stage, but of lifelong processes.
Suicide rates rise with age across the life course for both men and women in most countries that report mortality statistics to the World Health Organization (Van Orden etal., Reference Van Orden, Conwell and Silva2019). Furthermore, older adults are the most rapidly growing segment of the population worldwide. It is likely, therefore, that the global burden of disease associated with suicidal behavior in later life will rise dramatically in coming decades. As efforts continue to understand the factors that place older people at risk for suicide, the evidence provided by Wongpakaran etal. (Reference Wongpakaran, Oonarom, Wongpakaran and Kuntawong2020) suggests the need to complement research on age-acquired factors with studies informed by lifespan developmental theory. Examining the influence of early-life experiences, both adverse events and the health promoting effects of interpersonal relationships and supports, as well as fluctuations and development in risk across the lifespan, will enhance our understanding of risk for suicide a lifetime later. That understanding may in turn lead to new approaches to upstream preventive interventions that increase resilience of adults to the age-related stressors they encounter in older adulthood, impacting not only risk as it presents at that time (e.g. late-life social disconnectedness) but also the developmental processes that may drive that risk (e.g. attachment disorders).
Conflict of interest
The authors have no conflicts of interest to report.