Awareness of Age-Related Change as a Behavioral Determinant of Survival Time in Very Old Age

 Turns out, that may not be just flowery talk—in fact, a July 2022 article in the peer-reviewed journal Frontiers in Psychology cited 2021 research that suggested when we perceive  as "gain-related" (in other words, believing that life gets better with age), that perception alone "was predictive of longer survival" when it was compared to average all-cause mortality rates among people who saw aging as "loss-related," or negative.(Life is definitely better as I age, I got divorced enhancing my happiness immeasurably. I'm retired and comfortably well off. And healthy as I can be post stroke. I'm going to live a long time yet.)

People really need to start listening to and repeating the Ani DeFranco song. I'm happy.
"If Yr Not"
If you’re not, if you’re not
If you’re not, if you’re not
If you’re not getting happier as you get older
Then you’re fuckin’ up

Awareness of Age-Related Change as a Behavioral Determinant of Survival Time in Very Old Age

Roman Kaspar^1* Hans-Werner Wahl² Manfred Diehl³

• ¹Cologne Center for Ethics, Rights, Economics, and Social Sciences of Health, University of Cologne, Cologne, Germany
• ²Network Aging Research of Heidelberg University, Heidelberg, Germany
• ³Department of Human Development and Family Studies, Colorado State University, Fort Collins, CO, United States

Although research on the association between subjective views of aging (VOA) and survival is scarce, more negative VOA have been found to be associated with increased all-cause mortality, even after controlling for possible confounders. Longitudinal studies on the predictive association of VOA with survival in individuals aged 80 years or older are, however, very limited. Thus, the aim of this study was to link adults’ awareness of age-related change (AARC), a multidimensional measure of adults’ subjective VOA, to survival time across a 3.5-year observation interval in advanced old age. To put the AARC construct in context, the study also considered related psychosocial concepts (i.e., perceived control and appraisal of life) essential for coping with late-life challenges as potential behavioral predictors of longevity. Data came from a representative panel study that included persons living in community and institutional settings. A total of 1,863 interviews were conducted at wave 1. This study used meta-data from wave 2 fieldwork 2 years after the initial assessment and death records obtained during panel maintenance after 3.5 years to estimate determinants of survival. Results showed that loss-related VOA indicated increased risk to survival, whereas gain-related VOA were predictive of longer survival. Both perceived age-related losses and perceived age-related gains exerted a significant independent effect on late-life mortality over and above socio-demographic background characteristics, perceived control, engagement with life, as well as health status. These findings suggest that the multidimensional examination of very old adults’ VOA may help to better understand successful longevity in the Fourth Age.

Introduction

Advanced old age is frequently seen as the most vulnerable period of the human lifespan because many adults experience multimorbidity, functional disability, motor and sensory impairment, significant cognitive decline, and frailty. Psychological conceptualizations of what has also been named the “Fourth Age” to a large extent echo this biomedical loss perspective of very late life, resulting in what Baltes and Smith (2003) have described as the “dilemmas of the fourth age” to be contrasted with the successes of the Third Age (Baltes and Smith, 2003; Wahl and Ehni, 2020). Still, psychological resources to counteract and cope with increasing health-related and functional losses seem not to be completely exhausted in the Fourth Age. For example, the phenomenon of the subjective age bias (i.e., individuals feeling significantly younger than their chronological age) persists into advanced old age and the gap between felt age and chronological age, in fact, increases considerably in very old age (Pinquart and Wahl, 2021). In addition, subjective health evaluations which are of utmost importance for quality of life at large depict relatively little or even no decline, although objective health tends to decline in pronounced ways (Wettstein et al., 2016). Both findings have been shown to be quite robust and suggest that very old adults manage to distance themselves, at least to some extent, quite effectively from the down sides of the Fourth Age.

This article addresses whether psychological resources also account for significant variance in predicting survival in advanced old age. An important background for addressing this question is that the predictive strength of social factors, such as income, education, and marital status, for survival is much weaker in those 80+ than in younger age groups (Goldman et al., 1995; Dupre, 2007). In addition, a recent analysis based on large samples from the National Health and Nutrition Examination Survey (NHANES) data infrastructure revealed that although instrumental activities of daily living (IADL) and self-rated health were the most important predictors of survival in those in the Third (65–79 years) and the Fourth Age (80 years and older), their incremental predictive value decreased considerably from the Third to the Fourth Age group (Goldman et al., 2017). Finally, there is a general scarcity of studies identifying psychological predictors of survival in very old age. That is, data are missing whether the “age-as leveler” hypothesis (Goldman et al., 2017) not only applies to social and health-related factors, but also to psychological resources/risk factors. In other words, is the high likelihood of biological frailty in advanced old age canceling out the effects of psychological factors that are important predictors of survival in earlier stages of old age?

Against this background, the primary aim of this study was to link awareness of age-related change (AARC), an established concept to assess older adults’ subjective views of aging (VOA) in a multidimensional way (Diehl et al., 2021), to survival time across a 3.5-year interval in a large sample of individuals aged 80 years and older. To put the AARC concept in perspective, the study also considered related psychosocial resources essential for coping with the challenges of the Fourth Age, such as retaining a positive appraisal of life (e.g., purpose in life and optimism) and a sense of control over one’s life (e.g., mastery and perceived restrictions), as potential behavioral determinants of longevity. These psychosocial resources complemented established survival predictors (e.g., age, gender, ADL, and subjective health) to provide an overall more comprehensive picture of predicting survival in very late life.

Considering in our study the connection between subjective VOA and late-life survival is based on the observation that individuals reflect on their own development and try to understand their own aging as they move across the adult lifespan (Brandtstädter and Rothermund, 2002; Diehl et al., 2015; Kornadt et al., 2019). Thus, aside from using chronological age as a marker of their position in the life course (Settersten and Hagestad, 2015), individuals also draw on their perceptions, and behavioral experiences (Miche et al., 2014b) to establish VOA as part of their identity (Diehl et al., 2021).

A vast body of research has documented that more negative VOA are associated with a range of unfavorable developmental outcomes, such as poorer physical and mental health, and poorer cognitive functioning, including cognitive pathology (for review, see Westerhof et al., 2014; Diehl et al., 2015, 2021; Wurm et al., 2017). A central aspect for this article is that more negative VOA have also been found to be associated with increased all-cause mortality, even after controlling for confounding variables (Kotter-Grühn et al., 2009; Westerhof et al., 2014; Stephan et al., 2018). Yet, to the best of our knowledge, only two earlier publications have addressed the association between VOA in very old age and survival and both were based on data from the same study (i.e., Berlin Aging Study). Maier and Smith (1999) found in a sample of individuals 70 years and older at baseline that lowered dissatisfaction with one’s own aging remained a significant predictor of survival time based on survival status data gathered 3 to 6 years after baseline assessment. Importantly, no effect of chronological age was observed; that is, dissatisfaction with aging maintained its role as a significant predictor also in those 85 years and older. Kotter-Grühn et al. (2009) showed that dissatisfaction with one’s own age was more strongly related to time-to-death, whereas, the degree of feeling younger than one’s own age was more closely associated with chronological age (i.e., distance from birth). In summary, longitudinal findings on VOA in very old age are very limited. Furthermore, given the importance of multidimensional assessment of VOA (Diehl et al., 2021) reflecting losses but also gains, it seems critical to examine multiple, distinct dimensions of VOA. Finally, to better understand the relevance of VOA for survival in advanced old age, it is also important to consider them in combination with other essential indicators of psychosocial functioning with significance for survival.

Views of aging have been assessed in multiple ways in the literature, including felt age, attitudes toward own aging, or aging satisfaction (for an overview, see Diehl et al., 2014; Klusmann et al., 2020). This study relied on a relatively recent and multi-dimensional conceptualization and measurement of VOA. Specifically, Diehl and Wahl’s (2010) construct of awareness of age-related change (AARC) was used in the present study. Diehl and Wahl (2010) defined AARC as “all those experiences that make a person aware that his or her behavior, level of performance, or ways of experiencing his or her life have changed as a consequence of having grown older (i.e., increased chronological age)” (p. 340). We assumed that such an experience-based reflection on one’s own aging should be particularly pronounced in very old age, as this is a period of life when severe age-related losses may become normative for many individuals (Baltes and Smith, 2003; Wahl and Ehni, 2020). We also expected that the two major dimensions of AARC (i.e., perceived age-related gains and perceived age-related losses) would be a particularly well-suited construct in the VOA domain to predict survival in very old adults.

First, Diehl and Wahl’s (2010) conceptualization of AARC and its measurement relies on actual perceptions of changed behavior, changed performance, or changed personal experiences. Thus, in contrast to other measurement approaches (e.g., the attitudes toward own aging approach, ATOA; see Diehl et al., 2014; Miche et al., 2014a) the AARC questionnaire avoids general ratings of a person’s perceptions of aging and asks for ratings regarding specific behaviors and experiences in critical life domains. Given the many day-to-day changes coming with very old age and accumulating challenges in gait, sensory, motor functioning, and out-of-home behavior (Baltes and Smith, 2003), the AARC questionnaire seemed particularly suited to assess VOA in very old adults. At the same time, given its focus on actual perceptions of day-to-day behaviors, AARC may be particularly sensitive to capture self-perceived changes in functions and performances that may be indicative of serious declines and may signal impending death.

Second, the AARC conceptualization was from its inception designed as a multidimensional approach, differentiating between both positive (AARC-Gains) and negative (AARC-Losses) perceptions and interpretations of events, behaviors and sensations across various life domains (e.g., health, social relationships, leisure, and lifestyle). Previous research in younger age groups has shown that perceived age-related gains and losses co-exist even within behavioral domains and have different antecedents and different associations with developmental outcomes, including depression, psychological well-being, and self-rated health (Miche et al., 2014b; Brothers et al., 2016, 2019, 2021; Dutt et al., 2018a,b; Kaspar et al., 2019; Diehl et al., 2021). Thus, AARC explicitly addresses gains as a developmental option in very old age (Baltes and Smith, 2003; Baltes et al., 2006). More concretely, the concept of AARC captures, on the one hand, the “success” of having survived many peers, but on the other hand also the fact that many individuals have to cope with serious declines and potential impairments in biological and functional capacity. In very old age, however, age-related losses may become more pronounced and it may become increasingly harder to appreciate the benefits that aging brought about. In fact, the strength and vulnerability integration model (SAVI, Charles, 2010) proposes that age-related gains might no longer be sufficient to retain optimal functioning and well-being in the Fourth Age. With respect to the prediction of survival, however, perceptions of age-related gains may be particularly important, because they may reflect motivational resources that may be activated in drawing on remaining reserve capacities and sources of resilience.

Important for assessing AARC in very old individuals, Kaspar et al. (2019) developed a short form of the AARC questionnaire specifically for use in large-scale surveys and in populations in which the administration of lengthy questionnaires is not feasible. This short form was used in the present study. Like for the long form (Brothers et al., 2019), the two-factor solution was confirmed using confirmatory factor analysis and independent samples (Kaspar et al., 2019; Sabatini et al., 2020). Findings for the AARC measure suggest that VOA in very old age are both a result of change in health status and engagement in life and a predictor of future health status (Spuling et al., 2013; Dutt et al., 2018a; Kaspar et al., 2021). However, no studies are currently available addressing the role of the AARC concept with regard to survival.

In addition to AARC as a predictor of survival, we focused on two areas of psychosocial resources that have shown an association with survival in younger age groups and, therefore, may also be relevant for predicting survival in advanced old age. First, we addressed how very old individuals manage to maintain a self-view of a purposeful, valuable life, as seen from an individual and societal point of view. In a meta-analysis of ten prospective studies with more than 136,000 participants, Cohen et al. (2016) synthesized evidence for a robust link between purpose in life (e.g., being useful to others, life engagement, life meaning) and all-cause mortality. Similarly, the feeling of being needed was found to be the single most important aspect of positive life orientation to predict survival in a large Finnish study of community-dwelling individuals aged 75 years or older (Tilvis et al., 2012).

Second, another major psychological challenge is to what extent very old adults are able to exert control over their lives and keep track of current societal developments. This point addresses the core challenge whether and to what extent individuals in the Fourth Age can maintain feelings of agency and avoid feelings of being dependent on others. With respect to societal development, feeling distant and disconnected from major trends, such as globalization or communication technology, may result in possible alienation and perceived obsolescence. Adverse effects of alienation have been described in the context of suicide ideation in subpopulations with mood disorder (Moore, 1997). However, it is obvious to also expect more general negative consequences for late-life survival via reduced motivation for or limited perceived control over health-related behavior in very old age. Multiple studies have established an association between control beliefs and mortality hazard in older adult samples (Infurna et al., 2011, 2013; Wiest et al., 2013; Infurna and Okun, 2015; Duan-Porter et al., 2017; Hülür et al., 2017). For example, Infurna et al. (2011) noted that perceived control may be related to all-cause mortality through various mechanisms, including the absence of health promoting behaviors, ineffective emotion regulation, low social integration, and the absence of stress buffering effects. However, most studies considered a broad age range and used measures of perceived control administered as early as midlife to predict subsequent 8-, 11-, or 19-year mortality. Effects of control beliefs as a risk factor for mortality are likely to be different in a highly selected subpopulation of individuals that have already outlived most of their birth cohort peers. Because Infurna and Okun (2015) found that perceived control decreased with age and change in social participation, control beliefs may be changing substantially during the Fourth Age and more studies on changes in perceived control in very old individuals are needed. In addition, most studies have focused on the concept of internal control or mastery. We assumed that a multidimensional understanding of control that explicitly includes both feelings of mastery and dependency on powerful others or chance could be particularly helpful in very old age. Very old individuals may experience increased risk of chronic health conditions (Bercovitz et al., 2019), social losses, or a shifting potential for agency (Wahl et al., 2012), and therefore may become increasingly confronted with changes beyond their personal control.

The aims of this study were to examine the contribution of very old adults’ subjective VOA as predictors of survival time across a 3.5-year observation interval. Individuals’ subjective VOA, as operationalized in terms of AARC-Gains and AARC-Losses, were incorporated into a set of established socio-behavioral predictors of survival. We expected subjective VOA to show substantial associations with survival time. Moreover, we expected positive (AARC-Gains) and negative facets (AARC-Losses) of individuals’ subjective VOA to contribute independently to the prediction of survival in those in advanced old age, because these predictors have been shown to be differentially related to developmental outcomes in younger age groups. We expected a remaining increment of predictive power due to AARC even after controlling for other major survival-relevant psychological resources. Therefore, we assumed that attribution of perceived change to aging itself should evolve from a process of integrating knowledge of conditions more prevalent with age (i.e., associations with health status change) and experiences in handling such change (e.g., associations with perceived control).

Materials and Methods

Participants and Procedures

Data came from a representative panel study on quality of life (QoL) and well-being of very old adults conducted in Germany’s most populous state, North-Rhine Westphalia (Wagner et al., 2018). For the first wave of measurement in 2017/2018, a random community sample of persons aged 80 years and older was selected in a multi-stage sampling procedure, assuring adequate coverage of persons living both in private households and institutional settings. Persons in older age groups (i.e., 85 years and older) and men were oversampled to allow for precise estimation of population parameters also in these smaller subpopulations. A detailed discussion of the sampling design and efficiency as well as representativity of the weighted sample is available elsewhere (Hansen et al., 2021).

A total of 1,863 computer-assisted personal interviews were conducted at participants’ homes to assess a wide array of individual QoL resources (e.g., social and health) and subjective QoL outcomes (e.g., valuation of life). The study protocol also included objective testing such as a screening for mild cognitive impairment (MCI). Informed consent was given by all participants after written and verbal explanation of the study aims and procedures. Mean age of the realized sample at the time of the interview at wave 1 was 87.0 years (SD = 4.5 years; range: 80.1 to 102.9 years). A total of 211 interviews (11.3%) were conducted in nursing homes. The sample included 176 interviews with proxy informants (e.g., partner 48.3%, adult child 23.9%, other 27.8%) where target persons were willing to be included in the study but were not able to conduct the 90 min interview themselves due to severe mental or physical health constraints.

Based on consent, a total of 1,612 wave 1 participants were re-contacted in 2019/2020. Personal contacting during fieldwork revealed that 237 individuals had died since the first interview. Additional information on survival status and date of death was collected 1.5 years after wave 2 during regular panel maintenance work. By March 5, 2021 a total of 391 (24.2%) of respondents had died. The study was approved by the ethical board of the medical faculty at the University of Cologne (Protocol #: 17-169).

Measures

Awareness of Age-Related Change

The 10-item short form of the Awareness of Age-Related Change scale (AARC-SF; Kaspar et al., 2019) was used as a brief measure of participants’ subjective VOA. The AARC-SF is multidimensional in capturing change across five behavioral domains: Health and physical functioning (PHYS), cognitive functioning (COG), interpersonal relations (INT), social-cognitive and social-emotional functioning (SC/SE), and lifestyle and engagement (LIFE). Half of the 10 items assess positive (gain-related) and half assess negative (loss-related) perceptions of age-related changes, respectively. The item stem is, “With my increasing age, I realize that …” and the response format ranges from 1 (not at all) to 5 (very much). A sample gain item (INT+ domain) is, “…I appreciate relationships and people much more.” A sample loss item (LIFE- domain) is, “…I have to limit my activities.” Kaspar et al. (2019) reported favorable psychometric properties and evidence for concurrent and discriminant validity. Composite reliability (see Revelle and Zinbarg, 2009) of the AARC-Gains and AARC-Losses scales in the current sample was acceptable given the built-in domain heterogeneity of the brief scales (MacDonald’s ω = 0.68 and 0.81, respectively).

Psychosocial Resources

Additional psychosocial resources predictive of survival in very old age were assessed in terms of appraisal of life (i.e., valuation of life, perceived obsolescence, feeling needed) and control (i.e., internal and external control foci).

The Valuation of Life Scale (VOL; Lawton et al., 2001) was used as a comprehensive measure of emotional and behavioral aspects of attachment to life in old age. The scale has 13 statements (e.g., “Life has meaning for me,” “I feel hopeful right now”) and a 3-point response scale 0 (no), 1 (neither/nor), 2 (yes) is suggested for use in very old respondents (Rott et al., 2001; Jopp et al., 2008, 2013). Gitlin et al. (2016) documented a two-factor structure of the VOL, which was replicated in the current sample. The two factors corresponded with an optimistic outlook in life (McDonald’s ω = 0.82) and personal engagement (McDonald’s ω = 0.87). Two items (i.e., “increasingly difficult to come to terms with today’s way of living”; “growing lack of fit between own values and those of society these days”) from the perceived obsolescence subscale of the Future Time Perspective Scale by Brandtstädter et al. (1997) and one item (i.e., “hard to stay oriented because society is changing so quickly these days”) from the anomia scale suggested by Gümüs et al. (2014) were used to measure respondents’ reflection of correspondence with values held by current society. Items were answered on a 4-point scale from 1 (not at all) to 4 (very much). Scale consistency was moderate for this perceived obsolescence composite in the current sample (MacDonald’s ω = 0.69) and comparable to Cronbach’s alpha of 0.72 reported for the 5-item obsolescence subscale by Brandtstädter et al. (1997). In addition to individual values, subjective perceptions of appraisal by society were assessed using the single item “feel needed by society” with response options 1 (not at all) to 4 (very much).

Perceived control was assessed using the Internal and External Control Beliefs scale (IE-4, Kovaleva et al., 2012), with two of the four items targeting internal control and two items targeting the external others and chance facets of external control beliefs suggested by Levenson (1972). Evidence of construct validity has been reported for this brief instrument by Kovaleva et al. (2012). Psychometric results for this dataset show both satisfactory consistency of the internal scale (MacDonald’s ω = 0.76) and uniqueness of the chance and powerful others indicators in the realm of external control beliefs, supporting a three factor interpretation.

Health and Cognitive Status

Adults’ self-reported performance on Basic Activities of Daily Living (ADL; Katz et al., 1963) and Instrumental Activities of Daily Living (IADL; Lawton and Brody, 1969) was used as a measure of everyday functioning. Specifically, we used five items to assess ADL (e.g., getting dressed and walking) and seven items to assess IADL (e.g., preparing meals and handling finances) with response options 0 (not possible without help), 1 (some help needed), 2 (no help needed). Reliability of the ADL and IADL scales in the current sample was high (MacDonald’s ω = 0.92 and 0.93, respectively).

Further, the number of self-reported currently treated health conditions was used as an indicator of multimorbidity. Hence, this measure refers to a subset of medical conditions with high salience for the individual in everyday life. This index was modified from the Self-Administered Comorbidity Questionnaire (SCQ; Sangha et al., 2003) to include medical conditions particularly relevant in old age (Wiest et al., 2014). The 19 conditions included were: Heart disease (e.g., insufficiency), heart attack, hypertension, respiratory or lung disease, diabetes, gastrointestinal disease, kidney disease, liver disease, hemophilia (e.g., anemia), cancer, mental disease (e.g., phobia and depression), bone and joint disease (e.g., osteoporosis, arthrosis, and arthritis), back pain, urinary disorder, insomnia, hearing impairment, visual impairment, neurological disease (e.g., Parkinson’s and dementia), and stroke.

In terms of cognitive status, the DemTect was developed as a brief screening tool for MCI and early stages of dementia (Kalbe et al., 2004). The test has subtests assessing immediate/delayed word recall, digit span memory, number transformation, and verbal fluency. Favorable diagnostic properties in identifying beginning cognitive decline and MCI have been reported in comparison to alternative screening tools such as the MMSE (Kalbe et al., 2013) and age-specific scoring instructions for persons 80 years or older have been reported by Kessler et al. (2014). In the case of proxy interviews, cognitive status was reported with the Global Deterioration Scale (GDS; Reisberg et al., 1982) in seven stages from 1 (no cognitive impairment) to 7 (most severe). Reisberg et al. (2011) aligned GDS stage 3 to correspond to a clinical presentation of MCI (Reisberg et al., 2011).

Sample Selectivity

Characteristics of the very old population estimated based on (1) wave 1 participants, (2) those willing to be contacted again for future waves, and (3) respondents still alive approximately 3.5 years after wave 1 are shown in Table 1 to estimate potential selectivity of the analysis sample and risk factors to longevity.