Changes in views on aging in later adulthood: the role of cardiovascular events

  • Susanne WurmEmail author
  • Maja Wiest
  • Julia K. Wolff
  • Ann-Kristin Beyer
  • Svenja M. Spuling
Original Investigation


A number of longitudinal studies have pointed to the long-term impact of different views on aging (VoA) on health in later life, whereas the reverse relationship has rarely been examined. Serious cardiovascular events such as myocardial infarction or stroke are life-threatening events which might in turn lead to changes in VoA. The present longitudinal study examined the effect of a cardiovascular event (CVE) on VoA over a three-year period using pooled data from three waves of the German Ageing Survey (2008, 2011, 2014, age range: 40–95 years). In order to account for alternative explanations for changes in VoA, individuals without CVE (n = 200) were matched to individuals who experienced a CVE (n = 202) using a propensity score matching procedure. Compared to individuals without CVE, individuals who experienced a CVE showed adverse changes in three VoA indicators (self-perceptions of aging as associated with physical losses/with ongoing development; subjective age). These results suggest that CVE can in fact change how individuals view their own aging. According to previous studies, this can lead to future health changes and thus become a health-related downward spiral. Health promotion programs could, therefore, profit by adding specific VoA interventions for individuals who experienced a CVE.


Views on aging Self-perceptions of aging Subjective age Life event Cardiovascular event Longitudinal study 


A constantly rising number of longitudinal studies have provided broad evidence on the impact of different views on aging (VoA) on health-related outcomes. Individuals with more positive age stereotypes or self-perceptions of aging (SPA), and those who feel younger (subjective age) do not only live longer, but also maintain better physical and mental health, better cognitive, and physical functioning, and have a lower likelihood of experiencing falls or hospitalization (for an overview: Westerhof and Wurm 2018; Wurm et al. 2017). In addition, several studies show that more negative VoA are associated with a higher likelihood of experiencing acute medical events such as cardiovascular events (CVE; Cheng et al. 2012; Levy et al. 2009). While these findings provide a consistent picture of the consequences of VoA on health, much less is known about factors that change VoA in later life. The present paper addresses this question by investigating the impact of CVE on changes in SPA and subjective age in later adulthood. Thus, the paper switches the perspective from CVE as an outcome to that of a potential trigger for changes in VoA.

Prevalence and consequences of cardiovascular diseases

Worldwide, cardiovascular diseases are the most common cause of death (Naghavi et al. 2015) and by far the leading cause of disease burden in older adults (Prince et al. 2015). According to the Global Burden of Disease Study, ischemic heart diseases lead to the largest health loss globally, including an estimated 7.29 million acute myocardial infarctions and 110.55 million prevalent cases of ischemic heart diseases in 2015; stroke is the second leading cause of health loss with 8.97 million ischemic, hemorrhagic, and other strokes (Roth et al. 2017). Cardiovascular diseases—like other chronic diseases—are often accompanied by negative emotional reactions (e.g., de Ridder et al. 2008), but also changes in personality, such as decreases in extraversion, emotional stability, and conscientiousness (Jokela et al. 2014). In addition, acute life-threatening medical events such as a heart attack or stroke lead to posttraumatic stress disorder symptoms in about 12–15% of survivors (Edmondson 2014). Unlike external traumatic events such as combat that mostly refer to experiences in the past, medical traumas are often perceived as ongoing events due to the permanently impaired physiological system. This is accompanied by a heightened likelihood of permanent or recurrent health problems and chronic disease management such as life-long adherence to medication regimen (e.g., Smith et al. 2011). These fundamental consequences of cardiovascular diseases, in particular of those with an acute onset, suggest that VoA are also likely to change as a consequence of cardiovascular diseases.

Health-related changes in VoA

First indications of health-related changes in VoA come from longitudinal studies that examine the impact of different health factors on subjective age, pointing to younger subjective age in individuals with better self-rated and functional health (Hughes and Lachman 2018; Ward 2013). Conversely, older subjective age was found in individuals who experienced the incidence of a medical condition (Schafer and Shippee 2010). A longitudinal study on the reciprocal relationship of health and subjective age found that better self-rated health predicted younger subjective age and vice versa (Spuling et al. 2013).

Health also seems to effect changes in SPA. A 16-year longitudinal study of adults 65 + years showed that medical conditions and decreases in physical functioning were associated with more negative SPA (Sargent-Cox et al. 2012). Two studies examined the reciprocal relationships between health (i.e., medical conditions, self-rated health, cognitive functioning) and SPA and point to the relevance of good health for gain-related SPA and poor health for the view that aging is associated with physical losses (Seidler and Wolff 2017; Wurm et al. 2007). Together, these findings suggest that the incidence of a cardiovascular event could contribute to adverse changes in VoA, that is, to older subjective age and negative changes in SPA.

Interestingly, a cross-sectional study that also considered SPA in different life domains came in part to contrary findings (Bryant et al. 2012): In line with previous studies, worse physical health in older adults of 60 + years was related to the view that aging is associated with more negative physical changes and more psychosocial losses; however, worse physical health was unexpectedly related to the view that aging is associated with more psychological growth. The study’s authors assume that this counterintuitive finding may reflect effective coping responses. In sum, these findings underline that health-related changes could have differential impact on gain- and loss-related VoA which is in line with recent theoretical approaches that emphasize a multidimensional conceptualization of VoA (Kornadt et al. 2019; Wurm et al. 2017).

Impact of CVE on different VoA

The finding from previous studies that medical conditions and health-related changes are able to predict VoA over time suggests that the occurrence of a CVE could trigger changes in VoA. Unlike chronic health problems, a major characteristic of CVEs is their sudden onset which often poses an acute life-threatening situation. In particular, their suddenness and severity, but also the increased likelihood of long-term consequences for health such as a permanent medication intake, suggest that serious health events are likely to change VoA. Based on previous findings, it is likely that CVEs foster the view that aging is associated with physical losses. Physical health is also a strong predictor of subjective age (Kotter-Grühn et al. 2016) which suggests that individuals feel older rather than younger after a CVE.

However, it is less clear whether CVEs also contribute to an increase or a decrease in gain-related VoA. These VoA do not refer to health but to psychological growth and ongoing personal development. There is evidence for both possibilities: Whereas a study that addressed the longitudinal impact of medical conditions on psychological growth showed a negative effect (Wurm et al. 2007), the findings of Bryant and colleagues (2012) suggest that worse physical health might also be related to more psychological growth.

Although the findings of Bryant et al. (2012) are correlative in nature, they are in line with theories on posttraumatic growth (Calhoun and Tedeschi 2001) that have shown that naturalistic reminders of death can lead to positive changes such as a higher appreciation of life, changed sense of priorities, and recognition of personal strengths and new possibilities (Tedeschi and Calhoun 1996). Posttraumatic growth has been shown to occur in the context of life-threatening diseases, for instance, in case of myocardial infarction (Garnefski et al. 2008), cancer (Shand et al. 2015), or acquired brain injury (Rogan et al. 2013).

The present study

The present study examines the role of CVE for changes in subjective age and SPA in later life over a 3-year period. As suggested in recent studies (Kornadt et al. 2019; Wurm et al. 2017), we follow a multidimensional concept of VoA by considering loss- and gain-related SPA separately. In addition, we include subjective age as a global indicator of VoA. Based on previous findings on the impact of health on subjective age and the role of biological markers for subjective age (Kotter-Grühn et al. 2016), our first hypothesis assumes that CVEs are detrimental to subjective age.


Over the study period, the discrepancy between subjective and chronological age will decrease on average for individuals who experience a CVE during the study period as compared to individuals without CVE.

SPA were considered based on a multidimensional approach. The first facet refers to the self-view that aging is associated with physical losses. The occurrence of a CVE probably reinforces this view.


Over the study period, the loss-related self-perception that aging is associated with physical losses will increase for individuals who experience a CVE during the study period as compared to individuals without CVE.

The second facet refers to the gain-related view that aging is associated with ongoing development and personal growth. As this facet does not refer to health changes, but to the recognition of personal strengths and new possibilities, we consider findings from posttraumatic growth research and assume that CVE may contribute to a growth-related VoA. However, as studies on the association between medical conditions and gain-related VoA diverge, we cannot derive a directed hypothesis.


Over the study period, the gain-related self-perception that aging is associated with ongoing development and personal growth will change on average more strongly for individuals with CVE as compared to individuals without CVE.

CVEs are often associated with depressive symptoms and/or functional limitations (e.g., Almas et al. 2015; Sackley et al. 2019). In order to test the independent impact of CVE on subjective age and SPA, the analysis controls for health as well as typical risk factors of CVEs such as hypertension (Piepoli et al. 2016). Propensity score matching is used to compare participants who experienced a CVE with participants with similar demographic characteristics who did not experience a CVE within the study period.



Data come from the German Ageing Survey (DEAS), a population-based survey representative of adults in Germany aged 40 to 85 years started in 1996. Every 6 years a new baseline sample is randomly drawn from German registries and systematically stratified by age, sex, and place of residence. Already included participants are reassessed, if they previously provided informed consent and could be re-contacted. Starting in 2008, participants were assessed longitudinally every 3 years. Data collection at each measurement point consists of a 90-min computer-assisted interview and a drop-off questionnaire. A more comprehensive description of design and methods of the DEAS has been reported elsewhere (Klaus et al. 2017).

For the present study, we used pooled data from three measurement occasions, which were assessed every 3 years (i.e., 2008, 2011, and 2014). All participants with at least two consecutive survey participations were included (2008–2011, 2011–2014, or 2008–2011–2014; N = 4583). Data were pooled in order to have T1 and T2 measurement 3 years apart for every person, regardless of the exact years (2008, 2011, or 2014) of assessment. For participants with only two consecutive survey participations (2008–2011, or 2011–2014) these two participations were used as their individual baseline (T1) and follow-up (T2) measurements. For participants with three participations (2008–2011–2014), we proceeded as follows: If participants reported a CVE in only one of the time periods (i.e., either between 2008–2011 or 2011–2014), we used this time period for the definition of their individual T1 and T2; if participants reported the occurrence of a CVE in both time periods, we used the first time period (2008–2011) to define their individual T1 and T2. Finally, if participants did not report a CVE in any time period, we also used the first time period (2008–2011) as their individual T1 and T2.


Cardiovascular events

The occurrence of a serious health event between two measurement points was assessed with one question: “Have you yourself suffered a serious illness or had an accident in the past 3 years?” Answers were coded as 0 = no, 1 = yes. When more than one serious health event had occurred, interviewers asked the respondents to report only on the most serious one. In addition, respondents were asked to name or describe the kind of disease. Following the approach of Levy and colleagues (Levy et al. 2009), diseases such as angina attacks, myocardial infarctions, strokes, and transient ischemic attacks were considered as CVE. Taken together, n = 202 (4.4%) individuals reported a CVE between T1 and T2. In the survey, only the year (but not day or month) in which a CVE occurred was assessed. Descriptive analyses show that an almost equal amount of individuals (n = 108; 53.5%) experienced a CVE in 2008/2009, or in the second half of the study period, i.e., in 2010/2011 (n = 94; 46.5%). Year of occurrence was not considered as additional variable in the analyses as the comparison group did not experience a CVE and would have no valid data on such a variable.

Views on aging (VoA)

The three VoA indicators considered in this study were similarly assessed on all measurement occasions.

Subjective age was measured as the difference between felt age and chronological age divided through chronological age resulting in a proportional discrepancy score (Kotter-Grühn et al. 2016). This score represents how much younger/older people feel with reference to their age. Higher values (positive or negative) indicate a larger discrepancy between felt and chronological age. Negative values indicate younger, while positive values indicate older subjective age as compared with chronological age. Just like in previous studies (Stephan et al. 2015), extreme outliers (three standard deviations below or above the sample mean) were deleted (n = 157).

Gain- and loss-related SPA were assessed with two widely used subscales of the AgeCog-Battery (Steverink et al. 2001; Wurm et al. 2007). Each scale consisted of four items rated on a 4-point scale ranging from 1 (strongly disagree) to 4 (strongly agree). Items were averaged for the scales with higher values indicating more loss-related views in terms of physical losses (e.g., “Aging means to me that I am less healthy”), or more gain-related views in terms of personal growth and ongoing development (e.g., “Aging means to me that my capabilities are increasing”). At both measurement points, the internal consistency was good for loss-related SPA (Cronbach’s α, T1 = .80; T2 = .81) and gain-related SPA (Cronbach’s α, T1 = .81; T2 = .79).


Individual T1 values regarding the presence/absence of three chronic conditions (high cholesterol, diabetes, hypertension), smoking (no/yes), physical activity (“How often do you do endurance sport” rated on a 6-point scale ranging from “never” to “daily”), body mass index (BMI; body weight, in kilogram, divided by squared body height, in meters) as well as functional health (physical functioning subscale of the SF-36; Bullinger and Kirchberger 1998), and depressive symptoms (German version of the CES-D scale; Hautzinger and Bailer 1993) were used as controls as they are typical risk factors of CVEs. Age, gender, region (East or West Germany), and education (ISCED, UNESCO, 2011) were used as matching variables for propensity score matching.

Data analyses

Propensity score matching

The majority of DEAS participants did not experience a CVE. Apart from having no CVE, these people may differ in many ways from individuals who experienced a CVE. Therefore, we used SPSS 22 to perform propensity score matching based on age, gender, region, and education to reasonably compare individuals of the sample with CVE (n = 202) and without CVE. Propensity score matching procedures provide a common method to balance non-randomized groups in an observational dataset on a number of potentially confounding variables (Stuart 2010) using a two-step process. First, a propensity score is estimated (probability for participants to be part of the group with CVE based on the values of the observed covariates; Rosenbaum and Rubin 1983) by conducting a logistic regression analysis. Second, the generated propensity score is used to match the group with CVE to the group without CVE in such a way that the participants in both groups are equivalent on the observed covariates included in the estimation of the propensity score (in our study: age, gender, region, and education). We decided against the inclusion of health variables in the estimation of the propensity scores because the sample size of individuals with CVE was too small to use many variables for the matching process. Furthermore, including health indicators as matching variables has some limitations as some of these indicators were assessed in the drop-off questionnaire, which has more missing values as compared to the personal interview.

To ensure close matches, the nearest neighbor matching algorithm was adopted using a very low caliper width of .00001 (the maximum distance in propensity scores that two matches can be apart from each other) without replacements. A higher maximum distance would have resulted in a comparison group that still differs significantly on the matching variables from the individuals who experienced a CVE. Therefore, we decided to stick with the low caliper which resulted in a slightly lower sample size in the comparison group. Thus, within the group without CVE, n = 200 case-matched controls were identified, i.e., for two individuals with CVE no matching partner could be identified. Using t tests, we tested possible mean differences in the matching variables at T1 between the two groups to confirm balance between the two groups.

Changes in views on aging

For preliminary analyses, t tests were used in the final sample with n = 402 individuals (n = 202 with CVE and n = 200 matched individuals without CVE) in order to check possible mean differences in the three views on aging constructs, and the health indicators at T1 as well.1

Three latent difference score models (one for each VoA indicator) were conducted in Mplus 7.3 (Muthén and Muthén 1998–2010) in order to test whether changes in VoA between T1 and T2 were affected by the occurrence of a CVE (Fig. 1). Models were estimated using full information maximum likelihood estimation. We used the RMSEA (root-mean-square error of approximation) and CFI (comparative fit index) to evaluate the model fit. Values of RMSEA close to .08 (or smaller) and CFI > .90 indicate a good fit (Marsh et al. 2004). In each model, the respective VoA indicator at T1 was controlled for the above-described control variables; all variables were allowed to correlate among each other and also to correlate with CVE. Additionally, we added CVE (0 = no vs. 1 = yes) as predictor for the latent difference score in VoA between T1 and T2. Effects with p < .05 were interpreted as significant. If CVE is a significant predictor for the latent difference score of the VoA indicator, this means that change in VoA between T1 and T2 differed between participants with and without CVE.
Fig. 1

Schematic depiction of the latent difference score model for the change in each VoA indicator. BMI body mass index, CVE cardiovascular event, VoA views on aging indicator (either subjective age, AgeCog physical losses, or AgeCog ongoing development); Δ VoA = latent difference score VoA T1 regressed on VoA T2. All T1-variables were allowed to correlate with each other (and with the CVE variable); arrows are not depicted for reasons of simplicity


Sample description

Means and standard deviations of VoA indicators and matching variables of the matched sample are displayed in Table 1 separately for participants with and without CVE. There were no significant differences between participants with or without CVE at T1 (all ps > .05) in any of the matching variables, which indicates appropriate propensity score matching. On average, participants were 68.3 years old (SD = 9.4), 30% were female, and the majority reported a medium education level. The preliminary analyses showed no significant differences in any VoA indicator at T1 between the two groups of participants with or without a CVE at T1 (all ps > .05). Individuals with CVE showed lower T1-levels of self-rated health (p < . 01), had higher prevalence of hypertension (p < .05), and showed higher depressive symptoms (p < .05) as compared to individuals without CVE. However, they did not differ with regard to the prevalence of cholesterol, diabetes, smoking, mean physical activity, mean BMI, or mean functional health.
Table 1

Sample characteristics for the matched samples for the matching variables and VoA indicators


Means (SD) [range]/percent


CVE = no (n = 200)

CVE = yes (n = 202)

Age T1

67.79 (9.41) [40–89]

68.90 (9.35) [43–90]

Gender (female)



Region (Western Germany)













Subjective AgeaT1

− 0.13 (0.09) [− 0.46–0.10]

− 0.13 (0.10) [− 0.44–0.12]

Subjective AgeaT2

− 0.11 (0.09) [− 0.42–0.20]

− 0.09 (0.10) [− 0.39–0.17]

AgeCog physical losses T1

2.80 (0.54) [1–4]

2.89 (0.56) [1–4]

AgeCog physical losses T2

2.76 (0.52) [1–4]

3.03 (0.58) [1.25–4]

AgeCog ongoing development T1

2.83 (0.61) [1–4]

2.80 (0.58) [1–4]

AgeCog ongoing development T2

2.83 (0.56) [1.5–4]

2.61 (0.62) [1–4]

aMeasured as proportional discrepancy score

Changes in views on aging

Model fit indices of the latent difference score models as well as model parameters are presented in Table 2. All three latent difference score models yielded a good fit: RMSEA was below 0.08 and CFI was above 0.90.
Table 2

Fit statistics and relevant parameters of the three latent change models

Model parameter

Subjective age

AgeCog physical losses

AgeCog ongoing development

















p value




Intercept VoA T1

− 0.162*



Intercept Δ VoA


− 0.013

− 0.038

b for CVE → Δ VoA



− 0.154*

*p < .05; Δ VoA = latent difference score VoA T1 regressed on VoA T2; b = unstandardized regression weight

In line with the hypotheses, changes over time in VoA indicators differed according to the experience of a CVE in all three VoA models: CVE was a significant predictor for the latent change of subjective age (bCVE = 0.02, p = .017), loss-related SPA (bCVE = 0.166, p = .001), as well as gain-related SPA (bCVE = − 0.154, p = .002).

The discrepancy between subjective and chronological age decreased over the study period of 3 years which is reflected in a significant mean change of subjective age in individuals without CVE (∆subjective age(t1−t2) = 0.019, p = .003). As expected, this decrease was stronger for those individuals who experienced a CVE (∆subjective age(t1−t2) = 0.039). Post hoc analyses in which we computed the original latent change model only for those individuals who experienced a CVE revealed that the latent change of subjective age is significant in this subgroup (p < .001). Thus, hypothesis 1 (H1) is supported by the results.

The loss-related self-perception of aging did not significantly change over time in individuals without CVE (∆AgeCog physical losses(t1−t2) = − 0.013, p = .733). In contrast, mean change of loss-related self-perception of aging in fact increased in individuals who experienced a CVE (∆AgeCog physical losses(t1−t2) = 0.153). Post hoc analyses supported H2 and revealed a significant increase in loss-related SPA in individuals who experienced a CVE (p < .001).

Also, the gain-related self-perception of aging did not significantly change over time in individuals without CVE (∆AgeCog ongoing development(t1−t2) = − 0.038, p = .308). However, mean change over time in individuals who experienced a CVE was more negative (∆AgeCog ongoing development(t1−t2) = − 0.192). Post hoc analyses also supported H3 and revealed a significant decrease in gain-related SPA in individuals who experienced a CVE (p < .001).


This study examined whether and how different VoA are affected by a severe health event. Pre- and post-event measurements of individuals were investigated using a change score model to determine the effect of a CVE on VoA over a period of 3 years. Using large-scale survey data, it was possible to compare a case-matched control group with the same socio-demographic features to a group that experienced a CVE.

Based on previous research, we were expecting a negative effect of a CVE on subjective age and loss-related SPA, which was supported in the present study. Overall, subjective age decreased, which means that individuals felt older 3 years after T1, an effect which was stronger for individuals who experienced a CVE. However, on average, individuals still felt younger as compared to their chronological age which corresponds to other studies (e.g., Rubin and Berntsen 2006). The loss-related SPA also worsened for individuals who experienced a CVE: They associated aging more strongly with health-related losses over time. Remarkably, in the group without CVE, loss-related SPA remained stable over time. Gain-related SPA changed more strongly for those who experienced a CVE, namely in the negative direction. This means, aging was less associated with personal growth. In the group without CVE, gain-related SPA also remained stable.

Cardiovascular events are able to change VoA

VoA are subject to change over the lifespan induced by social, biological, or individual changes (Kornadt et al. 2019). Probably, these changes in VoA take place rather slowly, except in case of immediate (health) events.

Investigating the effect of serious health events on VoA is important to understand the complex interaction between VoA and health. So far, long-term longitudinal studies have shown that negative VoA have detrimental effects on health and likewise, worse health is associated with more negative VoA (for a review, see Westerhof et al. 2014). Also, one previous study has shown that negative VoA increase the likelihood of CVE (Levy et al. 2009). Addressing the reverse effect of CVE on VoA for the first time, we found a detrimental effect of CVEs on all considered VoA facets. The finding that loss- and gain-related VoA only changed for individuals who experienced a CVE during the three-year period suggests that crucial life events induce changes in SPA within a shorter period of time.

Possible explanations for the effect of CVE on VoA

CVEs are characterized by a sudden and often life-threatening onset, followed by high risk of long-term health impairments. The suddenness and seriousness of CVEs and accompanied changes in health are a possible reason why CVEs turned out to affect VoA within a shorter period of time. Keller and her colleagues (1989) found that not only illnesses such as arthritis, but also strokes or heart attacks are often considered to be caused by aging and less by other factors such as health behaviors or lifestyle choices. This can have detrimental consequences: Stewart et al. (2016) reported that individuals attributing a heart attack or stroke to old age were less likely to make lifestyle changes in order to avoid another heart attack or stroke and were more likely to be hospitalized over a follow-up period of 3 years. The attribution of CVE to own age might have led to the feeling that nothing can be done about these health problems, which can foster negative emotions that are also detrimental to health. Health events such as CVEs may affect VoA also as they function as a marker of age group membership for the individual. This could activate age stereotypes held by the individuals and trigger their internalization into the self-concept (Levy 2009; Weiss and Kornadt 2018), and thus may lead to worse VoA.

Interestingly, other age-associated events (widowhood, death of a parent) and age-specific events (transition to retirement) were found to be unrelated to changes in subjective age (Schafer and Shippee 2010; Ward 2013). Another study found no effect of hospitalization or bereavement on changes in age stereotypes over a 10-year period (Levy et al. 2015); at baseline, however, more negative age stereotypes were associated with a 50% greater likelihood of experiencing a hospitalization. Thus, both the type of event and whether self- vs. other-related VoA are examined seem to make a difference.

Future research on the impact of events on VoA

Apart from the present study, longitudinal research on the effect of age-associated or age-related events on SPA has been missing. It is likely that SPA are more prone to change by age-associated events than age stereotypes as they are self-directed and more strongly affected by personal experiences, but this needs to be further examined in future research.

Although the literature on posttraumatic growth and the previously described cross-sectional study on SPA and health (Bryant et al. 2012) point to positive associations between poor health and gain-related self-perceptions, we did not find longitudinal evidence for this association. This finding notwithstanding, there may be individuals showing more gain-related VoA following a CVE. Intraindividual differences in changes could not be addressed in our relatively small and heterogeneous sample in which we first aimed to examine whether CVEs make any difference at all on changes in VoA. This is why we compared similar groups of individuals with and without CVE. Other self-perceptions such as self-rated health have been found to remain stable or to improve after a serious health event for some individuals (Spuling et al. 2017; Wurm et al. 2008). Also, studies on resilience show that some but not all individuals adapt well to major health events with regard to well-being (deRoon-Cassini et al. 2010; Lucas 2007). This calls for future research examining differential changes in VoA after major health events, and also life events in general. It is conceivable that some resources (e.g., coping strategies) can differentiate between individuals who perceive an increase in loss-related vs. gain-related VoA. Future studies should therefore follow a differential approach, for instance, by including moderating factors.

The present study suggests that CVE can have negative effects on VoA. Targeted interventions to alter these VoA following a CVE may be a strategy to prevent a health-related downward spiral. Previous studies have shown that negative VoA promote lower expectations of the future (Voss et al. 2017) which in turn can be detrimental for health behavior and health-care behavior (Kim 2008; Meisner and Baker 2013). Recent randomized controlled trials provide evidence that both SPA and subjective age can be systematically changed by interventions (Beyer et al. 2019; Stephan et al. 2013; Wolff et al. 2014). Future studies should therefore examine when and for whom VoA interventions might be useful after a serious health event. Also, a differential approach could be advisable as for some individuals the attribution of a CVE to age seems to be unfavorable (Stewart et al. 2016), whereas for others, age attribution may serve to make sense of a traumatic experience and assume a self-protective function (Wrosch and Heckhausen 2005).

In addition, if a CVE occurs somewhat earlier in life, individuals are more likely to perceive this event as off-time which can have stronger detrimental effects compared to having a CVE in old age, in which health events are more likely and thus more often considered on-time and normal for this age (Neugarten 1996). Individuals in their 40 s or 50 s less likely attribute a CVE to their age. Thus, the effect of CVE on VoA may be different in this age group compared to older age groups. In the present study, we did not examine age-related differences due to small sample size; however, future research should address this question by using a different sampling strategy in order to contrast a larger number of middle-aged and older adults who experienced a CVE.

Strengths and limitations

In contrast to the majority of previous studies on VoA and health, the present study did not investigate the effect of VoA on health but considered the reverse direction of a serious health event on VoA. Taking advantage of large survey data, we were able to include pre- and post-event data as well as a control group that did not experience a CVE within the observed three-year period. We matched a subgroup (case-matched control group) of individuals without CVE to the sample of individuals who experienced a CVE. Using propensity score matching techniques, we were therefore able to compare longitudinal changes in VoA between two groups of individuals with similar socio-demographic features who differed regarding the (non-)experience of a CVE. Another strength of this study is the use of several indicators of VoA. So far, most studies examined subjective age or SPA, whereas the present study considered subjective age as well as loss- and gain-related SPA.

The study has also some limitations which need to be considered. As this study was based on a large-scale prospective study, only a small number of individuals experienced a CVE within 3 years and information on CVE was based on self-reports. A study in a clinical context on individuals with CVE could have provided a larger sample size and also medical data on the CVE. However, clinical studies usually do not have any information about individuals before the event and cannot be compared with a non-clinical population. However, due to the relatively small sample size, we were not able to contrast age groups or further investigate interindividual differences, for example with regard to possible posttraumatic growth. Likewise, we were not able to include health indicators as matching variables due to the relatively small sample size to draw matches from, and also due to the fact that some health variables were assessed in the drop-off questionnaire of the survey which has a higher rate of missing values. Also, it should be kept in mind that individuals with particularly severe CVE were probably more likely to drop out of the study. As these individuals could have experienced a stronger decline in their VoA, the present findings tend to underestimate rather than overestimate changes in VoA for individuals with CVE. Descriptive information about the year in which a CVE occurred showed that the percentage of individuals who experienced a CVE in the year of the T2-interview was smaller (10.4%) compared to the other years between T1 and T2 (between 23.3–36.1% for each measurement point).

Having only one measurement before and after a CVE limits the scope of data analysis. Using change score models, we were able to examine the effect of CVE on change in VoA directly. However, two measurement points will always give a limited picture; extended longitudinal follow-ups would be desirable for future studies and would also allow testing for nonlinear changes in VoA.

In addition, the present approach of comparing individuals with and without CVE did not allow controlling for the time distance between the CVE and the measure of VoA. Thus, individuals who experienced a CVE almost 3 years ago and individuals who experienced a CVE a few months ago were treated equally in the model. A measurement burst study would allow exploring changes in VoA after a CVE in more depth. Finally, a longer study period, for example, of 10 years would make it possible to examine the long-term effects of CVE on the development of VoA—as well as moderators thereof. Also, this may allow showing changes in VoA for the non-event group. Thus, whether VoA decline further in the future, remain stable, or improve after a longer time remains an open research question.


The present study provides first evidence for negative effects of CVE on VoA. This adds to the literature by showing that more negative VoA do not only increase the likelihood of experiencing a CVE (Levy et al. 2009) but that also conversely, VoA are negatively affected by a CVE. In order to prevent a health-related downward spiral after a CVE, it could be a promising approach to enhance health promotion and cardiac rehabilitation by including intervention modules on VoA. The finding that the occurrence of CVE was not associated with positive effects on VoA supports the need for interventions.


  1. 1.

    All following analyses were computed both with this sample (n = 402) and additionally with a reduced sample (n = 400) in which those two individuals were excluded for whom no case-matched control could be identified. Because the analyses came to similar results, we report only the findings for the total sample (n = 402).



This work is a result of the research network “Images of Aging: Via a dynamic life span model to new perspectives for research and practice,” funded by a Grant of the German Research Foundation (KL 3072/1-1). The German Ageing Survey (DEAS) was funded under Grant 301-6083-05/003*2 by the German Federal Ministry for Family, Senior Citizens, Women, and Youth. The content is the sole responsibility of the authors.


  1. Almas A, Forsell Y, Iqbal R, Janszky I, Moller J (2015) Severity of depression, anxious distress and the risk of cardiovascular disease in a Swedish population-based cohort. Plos One 10:e0140742. CrossRefGoogle Scholar
  2. Beyer A-K, Wolff JK, Freiberger E, Wurm S (2019) Are self-perceptions of ageing modifiable? Examination of an exercise programme with vs. without a self-perceptions of ageing-intervention for older adults. Psychol Health 34:661–676CrossRefGoogle Scholar
  3. Bryant C, Bei B, Gilson K, Komiti A, Jackson H, Judd F (2012) The relationship between attitudes to aging and physical and mental health in older adults. Int Psychogeriatr 24:1674–1683. CrossRefGoogle Scholar
  4. Bullinger M, Kirchberger I (1998) SF-36 Fragebogen zum Gesundheitszustand. Handanweisung Zeitschrift für Medizinische Psychologie 4:190–191. CrossRefGoogle Scholar
  5. Calhoun LG, Tedeschi RG (2001) Posttraumatic growth: The positive lessons of loss. In: Neimeyer RA (ed) Meaning reconstruction & the experience of loss. American Psychological Association, Washington, DC, US, pp 157–172. Doi: CrossRefGoogle Scholar
  6. Cheng ST, Yip LC, Jim OT, Hui AN (2012) Self-perception of aging and acute medical events in chronically institutionalized middle-aged and older persons with schizophrenia. Int J Geriatr Psychiatry 27:907–913. CrossRefGoogle Scholar
  7. de Ridder D, Geenen R, Kuijer R, van Middendorp H (2008) Psychological adjustment to chronic disease. Lancet 372:246–255. CrossRefGoogle Scholar
  8. deRoon-Cassini TA, Mancini AD, Rusch MD, Bonanno GA (2010) Psychopathology and resilience following traumatic injury: a latent growth mixture model analysis. Rehabil Psychol 55:1–11. CrossRefGoogle Scholar
  9. Edmondson D (2014) An enduring somatic threat model of posttraumatic stress disorder due to acute life-threatening medical events. Soc Personal Psychol Compass 8:118–134. CrossRefGoogle Scholar
  10. Garnefski N, Kraaij V, Schroevers MJ, Somsen GA (2008) Post-traumatic growth after a myocardial infarction: a matter of personality, psychological health, or cognitive coping? J Clin Psychol Med Settings 15:270–277. CrossRefGoogle Scholar
  11. Hautzinger M, Bailer M (1993) Allgemeine Depressions-Skala (ADS). Deutsche Form der “Center for Epidemiogical Studies Depression Scale (CES-D)”. MainzGoogle Scholar
  12. Hughes ML, Lachman ME (2018) Social comparisons of health and cognitive functioning contribute to changes in subjective age. J Gerontol B Psychol Sci Soc Sci 73:816–824. CrossRefGoogle Scholar
  13. Jokela M, Hakulinen C, Singh-Manoux A, Kivimäki M (2014) Personality change associated with chronic diseases: pooled analysis of four prospective cohort studies. Psychol Med 44:2629–2640. CrossRefGoogle Scholar
  14. Keller ML, Leventhal H, Prohaska TR, Leventhal EA (1989) Beliefs about aging and illness in a community sample. Res Nurs Health 12:247–255CrossRefGoogle Scholar
  15. Kim SH (2008) Older people’s expectations regarding ageing, health-promoting behaviour and health status. J Adv Nurs 65:84–91CrossRefGoogle Scholar
  16. Klaus D, Engstler H, Mahne K, Wolff JK, Simonson J, Wurm S, Tesch-Römer C (2017) Cohort Profile: The German Ageing Survey (DEAS). Int J Epidemiol 46:1105. CrossRefGoogle Scholar
  17. Kornadt AE, Kessler E-M, Wurm S, Bowen CE, Gabrian M, Klusmann V (2019) Views on aging: a life span perspective. Eur J Ageing. CrossRefGoogle Scholar
  18. Kotter-Grühn D, Kornadt AE, Stephan Y (2016) Looking beyond chronological age: current knowledge and future directions in the study of subjective age. Gerontology 62:86–93. CrossRefGoogle Scholar
  19. Levy BR (2009) Stereotype embodiment. A psychosocial approach to aging. Curr Dir Psychol Sci 18:332–336CrossRefGoogle Scholar
  20. Levy BR, Zonderman AB, Slade MD, Ferrucci L (2009) Age stereotypes held earlier in life predict cardiovascular events in later life. Psychol Sci 20:296–298CrossRefGoogle Scholar
  21. Levy BR, Slade MD, Chung PH, Gill TM (2015) Resiliency over time of elders’ age stereotypes after encountering stressful events. J Gerontol Ser B 70:886–890. CrossRefGoogle Scholar
  22. Lucas RE (2007) Adaptation and the set-point model of subjective well-being. Does happiness change after major life events? Curr Dir Psychol Sci 16:75–79CrossRefGoogle Scholar
  23. Marsh AP, Hau K-T, Wen Z (2004) In search of golden rules: comment on hypothesis-testing approaches to setting cutoff values for fit indexes and dangers in overgeneralizing Hu and Bentler’s (1999) findings. Struct Equ Model 11:320–341. CrossRefGoogle Scholar
  24. Meisner BA, Baker J (2013) An exploratory analysis of aging expectations and health care behavior among aging adults. Psychol Aging 28:99–104. CrossRefGoogle Scholar
  25. Muthén LK, Muthén BO (1998-2010) Mplus user’s guide. Sixth Edition edn. Muthén & Muthén, Los Angeles, CAGoogle Scholar
  26. Naghavi M, Wang H, Lozano R, Davis A, Liang X, Zhou M et al (2015) Global, regional, and national age–sex specific all-cause and cause-specific mortality for 240 causes of death, 1990–2013: a systematic analysis for the Global Burden of Disease Study 2013. Lancet 385:117–171. CrossRefGoogle Scholar
  27. Neugarten BL (1996) The meanings of age. The University of Chicago Press, ChicagoGoogle Scholar
  28. Piepoli MF et al (2016) 2016 European guidelines on cardiovascular disease prevention in clinical practice: the sixth joint task force of the European Society of Cardiology and Other Societies on Cardiovascular Disease Prevention in Clinical Practice. Eur Heart J 37:2315–2381. CrossRefGoogle Scholar
  29. Prince MJ, Wu F, Guo Y, Gutierrez Robledo LM, O’Donnell M, Sullivan R, Yusuf S (2015) The burden of disease in older people and implications for health policy and practice. Lancet 385:549–562. CrossRefGoogle Scholar
  30. Rogan C, Fortune DG, Prentice G (2013) Post-traumatic growth, illness perceptions and coping in people with acquired brain injury. Neuropsychol Rehabil 23:639–657. CrossRefGoogle Scholar
  31. Rosenbaum PR, Rubin DB (1983) The central role of the propensity score in observational studies for causal effects. Biometrika 70:41–55. CrossRefGoogle Scholar
  32. Roth GA et al (2017) Global, regional, and national burden of cardiovascular diseases for 10 causes, 1990 to 2015. J Am Coll Cardiol 70:1–25. CrossRefGoogle Scholar
  33. Rubin DC, Berntsen D (2006) People over forty feel 20% younger than their age: subjective age across the lifespan. Psychon Bull Rev 13:776–780CrossRefGoogle Scholar
  34. Sackley CM, Mant J, McManus RJ, Humphreys G, Sharp L, Mares K, Savva GM (2019) Functional and emotional outcomes after transient ischemic attack: a 12-month prospective controlled cohort study. Int J Stroke. CrossRefGoogle Scholar
  35. Sargent-Cox KA, Anstey KJ, Luszcz MA (2012) Change in health and self-perceptions of aging over 16 years: the role of psychological resources. Health Psychol 31:423–432. CrossRefGoogle Scholar
  36. Schafer MH, Shippee TP (2010) Age identity in context. Soc Psychol Q 73:245–264. CrossRefGoogle Scholar
  37. Seidler AL, Wolff JK (2017) Bidirectional associations between self-perceptions of aging and processing speed across 3 years. GeroPsych 30:49–59. CrossRefGoogle Scholar
  38. Shand LK, Cowlishaw S, Brooker JE, Burney S, Ricciardelli LA (2015) Correlates of post-traumatic stress symptoms and growth in cancer patients: a systematic review and meta-analysis. Psychooncology 24:624–634. CrossRefGoogle Scholar
  39. Smith SC Jr et al (2011) AHA/ACCF secondary prevention and risk reduction therapy for patients with coronary and other atherosclerotic vascular disease: 2011 update: a guideline from the American Heart Association and American College of Cardiology Foundation. J Am Coll Cardiol 58:2432–2446. CrossRefGoogle Scholar
  40. Spuling S, Miche M, Wurm S, Wahl H-W (2013) Exploring the causal interplay of subjective age and health dimensions in the second half of life: A cross-lagged-panel analysis. Zeitschrift für Gesundheitspsychologie 21:5–15. CrossRefGoogle Scholar
  41. Spuling S, Wolff JK, Wurm S (2017) Response shift in self-rated health after serious health events in old age. Soc Sci Med 192:85–93. CrossRefGoogle Scholar
  42. Stephan Y, Chalabaev A, Kotter-Grühn D, Jaconelli A (2013) “Feeling younger, being stronger”: an experimental study of subjective age and physical functioning among older adults. J Gerontol Ser B Psychol Sci Soc Sci 68:1–7. CrossRefGoogle Scholar
  43. Stephan Y, Sutin AR, Terracciano A (2015) “Feeling younger, walking faster”: subjective age and walking speed in older adults. Age 37:86CrossRefGoogle Scholar
  44. Steverink N, Westerhof GJ, Bode C, Dittmann-Kohli F (2001) The personal experience of aging, individual resources, and subjective well-being. J Gerontol B Psychol Sci Soc Sci 56:P364–P373. CrossRefGoogle Scholar
  45. Stewart TL, Chipperfield JG, Perry RP, Hamm JM (2016) Attributing heart attack and stroke to “old age”: implications for subsequent health outcomes among older adults. J Health Psychol 21:40–49. CrossRefGoogle Scholar
  46. Stuart EA (2010) Matching methods for causal inference: a review and a look forward. Stat Sci Rev J Inst Math Stat 25:1–21. CrossRefGoogle Scholar
  47. Tedeschi RG, Calhoun LG (1996) The posttraumatic growth inventory: measuring the positive legacy of trauma. J Trauma Stress 9:455–472. CrossRefGoogle Scholar
  48. Voss P, Kornadt AE, Rothermund K (2017) Getting what you expect? Future self-views predict the valence of life events. Dev Psychol 53:567–580. CrossRefGoogle Scholar
  49. Ward RA (2013) Change in perceived age in middle and later life. Int J Aging Hum Dev 76:251–267. CrossRefGoogle Scholar
  50. Weiss D, Kornadt AE (2018) Age-stereotype internalization and dissociation: contradictory processes or two sides of the same coin? Curr Dir Psychol Sci 27:447–483. CrossRefGoogle Scholar
  51. Westerhof GJ, Wurm S (2018) Subjective aging and health. In: Knight BG, Wahl H-W (eds) Oxford research encyclopedia of psychology and aging. Oxford University Press, Oxford, pp 1–30. CrossRefGoogle Scholar
  52. Westerhof G et al (2014) The influence of subjective aging on health and longevity: a meta-analysis of longitudinal data. Psychol Aging 29:793–802. CrossRefGoogle Scholar
  53. Wolff JK, Warner LM, Ziegelmann JP, Wurm S (2014) What do targeting positive views on ageing add to a physical activity intervention in older adults? Results from a randomised controlled trial. Psychol Health 29:915–932CrossRefGoogle Scholar
  54. Wrosch C, Heckhausen J (2005) Being on-time or off-time: developmental deadlines for regulating one’s own development. In: Perret-Clermont A-N (ed) Thinking time: a multidisciplinary perspective on time. Hogrefe & Huber Publishers, Cambridge, pp 110–123Google Scholar
  55. Wurm S, Tesch-Römer C, Tomasik MJ (2007) Longitudinal findings on aging related cognitions, control beliefs, and health in later life. J Gerontol Ser B Psychol Sci. 62B:P156–P164. CrossRefGoogle Scholar
  56. Wurm S, Tomasik MJ, Tesch-Römer C (2008) Serious health events and their impact on changes in subjective health and life satisfaction: the role of age and a positive view on aging. Eur J Ageing 5:117–127. CrossRefGoogle Scholar
  57. Wurm S, Diehl M, Kornadt AE, Westerhof G, Wahl H-W (2017) How do views on aging affect health outcomes in adulthood and late life? Explanations for an established connection. Dev Rev 46:27–43. CrossRefGoogle Scholar

Copyright information

© Springer Nature B.V. 2019

Authors and Affiliations

  1. 1.Institute for Community Medicine, Department of Social Medicine and PreventionUniversity Medicine GreifswaldGreifswaldGermany
  2. 2.Freie Universität BerlinBerlinGermany
  3. 3.IGES Institute BerlinBerlinGermany
  4. 4.Charité - Universitätsmedizin BerlinBerlinGermany
  5. 5.German Centre of Gerontology (DZA)BerlinGermany

Personalised recommendations