Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Population Priorities for Successful Aging: A Randomized Vignette Experiment

Population Priorities for Successful Aging: A Randomized Vignette Experiment Objectives: Aging populations have led to increasing interest in “successful aging” but there is no consensus as to what this entails. We aimed to understand the relative importance to the general population of six commonly-used successful aging dimensions (disease, disability, physical functioning, cognitive functioning, interpersonal engagement, and productive engagement). Method: Two thousand and ten British men and women were shown vignettes describing an older person with randomly determined favorable/unfavorable outcomes for each dimension and asked to score (0–10) how successfully the person was aging. Results: Vignettes with favorable successful aging dimensions were given higher mean scores than those with unfavorable dimensions. The dimensions given greatest importance were cognitive function (difference [95% confidence interval {CI}] in mean scores: 1.20 [1.11, 1.30]) and disability (1.18 [1.08, 1.27]), while disease (0.73 [0.64, 0.82]) and productive engage- ment (0.58 [0.49, 0.66]) were given the least importance. Older respondents gave increasingly greater relative importance to physical function, cognitive function, and productive engagement. Discussion: Successful aging definitions that focus on disease do not reflect the views of the population in general and older people in particular. Practitioners and policy makers should be aware of older people’s priorities for aging and understand how these differ from their own. Keywords: Attitudes, Cognition, Health, Interpersonal relations, Successful aging Industrialized populations are aging, (Christensen et  al., 2010) and are more likely to be in paid employment (Spijker 2009) prompting debate about whether growing propor- & MacInnes, 2013) or volunteering (Morrow-Howell, tions of older individuals require increasing investment in 2010), resulting in a growing interest in the notion of “suc- health and long-term care.(Bloom et  al., 2015) Early re- cessful aging” (Araújo, Ribeiro, Teixeira, & Paúl, 2016; search and policy often concentrated on more unfavorable Bowling, 2007; Katz & Calasanti, 2015; Martin et  al., aspects of aging, particularly at the population level, result- 2015; Martinson & Berridge, 2015; Nimrod & Ben-Shem, ing in anxiety and negativity about its potential impact on 2015; Stowe & Cooney, 2015). In addition, older people society (Baltes & Carstensen, 1996). However, more recent are often more positive about the aging process than those evidence suggests that, compared with their peers in pre- involved in their care, demonstrating high levels of adjust- vious cohorts, older people today have better physical and ment, acceptance, and resilience (Manning, Carr, & Kail, cognitive functioning (Christensen et  al., 2013; Vaupel, 2016). These views are consistent with recent challenges to © The Author(s) 2018. Published by Oxford University Press on behalf of The Gerontological Society of America. 293 This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. Downloaded from https://academic.oup.com/psychsocgerontology/article/75/2/293/5033526 by DeepDyve user on 16 July 2022 294 Journals of Gerontology: PSYCHOLOGICAL SCIENCES, 2020, Vol. 75, No. 2 the current World Health Organization (WHO) definition 2012; Montross et  al., 2006; Strawbridge, Wallhagen, & of health as “a state of complete physical, mental and social Cohen, 2002; Young, Frick, & Phelan, 2009). In spite of well-being,” which recommend an alternative formulation decades of research, there is still no firm consensus as to in terms of individuals’ ability to adapt and self-manage what successful aging entails, with recent special issues of (Huber et  al., 2011). However, there is a danger that the Journals of Gerontology: Social Sciences (Pruchno & Carr, attitudes of practitioners and policy makers involved with 2017) and The Gerontologist (Pruchno, 2015) devoted to older people are based on out-of-date and potentially mis- the question. In addition, the development of appropriate leading information and differences in the beliefs of older metrics has been identified as a research priority by WHO people and professionals are particularly pertinent in the (World Health Organisation, 2015). However, the most context of shared decision making and patient-centered widely adopted multidimensional model of successful aging care. Although the value of patients’ opinions in shap- was proposed by Rowe and Kahn (Rowe & Kahn, 1997) ing and informing clinical practice is well recognized in and incorporates six dimensions: (a) avoidance of disease; principle, recent results from the MAGIC (Making Good (b) avoidance of disability; (c) maintenance of good phys- Decisions in Collaboration) programme (Joseph-Williams ical function; (d) maintenance of good cognitive function; et al., 2017) highlight that, in practice, some clinicians “fail (e) good interpersonal social engagement (contacts and to recognize that patients’ values, opinions or preferences transactions with others); and (f) good productive engage- ... may differ from their own.” Moreover, the authors report ment (engagement in activities of value to society such as that older people may be particularly reluctant to share working or volunteering). Conventionally, according to this their views. If policy and practice are to support people to definition an individual is considered to be aging success- age successfully, a greater understanding of the extent to fully if they meet all six criteria. This straightforward char- which people value different aspects of aging is required. acterization moves beyond the biomedical to include social Clinicians, researchers, and policy makers worldwide and productive engagement, which have been shown to be agree that “successful aging” is an important goal (Bloom of substantial importance to older people (Bowling, 2007; et  al., 2015; Commission of the European Communities, Cosco, Prina, Perales, Stephan, & Brayne, 2013; Depp, 2009; United Nations, 2002) but its meaning remains un- Glatt, & Jeste, 2007), and positive associations have been clear. A  wide array of successful aging definitions have reported between this definition of successful aging and been proposed in the literature (Lupien & Wan, 2004). self-reported well-being (Strawbridge et al., 2002), health, Some focus on specific domains, for example, biomedical and life satisfaction (Whitley, Popham, & Benzeval, 2016) aging, covering compression of morbidity and genetic fac- in older people. However, the extent to which it reflects tors, while others consider cognitive or psychosocial aging, perceptions of successful aging in the general population with an emphasis on subjective well-being and personality. continues to be widely debated (Bowling & Iliffe, 2011; While these models provide insights into particular aging Ferri, James, & Pruchno, 2009; Martinson & Berridge, processes and are valuable in developing the specific poli- 2015; Montross et  al., 2006; Phelan, Anderson, LaCroix, cies that underpin them, they can also be limited in their & Larson, 2004; Stowe & Cooney, 2015; Strawbridge ability to predict or explain other aspects of aging and this et al., 2002; Young et al., 2009) and the relative importance has led to the development of multidimensional models of each dimension is unknown. Rowe and Kahn (Rowe & that include multiple aging dimensions.(Lupien & Wan, Kahn, 2015) have also entered this debate, acknowledging 2004) Again, many different multidimensional models have the limitations of their model but supporting the notion been proposed, some focusing on successful aging as an that “its extensive use in scientific enquiry warrants modi- adaptive process, such as the Selection, Optimization and fication over disposal.” In their discussion, they propose Compensation (SOC) model proposed by Baltes and Baltes new priorities for research, including the need to take a (Baltes & Baltes, 1990), and others focusing on successful lifecourse perspective to aging, to focus more on the poten- aging as measureable state, such as the MacArthur model tial benefits of an aging society, and to consider successful proposed by Rowe and Kahn (Rowe & Kahn, 1997). aging not only at the level of the individual but also at the There are also differences in multidimensional models level of society. proposed by different groups. For example, while the ma- In order to promote successful aging at the societal jority of operational definitions of successful aging include level, it is vital to understand what the general popula- physiological factors such as disease, disability, and phys- tion consider to be successful aging. Rather than propose ical function (Cosco et  al., 2014a; Depp & Jeste, 2006), another new successful aging model for additional debate, considerably fewer include dimensions known to be of we aim instead to understand population attitudes toward value to older people, such as functioning, social engage- the most commonly employed existing model with a view ment, well-being, independence, and acceptance (Cosco to identifying potential modifications that might make it et  al., 2013). This disparity is evidenced by a number of more relevant to the general population. Existing work studies indicating that many older people who consider aimed at understanding how the general population regard themselves to be aging successfully do not meet clinician/ successful aging has been primarily qualitative, considering researcher-defined criteria (McLaughlin, Jette, & Connell, responses to open-ended questions such as “How would Downloaded from https://academic.oup.com/psychsocgerontology/article/75/2/293/5033526 by DeepDyve user on 16 July 2022 Journals of Gerontology: PSYCHOLOGICAL SCIENCES, 2020, Vol. 75, No. 2 295 you define successful aging?” (Bowling, 2007; Cosco et al., and 7. Each wave carries a number of experiments based 2013), or asking participants to rank lists of researcher- on an annual competition and the current vignette experi- defined dimensions (Cosco et al., 2014b; Depp et al., 2007). ment was included in the 9th (IP9). At IP9 one-third of While results from these studies are useful, these approaches the sample was allocated to face-to-face interviewing and are not sufficiently systematic or robust to make inferences two-thirds to sequential mixed mode (households were first about the general population. An alternative, well-recog- offered a web interview and, if they did not take this up, nized approach is to use standardized vignettes (descrip- were then allocated a face-to-face interview) with mop- tions of a fictitious third party) in which factors used in the up interviews carried out by telephone. Respondents are description are randomized to assess their relative impact given a financial incentive to thank them for taking part. on individuals’ responses, independent of their own char- Ethics review is conducted by the University of Essex Ethics acteristics (Auspurg & Hinz, 2015). In our experiment, Committee. Full details of the design and experiments in participants in a large U.K. population sample were asked IP9 can be found in (Jäckle et al., 2017). to rate the successful aging of a (hypothetical) third party. Vignettes were based on the six successful aging dimen- This approach has not, to our knowledge, been used previ- sions each with two possible outcomes (favorable vs unfa- ously in this context and provides a unique, unconfounded, vorable), resulting in a total of 2  = 64 possible vignettes. empirical assessment of the relative importance of different Each respondent was presented with a set of three vignettes dimensions of successful aging to the general population. In to allow comparison while avoiding the task becom- addition, very few existing studies consider how views of ing tedious or arduous. A  2 factorial design was used to successful aging differ between men and women or younger randomly (without replacement) generate these vignettes, and older people (Charbonneau-Lyons, Mosher-Ashley, & ensuring that all combinations of favorable/unfavorable Stanford-Pollock, 2002; Collings, 2001; Cosco et al., 2015; dimensions were equally represented across all respondent Jopp et al., 2015) and there is therefore very limited infor- characteristics (Auspurg & Hinz, 2015). In addition, the mation about wider societal attitudes to aging and older randomization was designed to ensure that each respond- people, which is likely to have substantial influence on ent was presented with at least one male and one female policy discussions. There is also evidence that individu- vignette. Each vignette described a 75-year old with favora- als’ attitudes to aging change as they grow older (Phelan ble/unfavorable outcomes for each of the six dimensions. et  al., 2004; Tate, Swift, & Bayomi, 2013), although the The vignettes aimed to use neutral language, e.g., linking nature of these age-related changes is not well understood. word “and” rather than “but,” to avoid directing responses. Our study population includes respondents aged 16 years Definitions of favorable and unfavorable dimensions are and over, allowing exploration of perceptions of successful presented in Table 1. These definitions were based on spe- aging throughout the lifecourse and according to respond- cific rather than general conditions and limitations to main- ent characteristics. Existing evidence in this regard is very tain realism and engagement with the exercise, e.g., focusing limited. However, evidence from the medical sociological on “has difficulties climbing stairs,” rather than the broader literature on lay concepts of health (Blaxter, 1990) suggests “has a disability.” In addition, they were chosen to be eas- that, for example, older people might be more likely to pri- ily recognized, understood, and realistic in the context of oritize functioning while younger individuals might focus aging, e.g., considering productive engagement in terms of on disease and that men might focus on physical aspects volunteering rather than paid employment. Finally, defini- of disease while women will be more concerned with tions aimed to be similar in terms of severity and open to social factors. Our research aims were to gain a greater interpretation in terms of their potential impact on success- understanding of societal views of successful aging by: (a) ful aging. For example, diabetes was chosen as the chronic determining the relative importance placed by the general disease of interest as it is a leading cause of morbidity but population on the six Rowe-Kahn dimensions of success- can be successfully managed, whereas cancer might be ful aging and (b) understanding how perceptions of aging regarded as more likely to be terminal and therefore more vary according to respondent characteristics such as age severe. After each vignette, respondents were asked “How and gender. successfully is [Name] aging?,” giving a score from 0 (not successfully) to 10 (very successfully). An example set of vignettes is shown in Figure 1 along with the introductory Methods text presented to respondents. The Understanding Society Innovation Panel (IP) (Jäckle, Data from the experiment were analyzed using Gaia, Al Baghal, Burton, & Lynn, 2017) is a stratified, standard methods (Auspurg & Hinz, 2015). The relative geographically clustered sample of postcode sectors in importance of each vignette dimension in determining the Great Britain (south of the Caledonian Canal) with ran- successful aging score was assessed by comparing scores dom selection of addresses within each sampled sector. It for all vignettes in which the dimension was favorable is designed to be representative of the British population. with scores for all vignettes in which it was unfavor- All household members over 16  years are invited to take able, regardless of the values of the other dimensions. part annually with refreshment samples added at waves 4 Although, in the context of successful randomization, a Downloaded from https://academic.oup.com/psychsocgerontology/article/75/2/293/5033526 by DeepDyve user on 16 July 2022 296 Journals of Gerontology: PSYCHOLOGICAL SCIENCES, 2020, Vol. 75, No. 2 Table 1. Favorable and Unfavorable Rowe-Kahn Successful Aging Dimensions Used in the Vignettes Successful aging dimension Favorable Unfavorable Details Disease No long-term illness Diabetes Diabetes is a common disease of old age that is well known, doesn’t typically affect physical functioning, and avoids the potential life-limiting connotations of, for example, cancer or heart disease. Disability No difficulties climbing Difficulties climbing Difficulties with stairs is included in many health and stairs stairs disability scales, e.g., SF-36, Lambeth Diasability Screening Questionnaire, OECD Long-term Disability Questionnaire (McDowell, 2006). Physical functioning Opens food packages Struggles to open food Problems with opening food packaging is included in several easily packaging functional status scales, e.g., Functional Status Index, Stanford Health Assessment Questionnaire (McDowell, 2006). Cognitive functioning No problems Problems remembering Memory forms an integral part of many cognitive tests and, remembering in the context of aging, loss of memory is a prominent feature of dementia. Interpersonal engagement Regularly sees friends Rarely sees friends and Frequency of contact with family and friends is commonly and family family used in social health scales, e.g., RAND Social Health Battery, Katz Adjustment Scale (McDowell, 2006). Productive engagement Often volunteers Doesn’t volunteer Volunteering is a common form of productive engagement in the age group covered by the vignette, who are gener- ally past retirement age. in Supplementary Table  2, results from these regression models were very similar to those based on a simple com- parison of means. Coefficients from the regression models for each dimension of interest measure the difference be- tween the mean successful aging score across vignettes in which the dimension was favorable and the mean score across vignettes in which the dimension was unfavorable, with appropriate adjustments for the other dimensions and the study design. For example, the coefficient for (absence of) disease represents the (adjusted) difference between the mean score of all vignettes in which the in- dividual was described as having no long-term illness and the mean score of all vignettes in which the individual was Figure  1. Introductory text and example vignettes as presented to described as having diabetes. As each successful aging di- respondents. mension was presented in the same way (favorable versus unfavorable), it is valid to make direct comparisons be- simple comparison of means can be used, it is more usual tween them (Auspurg & Hinz, 2015) and the outcome (Auspurg & Hinz, 2015) to employ a multivariable (least measures from the models therefore represent the relative squares) regression model in which all vignette dimen- importance of each favorable dimension in determining sions are included simultaneously as independent binary the successful aging score. Formal comparisons of the (favorable versus unfavorable) predictors of the success- relative importance of different dimensions were made ful aging score. Moreover, when, as here, respondents post-estimation by considering linear combinations of re- are presented with multiple vignettes, random effects gression coefficients (e.g., β – β ). disease disability models are used to account for the hierarchical nature Analyses were repeated stratified by respondent gender, of the data (vignettes clustered within respondents) and age group, long-standing illness, marital status, employ- the order in which vignettes are presented. In addition, in ment status, financial difficulties, satisfaction with health, view of the survey design, the current analyses were also satisfaction with income, satisfaction with leisure time, sat- adjusted for sample and data collection mode and robust isfaction with life, and by vignette gender to explore what standard errors were calculated to allow for clustering impact these factors had on the relative importance attrib- within households and postcode sectors. As demonstrated uted to each dimension. Formal statistical tests of effect Downloaded from https://academic.oup.com/psychsocgerontology/article/75/2/293/5033526 by DeepDyve user on 16 July 2022 Journals of Gerontology: PSYCHOLOGICAL SCIENCES, 2020, Vol. 75, No. 2 297 modification by these factors were carried out by includ- favorable dimensions in each vignette varied from none to ing appropriate interaction terms in the regression models. six in approximately equal proportions. The success of the The six age groups included five younger than the person randomization is demonstrated by the similarities in the described in the vignette (<35, 35–44, 45–54, 55–64, percentage of positive dimensions across all respondent 65–74), representing those anticipating the scenario with characteristics. In addition, favorable dimensions were ap- varying proximity, and one the same age or older (75+), proximately equally allocated across vignettes describing considering the scenario concurrently or in retrospect. men and women. The scores given to the vignettes are sum- In sensitivity analyses, all analyses were repeated using marized in Figure  2 along with the range of scores given a subgroup of respondents for whom inverse probability by each respondent across the three vignettes (i.e., the dif- weights were available. These weights are calculated by the ference between the highest and lowest scoring vignette Understanding Society Team to adjust for differential non- presented to the individual respondent). Individual vignette response, unequal selection probabilities, and differential scores ranged from 0 to 10 and were somewhat skewed sampling error so that findings from the Innovation Panel toward the upper (more successful) end of the range with can be generalizable to the British population (Jäckle et al., a mean (standard deviation [SD]) score across all vignettes 2017). Analyses using these weights were very similar to of 6.2 (2.3). The range of scores given by each respondent those presented here. An outline of the design and analysis across the three vignettes also varied from 0 to 10, with a was prepared and approved before data collection and is mean (SD) range of 2.8 (2.1). The good spread of vignette held by the Understanding Society Team. scores and respondent ranges indicate that respondents distinguished between the vignettes and did not simply al- locate an average score to them all. In general, there was Results little evidence of systematic differences in mean scores according respondent characteristics and vignette gender A total of 1,508 eligible households were invited to par- (Supplementary Table 1) although there was some evidence ticipate in IP9 and 1,277 (85%) did so (Supplementary to suggest that, overall, women allocated somewhat higher Figure  1). Within participating households, there were scores than men (mean [SD] score: 6.4 (2.3) vs 6.0 [2.3]) 2,545 eligible adults, 2,143 (84%) of whom took part in and that scores decreased slightly with respondent age (e.g., either web (N  =  1,123) or face-to-face (N  =  1,020) inter- 6.3 [2.2] vs 5.9 [2.4] in <35 vs 75+ year olds, respectively). views; an additional 31 respondents had telephone inter- The importance given to each of the successful aging views. Of those interviewed via the web or face-to-face, dimensions, based on coefficients from regression models, 2,010 (94%) took part (unaided) in the self-complete sec- is presented in Figure 3. Numbers giving rise to this figure tion, which contained the vignettes. The ages of those who are presented in Supplementary Table 2 along with stand- took part ranged from 16 to 93  years. Characteristics of ardized effect sizes. Vignettes in which a particular dimen- the respondents who were presented with the vignettes are sion was favorable were consistently allocated higher scores presented in Supplementary Table 1. In total, 1,986 (99%) than those in which the same dimension was unfavorable, gave a score to all three and 24 (1%) to two or fewer. All scored vignettes were included in the analyses, giving a total of 5,967 completed overall. As would be expected from the design of the experiment, approximately half of all dimensions were favorable and the total of number of Figure 3. Relative importance of dimensions in determining successful aging score (based on difference (95% confidence interval) in mean suc- Figure 2. Vignette scores given in response to question “How success- cessful aging score from regression model comparing vignettes with fully is [Name] aging?” (N = 5,967) and range of scores given by each favorable versus unfavorable dimensions) for all respondents com- respondent (N = 2,010). bined plus, separately, male and female respondents. Downloaded from https://academic.oup.com/psychsocgerontology/article/75/2/293/5033526 by DeepDyve user on 16 July 2022 298 Journals of Gerontology: PSYCHOLOGICAL SCIENCES, 2020, Vol. 75, No. 2 with confidence intervals for the difference in mean scores excluding 0 (representing no impact of the dimension on successful aging scores) in every case. However, the rela- tive importance of the dimensions varied. Differences in the weights given to the different successful aging dimen- sions are presented in Supplementary Table 3. The dimen- sions given the greatest importance by respondents were cognitive function and disability; vignettes in which these dimensions were favorable were allocated successful aging scores that were 1.20 (95% confidence interval [CI]: 1.11, 1.30) and 1.18 (1.08, 1.27) points respectively higher than those in which the dimensions were unfavorable, with iden- tical corresponding standardized effect sizes of 0.56 (0.51, 0.61). Interpersonal engagement was also given relatively high importance (difference in mean scores: 0.99 [0.89, 1.08]; standardized effect size: 0.47 [0.42, 0.52]), although lower than disability and cognitive function (p for differ- ence with cognitive function < .001). Disease and physical function were given similar importance overall (difference in mean scores: 0.73 [0.64, 0.82] and 0.81 [0.73, 0.90]; standardized effect size: 0.32 [0.27, 0.37] and 0.37 [0.32, 0.42], respectively) and, again, this was markedly lower than disability, cognitive function and interpersonal engage- Figure 4. Relative importance of dimensions in determining successful ment (e.g., p for difference between disease and disability aging score (based on difference in mean successful aging score from < .001). The dimension given least weight was productive regression model comparing vignettes with favorable versus unfavora- engagement (difference in mean scores: 0.58 [0.49, 0.66]; ble dimensions) by age group. In each panel, differences are presented standardized effect size: 0.27 [0.22, 0.32], p for difference for all six successful aging dimensionswith differences (95% CI) for the with other dimensions < .02). dimension of interest in bold. Responses to vignettes were consistent across vignette gender, and the majority of respondent characteristics (Supplementary Table 4). However, there was some evidence dimensions, disease was generally given low importance that women gave more importance to productive engage- and this fell slightly, but not markedly, with increasing age ment than men (difference in mean scores for women and so that mean differences between favorable and unfavor- men: 0.70 [0.58, 0.81] vs 0.43 [0.30, 0.55], respectively; p able vignettes in the oldest age groups (65–74 and 75+) for interaction with gender  =  .002), although productive were the smallest overall (difference in mean scores: 0.71 engagement remained the dimension given least impor- [0.47, 0.95] and 0.58 [0.23, 0.93], respectively; p for inter- tance by both genders. In addition, there was a suggestion action with age group = .23). In contrast, disability was one that respondents who were married or living with a partner of the dimensions given the greatest importance at almost gave somewhat less importance to (absence of) disease than all ages, and the most important among 65–74-year olds those living alone (difference in mean scores: 0.64 [0.52, (difference in mean scores: 1.42 [1.19, 1.66]), although it 0.76] vs 0.85 [0.72, 0.99], respectively; p = .02). Responses was given somewhat less weight in 75+ year olds (differ- also differed somewhat between respondents who were ence in mean scores: 0.85 [0.49, 1.22]; p for interaction retired versus those who were employed/unemployed but with age group  =  .39). Physical function was given rela- these differences were due to variation in responses by age tively low weight by younger age groups (e.g., difference in and results for those who were employed and unemployed mean scores among <35-year olds: 0.66 [0.47, 0.84]) but were very similar. this increased with age, rising to one of the most important There were marked variations in the relative import- dimensions in 75+ year olds (difference in mean scores: ance attributed to different dimensions by respondents of 1.20 [0.88, 1.52]; p for interaction with age group = .003). different ages, particularly for physical and cognitive func- Cognitive function was consistently given high importance tion and productive engagement, as shown in Figure 4. In relative to other dimensions, particularly in those aged 45+, each panel, results are presented for all six successful aging and in 45–54, 55–64, and 75+ year olds was the most im- dimensions, with differences (95% CI) in mean successful portant overall (e.g., difference in mean scores in 75+ year aging scores between favorable and unfavorable vignettes olds: 1.39 [1.02, 1.76]; p for interaction with age group for the dimension of interest in bold. Numbers giving rise < .001). Interpersonal engagement was consistently in the to this figure are presented in Supplementary Table  4. middle of the dimensions in terms of importance (e.g., As previously observed in Figure  3, relative to other difference in mean scores in 45–54-year olds: 1.03 [0.81, Downloaded from https://academic.oup.com/psychsocgerontology/article/75/2/293/5033526 by DeepDyve user on 16 July 2022 Journals of Gerontology: PSYCHOLOGICAL SCIENCES, 2020, Vol. 75, No. 2 299 1.25]) and there was little evidence that this varied with them where appropriate. However, it is of note that results respondent age (p for interaction with age group  =  .61). from analyses weighted to be representative of the British Finally, although productive engagement increased slightly population were very similar to those presented here. The in importance with age, overall it was given less weight wide age range of respondents is a major strength of the ex- than the other dimensions and, in respondents aged less periment. However, in spite of the large sample size, it was than 65  years, differences in mean scores comparing fa- necessary to base age-stratified analyses on six age groups, vorable and unfavorable productive engagement were the the youngest including 16–34 years and the oldest 75–93- smallest overall (e.g., difference in means scores in less than year olds. These two age groups span almost 20 years each 35-year olds: 0.43 [0.25, 0.62]; p for interaction with age and there may be age-related differences within them that group = .01). are not captured in these analyses. Future work might focus on narrower age bands but this would require sub- stantially larger numbers of participants. It is also possible Discussion that the relative importance given to different dimensions was influenced by the success with which the definitions Successful aging scores given to the vignettes covered the captured them. Definitions were based on common fac- full range of possibilities and there was variation in the tors from existing, validated scales and were chosen to be scores allocated across the three vignettes presented to easily recognized, understandable, and relevant to older each respondent, indicating that respondents distinguished individuals. In addition, the perceived severity of the defini- between the different scenarios. Scores were consistently tions may have impacted on the results; for example a more higher for vignettes describing dimensions in favorable life-limiting disease e.g., cancer, or a more severe disability, rather than unfavorable terms although the relative impor- e.g., being in a wheelchair, might have been given greater tance of each dimension varied. Disease (presence/absence importance than those described here. However, vignette of diabetes) was one of the dimensions given least weight definitions were selected to be similar in terms of their (lim- in this experiment and the weight decreased with increas- ited) impact on activities of daily living. It is also worth ing age so that, among respondents aged 65+, disease was noting that the relative importance given to the different regarded as the least important overall. Productive engage- dimensions in the present study are broadly consistent with ment (volunteering) was also consistently less important existing literature. For example, a recent review of qualita- than other dimensions, particularly among men, although tive studies highlights the greater emphasis placed on psy- scores increased at older ages. In contrast, disability (dif- chosocial factors compared with physical health by older ficulties climbing stairs) and cognitive function (problems people (Cosco et al., 2013). Finally, many successful aging remembering) were given the greatest importance at all definitions, including the Rowe-Kahn model, have been ages, with the exception of a drop in the disability weight criticized for not going far enough in capturing the priori- among those aged 75+. Physical function (difficulties open- ties of older people, for example, well-being and autonomy ing food packaging) was given relatively low weight by (Ferri et al., 2009; Martinson & Berridge, 2015; Montross younger respondents but increased in importance in those et  al., 2006; Young et  al., 2009). Although it would have aged 65+. Interpersonal engagement (meeting family and been possible to include other dimensions such as these in friends regularly) was consistently weighted in the middle. our vignettes, this would have substantially increased the The relative importance given to the different dimensions number required. In addition, our aim in this experiment were generally consistent across respondent characteris- was to specifically understand societal attitudes to the most tics other than age, and the gender of the vignette had no commonly employed existing model of successful aging impact on the results. rather than create a new one that incorporates additional Existing work considering societal attitudes toward suc- dimensions. cessful aging has largely focused on qualitative responses to The value of patient preferences in directing clinical open-ended questions (Bowling, 2007; Cosco et al., 2013). practice is well established but practitioners’ views may dif- The use of vignettes, in which respondents consider a ficti- fer from those of their patients and this may be a particular tious third party, encourages individuals to consider success- problem in the context of aging as older patients are often ful aging as a broad hypothetical concept rather than asking reluctant to share their views (Joseph-Williams et al., 2017). whether they themselves are aging successfully. In addition, The majority of vignettes in the current study described an although respondents’ circumstances may influence their individual with at least one unfavorable dimension who, responses to vignettes the randomization of dimensions according to standard definitions, would be considered not across vignettes ensures a balanced design, meaning that to be aging successfully. However, the mean score across all potential biases and confounding arising from differences in vignettes was well above the midpoint of the scale (toward individual circumstances are eliminated. However, the ex- “aging successfully”), suggesting that the general popula- periment also has some limitations. The Innovation Panel is tion have a positive view of aging, even in the context of a household survey and individuals living in institutions are disease, disability, or limitations of functioning and social not included, although if individuals from previous waves engagement. The importance given to different successful move into an institution attempts are still made to interview Downloaded from https://academic.oup.com/psychsocgerontology/article/75/2/293/5033526 by DeepDyve user on 16 July 2022 300 Journals of Gerontology: PSYCHOLOGICAL SCIENCES, 2020, Vol. 75, No. 2 aging dimensions was largely independent of respondents’ importance given to interpersonal engagement was very circumstances, e.g., there was no difference in the weights similar across all age groups, underlining the ubiquity of given to disease and disability among those with and with- this dimension throughout the lifecourse and highlighting out LSI and, similarly, no difference in the importance given the need for health and social care services to, not only to productive engagement in those who were employed or treat disease and poor functioning, but also create oppor- not. A  specific criticism of the Rowe-Kahn model is that tunities for social interaction. Moreover, it is important to it reinforces social inequalities by defining successful aging recognize that attitudes among older people may continue as a state more easily achieved by those with higher socio- to change as they age and that patient-centered care is an economic position (Katz & Calasanti, 2015; Martinson evolving process. For example, there was a sharp drop in & Berridge, 2015; Stowe & Cooney, 2015). However, our the importance given to disability in the 75+ group, fol- stratified analyses suggest that attitudes toward success- lowing steady rises at younger ages. This is an isolated ful aging are not socially patterned, with almost identical finding and could be due to chance but could also reflect results for those with and without financial difficulties and shifting attitudes toward disability in an age group who those who were satisfied or dissatisfied with their income. were “living the vignette” and, perhaps, beginning to expe- In addition, despite different experiences throughout the rience, and therefore recognize, physical decline. Clinicians life course, results were largely consistent across vignette wishing to base their practice on shared decision making and respondents’ gender, although women gave somewhat and patient-centered care should recognize potential dif- more weight to volunteering, consistent with previous ferences between their own priorities for successful aging sociological work (Blaxter, 1990), perhaps reflecting trad- and those of their patients and, while there is no substitute itional gender roles in this regard. In terms of individual for face-to-face discussion with patients, our results pro- successful aging dimensions, the consistently low import- vide guidelines as to how these may differ. ance given to disease reinforces qualitative findings in Results from this study support and extend existing older individuals (Cosco et  al., 2013), and extends these work, providing unconfounded estimates of the relative to younger ages. An isolated, possibly chance, finding from importance given to six successful aging dimensions by the current analysis suggests that respondents living alone a large U.K.  population sample and demonstrating how gave more weight to disease than those in relationships, these vary across the lifecourse. However, it is not clear possibly reflecting greater perceived vulnerability in this whether wider societal policies such as health and social group, although disease remained among the dimensions care or pension provision influence these results and it regarded as least important. In contrast, other biomedical would be of considerable interest to repeat this experi- dimensions, such as disability and cognitive function, along ment in other populations where these differ. In addition, with interpersonal social engagement, were given some of given changing attitudes with age, it would be beneficial the highest weights by respondents of all ages. Morbidity in to understand how major life events such as retirement or older age is regarded as an important factor in determining bereavement influence these results. As well as informing health, social and economic policies, but policy makers and clinicians and policy makers working directly with older clinicians should recognize the relatively low value placed people, our results are relevant to researchers interested on disease by the general population and acknowledge the in measuring successful aging and its determinants. The greater importance to individuals of good functioning and Rowe-Kahn definition of successful aging is a widely used social engagement. research tool with “success” traditionally defined as a Perhaps the most striking results presented here are dichotomy in which all six criteria are met  although, in those demonstrating how attitudes to successful aging vary practice, very few older people achieve this, despite con- with age. The majority of existing work on perceptions of sidering themselves to be aging well (McLaughlin et  al., successful aging has focused on older people, while many 2012; Montross et  al., 2006; Strawbridge et  al., 2002). researchers, clinicians and policy makers are younger than A more pragmatic approach has been proposed in which those under study. Understanding the views of younger the extent of success in aging is measured by summing the individuals and how these differ from those at older ages number of favorable dimensions (Bowling, 2007; Bowling has the potential to close the gap between the attitudes of & Iliffe, 2011; Whitley et al., 2016), and our results may clinicians and their patients and to promote shared deci- provide a more “cutting edge” approach (Gu et al., 2017) sion making. For example, both physical function and pro- in which favorable dimensions are weighted according to ductive engagement were viewed as relatively unimportant the priorities of the general population. Such a measure by those of working age (<65  years) but their weights would provide a more nuanced approach to success- increased among older respondents, consistent with previ- ful aging and acknowledges the importance of quality ous work on lay perceptions of health (Blaxter, 1990). This as well as quantity of life, consistent with challenges to highlights the potential for relatively common problems the notion of health as complete physical, mental and so- of older age, such as struggling with food packaging or cial well-being (Huber et al., 2011). The quality-adjusted lacking a meaningful role in society, to be dismissed by life year (QALY), which weights different health states those involved in the care of older people. In contrast, the according to patient preference, is a well-established Downloaded from https://academic.oup.com/psychsocgerontology/article/75/2/293/5033526 by DeepDyve user on 16 July 2022 Journals of Gerontology: PSYCHOLOGICAL SCIENCES, 2020, Vol. 75, No. 2 301 Baltes, M. M., & Carstensen, L. L. (1996). The process of suc- health outcome (Whitehead & Ali, 2010). A  similar cessful ageing. Ageing Society, 16, 397–422. doi:10.1017/ measure based on weights such as those presented here s0144686x00003603 might form the basis for a modified Rowe-Kahn model Blaxter, M. (1990). Health and lifestyles. New York: Routledge, (Rowe & Kahn, 2015) that better represents societal Chapman and Hall. attitudes toward successful aging and could be used to Bloom, D. E., Chatterji, S., Kowal, P., Lloyd-Sherlock, P., McKee, M., evaluate interventions and direct policy investments to Rechel, B.,…Smith, J. P. (2015). Macroeconomic implications of promote successful aging worldwide. population ageing and selected policy responses. Lancet (London, England), 385, 649–657. doi:10.1016/S0140-6736(14)61464-1 st Bowling, A. (2007). Aspirations for older age in the 21 cen- Supplementary Material tury: What is successful aging? International Journal of Supplementary data is available at The Journals of Aging & Human Development, 64, 263–297. doi:10.2190/ Gerontology, Series B: Psychological Sciences and Social L0K1-87W4-9R01-7127 Sciences online. Bowling, A., & Iliffe, S. (2011). Psychological approach to successful ageing predicts future quality of life in older adults. Health and Quality of Life Outcomes, 9, 13. doi:10.1186/1477-7525-9-13 Funding Charbonneau-Lyons, D. L., Mosher-Ashley, P. M., & Stanford- E. Whitley and F.  Popham are funded by the Medical Research Pollock, M. (2002). Opinions of college students and inde- Council (MC_UU_12017/13) and the Chief Scientist Office, Scottish pendent-living adults regarding successful aging. Educational Government (SPHSU 13). M. Benzeval is funded by the UK Economic Gerontology, 28, 823–833. doi:10.1080/03601270290099822 and Social Research Council (ES/K005146/1). The funders had no Christensen, K., Doblhammer, G., Rau, R., & Vaupel, J. W. (2009). role in study design, data collection and analysis, decision to publish, Ageing populations: the challenges ahead. Lancet, 374(9696), or preparation of the manuscript. 1196–1208. doi:10.1016/S0140-6736(09)61460-4 Christensen, K., Thinggaard, M., Oksuzyan, A., Steenstrup, T., Andersen-Ranberg, K., Jeune, B.,…Vaupel, J. W. (2013). Physical Acknowledgments and cognitive functioning of people older than 90  years: A  comparison of two Danish cohorts born 10  years apart. This paper makes use of data from the Understanding Society Innovation Panel. Understanding Society is an initiative funded by Lancet (London, England), 382, 1507–1513. doi:10.1016/ the Economic and Social Research Council and various Government S0140-6736(13)60777-1 Departments, with scientific leadership by the Institute for Social Collings, P. (2001). “If you got everything, it’s good enough”: and Economic Research, University of Essex, and survey delivery Perspectives on successful aging in a Canadian Inuit community. by NatCen Social Research and Kantar Public. The research data Journal of Cross-Cultural Gerontology, 16, 127–155. doi:10.10 are distributed by the UK Data Service (SN 6849). The authors are 23/A:1010698200870 grateful to Vittal Katikireddi for advice on common diseases and Commission of the European Communities. (2009). Dealing with limitations in older people, Tarek Al Baghal for advice on experi- the impact of an ageing population in the EU (2009 Ageing mental design and question wording, and John Gilchrist for graphic Report). Brussels: Commission of the European Communities. production. doi:10.2765/80301 Cosco, T. D., Lemsalu, L., Brehme, D. F., Grigoruta, N., Kaufmann, L. K., Meex, R.,…Brayne, C. (2015). Younger Europeans’ con- Author Contributions ceptualizations of successful aging. Journal of the American FP had the original idea for the study, which was developed with EW Geriatrics Society, 63, 609–611. doi:10.1111/jgs.13307 and MB. EW, with FP and MB, developed the vignettes and study Cosco, T. D., Prina, A. M., Perales, J., Stephan, B. C. M., & Brayne, design. EW analyzed the data and wrote the first draft of the paper. C. (2013). Lay perspectives of successful ageing: A  system- All authors have reviewed and approved the final version. EW is the atic review and meta-ethnography. BMJ Open, 3:e002710. guarantor. doi:10.1136/bmjopen-2013–002710 Cosco, T. D., Prina, A. M., Perales, J., Stephan, B. C., & Brayne, C. (2014a). Operational definitions of successful aging: A  sys- Conflict of Interest tematic review. International Psychogeriatrics, 26, 373–381. None reported. doi:10.1017/S1041610213002287 Cosco, T. D., Prina, A. M., Perales, J., Stephan, B. C. M., & Brayne, C. (2014b). Whose “successful ageing”? Lay- and researcher- References driven conceptualisations of ageing well. Eur J Psychiat, 28, 124–130. doi:10.4321/S0213-61632014000200006 Araújo, L., Ribeiro, O., Teixeira, L., & Paúl, C. (2016). Predicting Depp, C. A., Glatt, S. J., & Jeste, D. V. (2007). Recent advances in successful aging at one hundred years of age. Research on Aging, research on successful or healthy aging. Current Psychiatry 38, 689–709. doi:10.1177/0164027515603771 Reports, 9, 7–13. doi:10.1007/s11920-007-0003-0 Auspurg, K., & Hinz, T. (2015). Factorial survey experiments. Los Depp, C. A., & Jeste, D. V. (2006). Definitions and predictors of Angeles: SAGE Publications Inc. doi:10.4135/9781483398075 successful aging: A comprehensive review of larger quantitative Baltes, P. B., & Baltes, M. M. (1990). Successful aging: Perspectives studies. The American Journal of Geriatric Psychiatry, 14, 6–20. from the behavioral sciences. New York: Cambridge University doi:10.1097/01.JGP.0000192501.03069.bc Press. doi:10.1017/CBO9780511665684 Downloaded from https://academic.oup.com/psychsocgerontology/article/75/2/293/5033526 by DeepDyve user on 16 July 2022 302 Journals of Gerontology: PSYCHOLOGICAL SCIENCES, 2020, Vol. 75, No. 2 Ferri, C., James, I., & Pruchno, R. (2009). Successful aging: Definitions Morrow-Howell, N. (2010). Volunteering in later life: Research frontiers. The Journals of Gerontology, Series B: Psychological and subjective assessment according to older adults. Clinical Gerontologist, 32, 379–388. doi:10.1080/07317110802677302 Sciences and Social Sciences, 65, 461–469. doi:10.1093/geronb/ gbq024 Gu, D., Feng, Q., Sautter, J. M., Yang, F., Ma, L., & Zhen, Z. (2017). Concordance and discordance of self-rated and researcher- Nimrod, G., & Ben-Shem, I. (2015). Successful aging as a lifelong process. Educational Gerontology, 41, 814–824. doi:10.1080/0 measured successful aging: Subtypes and associated factors. The Journals of Gerontology, Series B: Psychological Sciences and 3601277.2015.1050904 Phelan, E. A., Anderson, L. A., LaCroix, A. Z., & Larson, E. B. (2004). Social Sciences, 72, 214–227. doi:10.1093/geronb/gbw143 Huber, M., Knottnerus, J. A., Green, L., van der Horst, H., Jadad, Older adults’ views of “successful aging”–how do they compare with researchers’ definitions? Journal of the American Geriatrics A. R., Kromhout, D.,…Smid, H. (2011). How should we define health? BMJ (Clinical research ed.), 343, d4163. doi:10.1136/ Society, 52, 211–216. doi:10.1111/j.1532-5415.2004.52056.x Pruchno, R. A. (2015). Successful aging: Contentious past, productive bmj.d4163 Jäckle, A., Gaia, A., Al Baghal, T., Burton, J., & Lynn, P. (2017). future. Gerontologist, 55, 1–4. doi:10.1093/geront/gnv002 Pruchno, R., & Carr, D. (2017). Successful aging 2.0: Resilience and Understanding Society – The UK Household Longitudinal Study, Innovation Panel, Waves 1–9, User Manual. Colchester: beyond. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 72, 201–203. doi:10.1093/geronb/ University of Essex. Jopp, D. S., Wozniak, D., Damarin, A. K., De Feo, M., Jung, S., & gbw214 Rowe, J. W., & Kahn, R. L. (1997). Successful aging. The Jeswani, S. (2015). How could lay perspectives on successful aging complement scientific theory? Findings from a U.S. And Gerontologist, 37, 433–440. doi:10.1093/geront/37.4.433 Rowe, J. W., & Kahn, R. L. (2015). Successful aging 2.0: Conceptual a German life-span sample. The Gerontologist, 55, 91–106. st doi:10.1093/geront/gnu059 expansions for the 21 Century. The Journals of Gerontology, Joseph-Williams, N., Lloyd, A., Edwards, A., Stobbart, L., Tomson, Series B: Psychological Sciences and Social Sciences, 70, 593– D., Macphail, S.,…Thomson, R. (2017). Implementing shared 596. doi:10.1093/geronb/gbv025 decision making in the NHS: Lessons from the MAGIC pro- Spijker, J., & MacInnes, J. (2013). Population ageing: The timebomb gramme. BMJ (Clinical research ed.), 357, j1744. doi:10.1136/ that isn’t? BMJ (Clinical research ed.), 347, f6598. doi:10.1136/ bmj.j1744 bmj.f6598 Katz, S., & Calasanti, T. (2015). Critical perspectives on success- Stowe, J. D., & Cooney, T. M. (2015). Examining Rowe and Kahn’s ful aging: Does it “appeal more than it illuminates”? The concept of successful aging: Importance of taking a life course Gerontologist, 55, 26–33. doi:10.1093/geront/gnu027 perspective. The Gerontologist, 55, 43–50. doi:10.1093/geront/ Lupien, S. J., & Wan, N. (2004). Successful ageing: From cell to self. gnu055 Philosophical Transactions of the Royal Society of London, Strawbridge, W. J., Wallhagen, M. I., & Cohen, R. D. (2002). Series B: Biological Sciences, 359, 1413–1426. doi:10.1098/ Successful aging and well-being: Self-rated compared with rstb.2004.1516 Rowe and Kahn. The Gerontologist, 42, 727–733. doi:10.1093/ Manning, L. K., Carr, D. C., & Kail, B. L. (2016). Do higher lev- geront/42.6.727 els of resilience buffer the deleterious impact of chronic illness Tate, R. B., Swift, A. U., & Bayomi, D. J. (2013). Older men’s lay defi- on disability in later life? The Gerontologist, 56, 514–524. nitions of successful aging over time: The Manitoba follow-up doi:10.1093/geront/gnu068 study. International Journal of Aging & Human Development, Martin, P., Kelly, N., Kahana, B., Kahana, E., Willcox, B. J., Willcox, 76, 297–322. doi:10.2190/AG.76.4.b D. C., & Poon, L. W. (2015). Defining successful aging: A tangible United Nations. (2002). Report of the Second World Assembly on or elusive concept? The Gerontologist, 55, 14–25. doi:10.1093/ Ageing A/CONF.197/9. New York: United Nations. geront/gnu044 Vaupel, J. W. (2010). Biodemography of human ageing. Nature, 464, Martinson, M., & Berridge, C. (2015). Successful aging and its dis- 536–542. doi:10.1038/nature08984 contents: A  systematic review of the social gerontology litera- Whitehead, S. J., & Ali, S. (2010). Health outcomes in economic ture. The Gerontologist, 55, 58–69. doi:10.1093/geront/gnu037 evaluation: The QALY and utilities. British Medical Bulletin, 96, McDowell, I. (2006). Measuring health: A  guide to rating scales 5–21. doi:10.1093/bmb/ldq033 and questionnaires. New York: Oxford University Press. Whitley, E., Popham, F., & Benzeval, M. (2016). Comparison of the doi:10.1093/acprof:oso/9780195165678.001.0001 Rowe-Kahn model of successful aging with self-rated health and McLaughlin, S. J., Jette, A. M., & Connell, C. M. (2012). An examin- life satisfaction: The west of Scotland twenty-07 prospective ation of healthy aging across a conceptual continuum: Prevalence cohort study. The Gerontologist, 56, 1082–1092. doi:10.1093/ estimates, demographic patterns, and validity. The Journals of geront/gnv054 Gerontology, Series A: Biological Sciences and Medical Sciences, World Health Organisation. (2015). World report on ageing and 67, 783–789. doi:10.1093/gerona/glr234 health. Geneva, Switzerland: World Health Organisation. Montross, L. P., Depp, C., Daly, J., Reichstadt, J., Golshan, S., Moore, Young, Y., Frick, K. D., & Phelan, E. A. (2009). Can successful aging D.,…Jeste, D. V. (2006). Correlates of self-rated successful and chronic illness coexist in the same individual? A  multidi- aging among community-dwelling older adults. The American mensional concept of successful aging. Journal of the American Journal of Geriatric Psychiatry, 14, 43–51. doi:10.1097/01. Medical Directors Association, 10, 87–92. doi:10.1016/j. JGP.0000192489.43179.31 jamda.2008.11.003 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png "The Journals of Gerontology - Series B: Psychological Sciences and Social Sciences" Oxford University Press

Population Priorities for Successful Aging: A Randomized Vignette Experiment

Download PDF
10 pages

Loading next page...
 
/lp/ou_press/population-priorities-for-successful-aging-a-randomized-vignette-IB0BNjIwTG

References (55)

Publisher
Oxford University Press
Copyright
Copyright © 2022 The Gerontological Society of America
ISSN
1079-5014
eISSN
1758-5368
DOI
10.1093/geronb/gby060
Publisher site
See Article on Publisher Site

Abstract

Objectives: Aging populations have led to increasing interest in “successful aging” but there is no consensus as to what this entails. We aimed to understand the relative importance to the general population of six commonly-used successful aging dimensions (disease, disability, physical functioning, cognitive functioning, interpersonal engagement, and productive engagement). Method: Two thousand and ten British men and women were shown vignettes describing an older person with randomly determined favorable/unfavorable outcomes for each dimension and asked to score (0–10) how successfully the person was aging. Results: Vignettes with favorable successful aging dimensions were given higher mean scores than those with unfavorable dimensions. The dimensions given greatest importance were cognitive function (difference [95% confidence interval {CI}] in mean scores: 1.20 [1.11, 1.30]) and disability (1.18 [1.08, 1.27]), while disease (0.73 [0.64, 0.82]) and productive engage- ment (0.58 [0.49, 0.66]) were given the least importance. Older respondents gave increasingly greater relative importance to physical function, cognitive function, and productive engagement. Discussion: Successful aging definitions that focus on disease do not reflect the views of the population in general and older people in particular. Practitioners and policy makers should be aware of older people’s priorities for aging and understand how these differ from their own. Keywords: Attitudes, Cognition, Health, Interpersonal relations, Successful aging Industrialized populations are aging, (Christensen et  al., 2010) and are more likely to be in paid employment (Spijker 2009) prompting debate about whether growing propor- & MacInnes, 2013) or volunteering (Morrow-Howell, tions of older individuals require increasing investment in 2010), resulting in a growing interest in the notion of “suc- health and long-term care.(Bloom et  al., 2015) Early re- cessful aging” (Araújo, Ribeiro, Teixeira, & Paúl, 2016; search and policy often concentrated on more unfavorable Bowling, 2007; Katz & Calasanti, 2015; Martin et  al., aspects of aging, particularly at the population level, result- 2015; Martinson & Berridge, 2015; Nimrod & Ben-Shem, ing in anxiety and negativity about its potential impact on 2015; Stowe & Cooney, 2015). In addition, older people society (Baltes & Carstensen, 1996). However, more recent are often more positive about the aging process than those evidence suggests that, compared with their peers in pre- involved in their care, demonstrating high levels of adjust- vious cohorts, older people today have better physical and ment, acceptance, and resilience (Manning, Carr, & Kail, cognitive functioning (Christensen et  al., 2013; Vaupel, 2016). These views are consistent with recent challenges to © The Author(s) 2018. Published by Oxford University Press on behalf of The Gerontological Society of America. 293 This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. Downloaded from https://academic.oup.com/psychsocgerontology/article/75/2/293/5033526 by DeepDyve user on 16 July 2022 294 Journals of Gerontology: PSYCHOLOGICAL SCIENCES, 2020, Vol. 75, No. 2 the current World Health Organization (WHO) definition 2012; Montross et  al., 2006; Strawbridge, Wallhagen, & of health as “a state of complete physical, mental and social Cohen, 2002; Young, Frick, & Phelan, 2009). In spite of well-being,” which recommend an alternative formulation decades of research, there is still no firm consensus as to in terms of individuals’ ability to adapt and self-manage what successful aging entails, with recent special issues of (Huber et  al., 2011). However, there is a danger that the Journals of Gerontology: Social Sciences (Pruchno & Carr, attitudes of practitioners and policy makers involved with 2017) and The Gerontologist (Pruchno, 2015) devoted to older people are based on out-of-date and potentially mis- the question. In addition, the development of appropriate leading information and differences in the beliefs of older metrics has been identified as a research priority by WHO people and professionals are particularly pertinent in the (World Health Organisation, 2015). However, the most context of shared decision making and patient-centered widely adopted multidimensional model of successful aging care. Although the value of patients’ opinions in shap- was proposed by Rowe and Kahn (Rowe & Kahn, 1997) ing and informing clinical practice is well recognized in and incorporates six dimensions: (a) avoidance of disease; principle, recent results from the MAGIC (Making Good (b) avoidance of disability; (c) maintenance of good phys- Decisions in Collaboration) programme (Joseph-Williams ical function; (d) maintenance of good cognitive function; et al., 2017) highlight that, in practice, some clinicians “fail (e) good interpersonal social engagement (contacts and to recognize that patients’ values, opinions or preferences transactions with others); and (f) good productive engage- ... may differ from their own.” Moreover, the authors report ment (engagement in activities of value to society such as that older people may be particularly reluctant to share working or volunteering). Conventionally, according to this their views. If policy and practice are to support people to definition an individual is considered to be aging success- age successfully, a greater understanding of the extent to fully if they meet all six criteria. This straightforward char- which people value different aspects of aging is required. acterization moves beyond the biomedical to include social Clinicians, researchers, and policy makers worldwide and productive engagement, which have been shown to be agree that “successful aging” is an important goal (Bloom of substantial importance to older people (Bowling, 2007; et  al., 2015; Commission of the European Communities, Cosco, Prina, Perales, Stephan, & Brayne, 2013; Depp, 2009; United Nations, 2002) but its meaning remains un- Glatt, & Jeste, 2007), and positive associations have been clear. A  wide array of successful aging definitions have reported between this definition of successful aging and been proposed in the literature (Lupien & Wan, 2004). self-reported well-being (Strawbridge et al., 2002), health, Some focus on specific domains, for example, biomedical and life satisfaction (Whitley, Popham, & Benzeval, 2016) aging, covering compression of morbidity and genetic fac- in older people. However, the extent to which it reflects tors, while others consider cognitive or psychosocial aging, perceptions of successful aging in the general population with an emphasis on subjective well-being and personality. continues to be widely debated (Bowling & Iliffe, 2011; While these models provide insights into particular aging Ferri, James, & Pruchno, 2009; Martinson & Berridge, processes and are valuable in developing the specific poli- 2015; Montross et  al., 2006; Phelan, Anderson, LaCroix, cies that underpin them, they can also be limited in their & Larson, 2004; Stowe & Cooney, 2015; Strawbridge ability to predict or explain other aspects of aging and this et al., 2002; Young et al., 2009) and the relative importance has led to the development of multidimensional models of each dimension is unknown. Rowe and Kahn (Rowe & that include multiple aging dimensions.(Lupien & Wan, Kahn, 2015) have also entered this debate, acknowledging 2004) Again, many different multidimensional models have the limitations of their model but supporting the notion been proposed, some focusing on successful aging as an that “its extensive use in scientific enquiry warrants modi- adaptive process, such as the Selection, Optimization and fication over disposal.” In their discussion, they propose Compensation (SOC) model proposed by Baltes and Baltes new priorities for research, including the need to take a (Baltes & Baltes, 1990), and others focusing on successful lifecourse perspective to aging, to focus more on the poten- aging as measureable state, such as the MacArthur model tial benefits of an aging society, and to consider successful proposed by Rowe and Kahn (Rowe & Kahn, 1997). aging not only at the level of the individual but also at the There are also differences in multidimensional models level of society. proposed by different groups. For example, while the ma- In order to promote successful aging at the societal jority of operational definitions of successful aging include level, it is vital to understand what the general popula- physiological factors such as disease, disability, and phys- tion consider to be successful aging. Rather than propose ical function (Cosco et  al., 2014a; Depp & Jeste, 2006), another new successful aging model for additional debate, considerably fewer include dimensions known to be of we aim instead to understand population attitudes toward value to older people, such as functioning, social engage- the most commonly employed existing model with a view ment, well-being, independence, and acceptance (Cosco to identifying potential modifications that might make it et  al., 2013). This disparity is evidenced by a number of more relevant to the general population. Existing work studies indicating that many older people who consider aimed at understanding how the general population regard themselves to be aging successfully do not meet clinician/ successful aging has been primarily qualitative, considering researcher-defined criteria (McLaughlin, Jette, & Connell, responses to open-ended questions such as “How would Downloaded from https://academic.oup.com/psychsocgerontology/article/75/2/293/5033526 by DeepDyve user on 16 July 2022 Journals of Gerontology: PSYCHOLOGICAL SCIENCES, 2020, Vol. 75, No. 2 295 you define successful aging?” (Bowling, 2007; Cosco et al., and 7. Each wave carries a number of experiments based 2013), or asking participants to rank lists of researcher- on an annual competition and the current vignette experi- defined dimensions (Cosco et al., 2014b; Depp et al., 2007). ment was included in the 9th (IP9). At IP9 one-third of While results from these studies are useful, these approaches the sample was allocated to face-to-face interviewing and are not sufficiently systematic or robust to make inferences two-thirds to sequential mixed mode (households were first about the general population. An alternative, well-recog- offered a web interview and, if they did not take this up, nized approach is to use standardized vignettes (descrip- were then allocated a face-to-face interview) with mop- tions of a fictitious third party) in which factors used in the up interviews carried out by telephone. Respondents are description are randomized to assess their relative impact given a financial incentive to thank them for taking part. on individuals’ responses, independent of their own char- Ethics review is conducted by the University of Essex Ethics acteristics (Auspurg & Hinz, 2015). In our experiment, Committee. Full details of the design and experiments in participants in a large U.K. population sample were asked IP9 can be found in (Jäckle et al., 2017). to rate the successful aging of a (hypothetical) third party. Vignettes were based on the six successful aging dimen- This approach has not, to our knowledge, been used previ- sions each with two possible outcomes (favorable vs unfa- ously in this context and provides a unique, unconfounded, vorable), resulting in a total of 2  = 64 possible vignettes. empirical assessment of the relative importance of different Each respondent was presented with a set of three vignettes dimensions of successful aging to the general population. In to allow comparison while avoiding the task becom- addition, very few existing studies consider how views of ing tedious or arduous. A  2 factorial design was used to successful aging differ between men and women or younger randomly (without replacement) generate these vignettes, and older people (Charbonneau-Lyons, Mosher-Ashley, & ensuring that all combinations of favorable/unfavorable Stanford-Pollock, 2002; Collings, 2001; Cosco et al., 2015; dimensions were equally represented across all respondent Jopp et al., 2015) and there is therefore very limited infor- characteristics (Auspurg & Hinz, 2015). In addition, the mation about wider societal attitudes to aging and older randomization was designed to ensure that each respond- people, which is likely to have substantial influence on ent was presented with at least one male and one female policy discussions. There is also evidence that individu- vignette. Each vignette described a 75-year old with favora- als’ attitudes to aging change as they grow older (Phelan ble/unfavorable outcomes for each of the six dimensions. et  al., 2004; Tate, Swift, & Bayomi, 2013), although the The vignettes aimed to use neutral language, e.g., linking nature of these age-related changes is not well understood. word “and” rather than “but,” to avoid directing responses. Our study population includes respondents aged 16 years Definitions of favorable and unfavorable dimensions are and over, allowing exploration of perceptions of successful presented in Table 1. These definitions were based on spe- aging throughout the lifecourse and according to respond- cific rather than general conditions and limitations to main- ent characteristics. Existing evidence in this regard is very tain realism and engagement with the exercise, e.g., focusing limited. However, evidence from the medical sociological on “has difficulties climbing stairs,” rather than the broader literature on lay concepts of health (Blaxter, 1990) suggests “has a disability.” In addition, they were chosen to be eas- that, for example, older people might be more likely to pri- ily recognized, understood, and realistic in the context of oritize functioning while younger individuals might focus aging, e.g., considering productive engagement in terms of on disease and that men might focus on physical aspects volunteering rather than paid employment. Finally, defini- of disease while women will be more concerned with tions aimed to be similar in terms of severity and open to social factors. Our research aims were to gain a greater interpretation in terms of their potential impact on success- understanding of societal views of successful aging by: (a) ful aging. For example, diabetes was chosen as the chronic determining the relative importance placed by the general disease of interest as it is a leading cause of morbidity but population on the six Rowe-Kahn dimensions of success- can be successfully managed, whereas cancer might be ful aging and (b) understanding how perceptions of aging regarded as more likely to be terminal and therefore more vary according to respondent characteristics such as age severe. After each vignette, respondents were asked “How and gender. successfully is [Name] aging?,” giving a score from 0 (not successfully) to 10 (very successfully). An example set of vignettes is shown in Figure 1 along with the introductory Methods text presented to respondents. The Understanding Society Innovation Panel (IP) (Jäckle, Data from the experiment were analyzed using Gaia, Al Baghal, Burton, & Lynn, 2017) is a stratified, standard methods (Auspurg & Hinz, 2015). The relative geographically clustered sample of postcode sectors in importance of each vignette dimension in determining the Great Britain (south of the Caledonian Canal) with ran- successful aging score was assessed by comparing scores dom selection of addresses within each sampled sector. It for all vignettes in which the dimension was favorable is designed to be representative of the British population. with scores for all vignettes in which it was unfavor- All household members over 16  years are invited to take able, regardless of the values of the other dimensions. part annually with refreshment samples added at waves 4 Although, in the context of successful randomization, a Downloaded from https://academic.oup.com/psychsocgerontology/article/75/2/293/5033526 by DeepDyve user on 16 July 2022 296 Journals of Gerontology: PSYCHOLOGICAL SCIENCES, 2020, Vol. 75, No. 2 Table 1. Favorable and Unfavorable Rowe-Kahn Successful Aging Dimensions Used in the Vignettes Successful aging dimension Favorable Unfavorable Details Disease No long-term illness Diabetes Diabetes is a common disease of old age that is well known, doesn’t typically affect physical functioning, and avoids the potential life-limiting connotations of, for example, cancer or heart disease. Disability No difficulties climbing Difficulties climbing Difficulties with stairs is included in many health and stairs stairs disability scales, e.g., SF-36, Lambeth Diasability Screening Questionnaire, OECD Long-term Disability Questionnaire (McDowell, 2006). Physical functioning Opens food packages Struggles to open food Problems with opening food packaging is included in several easily packaging functional status scales, e.g., Functional Status Index, Stanford Health Assessment Questionnaire (McDowell, 2006). Cognitive functioning No problems Problems remembering Memory forms an integral part of many cognitive tests and, remembering in the context of aging, loss of memory is a prominent feature of dementia. Interpersonal engagement Regularly sees friends Rarely sees friends and Frequency of contact with family and friends is commonly and family family used in social health scales, e.g., RAND Social Health Battery, Katz Adjustment Scale (McDowell, 2006). Productive engagement Often volunteers Doesn’t volunteer Volunteering is a common form of productive engagement in the age group covered by the vignette, who are gener- ally past retirement age. in Supplementary Table  2, results from these regression models were very similar to those based on a simple com- parison of means. Coefficients from the regression models for each dimension of interest measure the difference be- tween the mean successful aging score across vignettes in which the dimension was favorable and the mean score across vignettes in which the dimension was unfavorable, with appropriate adjustments for the other dimensions and the study design. For example, the coefficient for (absence of) disease represents the (adjusted) difference between the mean score of all vignettes in which the in- dividual was described as having no long-term illness and the mean score of all vignettes in which the individual was Figure  1. Introductory text and example vignettes as presented to described as having diabetes. As each successful aging di- respondents. mension was presented in the same way (favorable versus unfavorable), it is valid to make direct comparisons be- simple comparison of means can be used, it is more usual tween them (Auspurg & Hinz, 2015) and the outcome (Auspurg & Hinz, 2015) to employ a multivariable (least measures from the models therefore represent the relative squares) regression model in which all vignette dimen- importance of each favorable dimension in determining sions are included simultaneously as independent binary the successful aging score. Formal comparisons of the (favorable versus unfavorable) predictors of the success- relative importance of different dimensions were made ful aging score. Moreover, when, as here, respondents post-estimation by considering linear combinations of re- are presented with multiple vignettes, random effects gression coefficients (e.g., β – β ). disease disability models are used to account for the hierarchical nature Analyses were repeated stratified by respondent gender, of the data (vignettes clustered within respondents) and age group, long-standing illness, marital status, employ- the order in which vignettes are presented. In addition, in ment status, financial difficulties, satisfaction with health, view of the survey design, the current analyses were also satisfaction with income, satisfaction with leisure time, sat- adjusted for sample and data collection mode and robust isfaction with life, and by vignette gender to explore what standard errors were calculated to allow for clustering impact these factors had on the relative importance attrib- within households and postcode sectors. As demonstrated uted to each dimension. Formal statistical tests of effect Downloaded from https://academic.oup.com/psychsocgerontology/article/75/2/293/5033526 by DeepDyve user on 16 July 2022 Journals of Gerontology: PSYCHOLOGICAL SCIENCES, 2020, Vol. 75, No. 2 297 modification by these factors were carried out by includ- favorable dimensions in each vignette varied from none to ing appropriate interaction terms in the regression models. six in approximately equal proportions. The success of the The six age groups included five younger than the person randomization is demonstrated by the similarities in the described in the vignette (<35, 35–44, 45–54, 55–64, percentage of positive dimensions across all respondent 65–74), representing those anticipating the scenario with characteristics. In addition, favorable dimensions were ap- varying proximity, and one the same age or older (75+), proximately equally allocated across vignettes describing considering the scenario concurrently or in retrospect. men and women. The scores given to the vignettes are sum- In sensitivity analyses, all analyses were repeated using marized in Figure  2 along with the range of scores given a subgroup of respondents for whom inverse probability by each respondent across the three vignettes (i.e., the dif- weights were available. These weights are calculated by the ference between the highest and lowest scoring vignette Understanding Society Team to adjust for differential non- presented to the individual respondent). Individual vignette response, unequal selection probabilities, and differential scores ranged from 0 to 10 and were somewhat skewed sampling error so that findings from the Innovation Panel toward the upper (more successful) end of the range with can be generalizable to the British population (Jäckle et al., a mean (standard deviation [SD]) score across all vignettes 2017). Analyses using these weights were very similar to of 6.2 (2.3). The range of scores given by each respondent those presented here. An outline of the design and analysis across the three vignettes also varied from 0 to 10, with a was prepared and approved before data collection and is mean (SD) range of 2.8 (2.1). The good spread of vignette held by the Understanding Society Team. scores and respondent ranges indicate that respondents distinguished between the vignettes and did not simply al- locate an average score to them all. In general, there was Results little evidence of systematic differences in mean scores according respondent characteristics and vignette gender A total of 1,508 eligible households were invited to par- (Supplementary Table 1) although there was some evidence ticipate in IP9 and 1,277 (85%) did so (Supplementary to suggest that, overall, women allocated somewhat higher Figure  1). Within participating households, there were scores than men (mean [SD] score: 6.4 (2.3) vs 6.0 [2.3]) 2,545 eligible adults, 2,143 (84%) of whom took part in and that scores decreased slightly with respondent age (e.g., either web (N  =  1,123) or face-to-face (N  =  1,020) inter- 6.3 [2.2] vs 5.9 [2.4] in <35 vs 75+ year olds, respectively). views; an additional 31 respondents had telephone inter- The importance given to each of the successful aging views. Of those interviewed via the web or face-to-face, dimensions, based on coefficients from regression models, 2,010 (94%) took part (unaided) in the self-complete sec- is presented in Figure 3. Numbers giving rise to this figure tion, which contained the vignettes. The ages of those who are presented in Supplementary Table 2 along with stand- took part ranged from 16 to 93  years. Characteristics of ardized effect sizes. Vignettes in which a particular dimen- the respondents who were presented with the vignettes are sion was favorable were consistently allocated higher scores presented in Supplementary Table 1. In total, 1,986 (99%) than those in which the same dimension was unfavorable, gave a score to all three and 24 (1%) to two or fewer. All scored vignettes were included in the analyses, giving a total of 5,967 completed overall. As would be expected from the design of the experiment, approximately half of all dimensions were favorable and the total of number of Figure 3. Relative importance of dimensions in determining successful aging score (based on difference (95% confidence interval) in mean suc- Figure 2. Vignette scores given in response to question “How success- cessful aging score from regression model comparing vignettes with fully is [Name] aging?” (N = 5,967) and range of scores given by each favorable versus unfavorable dimensions) for all respondents com- respondent (N = 2,010). bined plus, separately, male and female respondents. Downloaded from https://academic.oup.com/psychsocgerontology/article/75/2/293/5033526 by DeepDyve user on 16 July 2022 298 Journals of Gerontology: PSYCHOLOGICAL SCIENCES, 2020, Vol. 75, No. 2 with confidence intervals for the difference in mean scores excluding 0 (representing no impact of the dimension on successful aging scores) in every case. However, the rela- tive importance of the dimensions varied. Differences in the weights given to the different successful aging dimen- sions are presented in Supplementary Table 3. The dimen- sions given the greatest importance by respondents were cognitive function and disability; vignettes in which these dimensions were favorable were allocated successful aging scores that were 1.20 (95% confidence interval [CI]: 1.11, 1.30) and 1.18 (1.08, 1.27) points respectively higher than those in which the dimensions were unfavorable, with iden- tical corresponding standardized effect sizes of 0.56 (0.51, 0.61). Interpersonal engagement was also given relatively high importance (difference in mean scores: 0.99 [0.89, 1.08]; standardized effect size: 0.47 [0.42, 0.52]), although lower than disability and cognitive function (p for differ- ence with cognitive function < .001). Disease and physical function were given similar importance overall (difference in mean scores: 0.73 [0.64, 0.82] and 0.81 [0.73, 0.90]; standardized effect size: 0.32 [0.27, 0.37] and 0.37 [0.32, 0.42], respectively) and, again, this was markedly lower than disability, cognitive function and interpersonal engage- Figure 4. Relative importance of dimensions in determining successful ment (e.g., p for difference between disease and disability aging score (based on difference in mean successful aging score from < .001). The dimension given least weight was productive regression model comparing vignettes with favorable versus unfavora- engagement (difference in mean scores: 0.58 [0.49, 0.66]; ble dimensions) by age group. In each panel, differences are presented standardized effect size: 0.27 [0.22, 0.32], p for difference for all six successful aging dimensionswith differences (95% CI) for the with other dimensions < .02). dimension of interest in bold. Responses to vignettes were consistent across vignette gender, and the majority of respondent characteristics (Supplementary Table 4). However, there was some evidence dimensions, disease was generally given low importance that women gave more importance to productive engage- and this fell slightly, but not markedly, with increasing age ment than men (difference in mean scores for women and so that mean differences between favorable and unfavor- men: 0.70 [0.58, 0.81] vs 0.43 [0.30, 0.55], respectively; p able vignettes in the oldest age groups (65–74 and 75+) for interaction with gender  =  .002), although productive were the smallest overall (difference in mean scores: 0.71 engagement remained the dimension given least impor- [0.47, 0.95] and 0.58 [0.23, 0.93], respectively; p for inter- tance by both genders. In addition, there was a suggestion action with age group = .23). In contrast, disability was one that respondents who were married or living with a partner of the dimensions given the greatest importance at almost gave somewhat less importance to (absence of) disease than all ages, and the most important among 65–74-year olds those living alone (difference in mean scores: 0.64 [0.52, (difference in mean scores: 1.42 [1.19, 1.66]), although it 0.76] vs 0.85 [0.72, 0.99], respectively; p = .02). Responses was given somewhat less weight in 75+ year olds (differ- also differed somewhat between respondents who were ence in mean scores: 0.85 [0.49, 1.22]; p for interaction retired versus those who were employed/unemployed but with age group  =  .39). Physical function was given rela- these differences were due to variation in responses by age tively low weight by younger age groups (e.g., difference in and results for those who were employed and unemployed mean scores among <35-year olds: 0.66 [0.47, 0.84]) but were very similar. this increased with age, rising to one of the most important There were marked variations in the relative import- dimensions in 75+ year olds (difference in mean scores: ance attributed to different dimensions by respondents of 1.20 [0.88, 1.52]; p for interaction with age group = .003). different ages, particularly for physical and cognitive func- Cognitive function was consistently given high importance tion and productive engagement, as shown in Figure 4. In relative to other dimensions, particularly in those aged 45+, each panel, results are presented for all six successful aging and in 45–54, 55–64, and 75+ year olds was the most im- dimensions, with differences (95% CI) in mean successful portant overall (e.g., difference in mean scores in 75+ year aging scores between favorable and unfavorable vignettes olds: 1.39 [1.02, 1.76]; p for interaction with age group for the dimension of interest in bold. Numbers giving rise < .001). Interpersonal engagement was consistently in the to this figure are presented in Supplementary Table  4. middle of the dimensions in terms of importance (e.g., As previously observed in Figure  3, relative to other difference in mean scores in 45–54-year olds: 1.03 [0.81, Downloaded from https://academic.oup.com/psychsocgerontology/article/75/2/293/5033526 by DeepDyve user on 16 July 2022 Journals of Gerontology: PSYCHOLOGICAL SCIENCES, 2020, Vol. 75, No. 2 299 1.25]) and there was little evidence that this varied with them where appropriate. However, it is of note that results respondent age (p for interaction with age group  =  .61). from analyses weighted to be representative of the British Finally, although productive engagement increased slightly population were very similar to those presented here. The in importance with age, overall it was given less weight wide age range of respondents is a major strength of the ex- than the other dimensions and, in respondents aged less periment. However, in spite of the large sample size, it was than 65  years, differences in mean scores comparing fa- necessary to base age-stratified analyses on six age groups, vorable and unfavorable productive engagement were the the youngest including 16–34 years and the oldest 75–93- smallest overall (e.g., difference in means scores in less than year olds. These two age groups span almost 20 years each 35-year olds: 0.43 [0.25, 0.62]; p for interaction with age and there may be age-related differences within them that group = .01). are not captured in these analyses. Future work might focus on narrower age bands but this would require sub- stantially larger numbers of participants. It is also possible Discussion that the relative importance given to different dimensions was influenced by the success with which the definitions Successful aging scores given to the vignettes covered the captured them. Definitions were based on common fac- full range of possibilities and there was variation in the tors from existing, validated scales and were chosen to be scores allocated across the three vignettes presented to easily recognized, understandable, and relevant to older each respondent, indicating that respondents distinguished individuals. In addition, the perceived severity of the defini- between the different scenarios. Scores were consistently tions may have impacted on the results; for example a more higher for vignettes describing dimensions in favorable life-limiting disease e.g., cancer, or a more severe disability, rather than unfavorable terms although the relative impor- e.g., being in a wheelchair, might have been given greater tance of each dimension varied. Disease (presence/absence importance than those described here. However, vignette of diabetes) was one of the dimensions given least weight definitions were selected to be similar in terms of their (lim- in this experiment and the weight decreased with increas- ited) impact on activities of daily living. It is also worth ing age so that, among respondents aged 65+, disease was noting that the relative importance given to the different regarded as the least important overall. Productive engage- dimensions in the present study are broadly consistent with ment (volunteering) was also consistently less important existing literature. For example, a recent review of qualita- than other dimensions, particularly among men, although tive studies highlights the greater emphasis placed on psy- scores increased at older ages. In contrast, disability (dif- chosocial factors compared with physical health by older ficulties climbing stairs) and cognitive function (problems people (Cosco et al., 2013). Finally, many successful aging remembering) were given the greatest importance at all definitions, including the Rowe-Kahn model, have been ages, with the exception of a drop in the disability weight criticized for not going far enough in capturing the priori- among those aged 75+. Physical function (difficulties open- ties of older people, for example, well-being and autonomy ing food packaging) was given relatively low weight by (Ferri et al., 2009; Martinson & Berridge, 2015; Montross younger respondents but increased in importance in those et  al., 2006; Young et  al., 2009). Although it would have aged 65+. Interpersonal engagement (meeting family and been possible to include other dimensions such as these in friends regularly) was consistently weighted in the middle. our vignettes, this would have substantially increased the The relative importance given to the different dimensions number required. In addition, our aim in this experiment were generally consistent across respondent characteris- was to specifically understand societal attitudes to the most tics other than age, and the gender of the vignette had no commonly employed existing model of successful aging impact on the results. rather than create a new one that incorporates additional Existing work considering societal attitudes toward suc- dimensions. cessful aging has largely focused on qualitative responses to The value of patient preferences in directing clinical open-ended questions (Bowling, 2007; Cosco et al., 2013). practice is well established but practitioners’ views may dif- The use of vignettes, in which respondents consider a ficti- fer from those of their patients and this may be a particular tious third party, encourages individuals to consider success- problem in the context of aging as older patients are often ful aging as a broad hypothetical concept rather than asking reluctant to share their views (Joseph-Williams et al., 2017). whether they themselves are aging successfully. In addition, The majority of vignettes in the current study described an although respondents’ circumstances may influence their individual with at least one unfavorable dimension who, responses to vignettes the randomization of dimensions according to standard definitions, would be considered not across vignettes ensures a balanced design, meaning that to be aging successfully. However, the mean score across all potential biases and confounding arising from differences in vignettes was well above the midpoint of the scale (toward individual circumstances are eliminated. However, the ex- “aging successfully”), suggesting that the general popula- periment also has some limitations. The Innovation Panel is tion have a positive view of aging, even in the context of a household survey and individuals living in institutions are disease, disability, or limitations of functioning and social not included, although if individuals from previous waves engagement. The importance given to different successful move into an institution attempts are still made to interview Downloaded from https://academic.oup.com/psychsocgerontology/article/75/2/293/5033526 by DeepDyve user on 16 July 2022 300 Journals of Gerontology: PSYCHOLOGICAL SCIENCES, 2020, Vol. 75, No. 2 aging dimensions was largely independent of respondents’ importance given to interpersonal engagement was very circumstances, e.g., there was no difference in the weights similar across all age groups, underlining the ubiquity of given to disease and disability among those with and with- this dimension throughout the lifecourse and highlighting out LSI and, similarly, no difference in the importance given the need for health and social care services to, not only to productive engagement in those who were employed or treat disease and poor functioning, but also create oppor- not. A  specific criticism of the Rowe-Kahn model is that tunities for social interaction. Moreover, it is important to it reinforces social inequalities by defining successful aging recognize that attitudes among older people may continue as a state more easily achieved by those with higher socio- to change as they age and that patient-centered care is an economic position (Katz & Calasanti, 2015; Martinson evolving process. For example, there was a sharp drop in & Berridge, 2015; Stowe & Cooney, 2015). However, our the importance given to disability in the 75+ group, fol- stratified analyses suggest that attitudes toward success- lowing steady rises at younger ages. This is an isolated ful aging are not socially patterned, with almost identical finding and could be due to chance but could also reflect results for those with and without financial difficulties and shifting attitudes toward disability in an age group who those who were satisfied or dissatisfied with their income. were “living the vignette” and, perhaps, beginning to expe- In addition, despite different experiences throughout the rience, and therefore recognize, physical decline. Clinicians life course, results were largely consistent across vignette wishing to base their practice on shared decision making and respondents’ gender, although women gave somewhat and patient-centered care should recognize potential dif- more weight to volunteering, consistent with previous ferences between their own priorities for successful aging sociological work (Blaxter, 1990), perhaps reflecting trad- and those of their patients and, while there is no substitute itional gender roles in this regard. In terms of individual for face-to-face discussion with patients, our results pro- successful aging dimensions, the consistently low import- vide guidelines as to how these may differ. ance given to disease reinforces qualitative findings in Results from this study support and extend existing older individuals (Cosco et  al., 2013), and extends these work, providing unconfounded estimates of the relative to younger ages. An isolated, possibly chance, finding from importance given to six successful aging dimensions by the current analysis suggests that respondents living alone a large U.K.  population sample and demonstrating how gave more weight to disease than those in relationships, these vary across the lifecourse. However, it is not clear possibly reflecting greater perceived vulnerability in this whether wider societal policies such as health and social group, although disease remained among the dimensions care or pension provision influence these results and it regarded as least important. In contrast, other biomedical would be of considerable interest to repeat this experi- dimensions, such as disability and cognitive function, along ment in other populations where these differ. In addition, with interpersonal social engagement, were given some of given changing attitudes with age, it would be beneficial the highest weights by respondents of all ages. Morbidity in to understand how major life events such as retirement or older age is regarded as an important factor in determining bereavement influence these results. As well as informing health, social and economic policies, but policy makers and clinicians and policy makers working directly with older clinicians should recognize the relatively low value placed people, our results are relevant to researchers interested on disease by the general population and acknowledge the in measuring successful aging and its determinants. The greater importance to individuals of good functioning and Rowe-Kahn definition of successful aging is a widely used social engagement. research tool with “success” traditionally defined as a Perhaps the most striking results presented here are dichotomy in which all six criteria are met  although, in those demonstrating how attitudes to successful aging vary practice, very few older people achieve this, despite con- with age. The majority of existing work on perceptions of sidering themselves to be aging well (McLaughlin et  al., successful aging has focused on older people, while many 2012; Montross et  al., 2006; Strawbridge et  al., 2002). researchers, clinicians and policy makers are younger than A more pragmatic approach has been proposed in which those under study. Understanding the views of younger the extent of success in aging is measured by summing the individuals and how these differ from those at older ages number of favorable dimensions (Bowling, 2007; Bowling has the potential to close the gap between the attitudes of & Iliffe, 2011; Whitley et al., 2016), and our results may clinicians and their patients and to promote shared deci- provide a more “cutting edge” approach (Gu et al., 2017) sion making. For example, both physical function and pro- in which favorable dimensions are weighted according to ductive engagement were viewed as relatively unimportant the priorities of the general population. Such a measure by those of working age (<65  years) but their weights would provide a more nuanced approach to success- increased among older respondents, consistent with previ- ful aging and acknowledges the importance of quality ous work on lay perceptions of health (Blaxter, 1990). This as well as quantity of life, consistent with challenges to highlights the potential for relatively common problems the notion of health as complete physical, mental and so- of older age, such as struggling with food packaging or cial well-being (Huber et al., 2011). The quality-adjusted lacking a meaningful role in society, to be dismissed by life year (QALY), which weights different health states those involved in the care of older people. In contrast, the according to patient preference, is a well-established Downloaded from https://academic.oup.com/psychsocgerontology/article/75/2/293/5033526 by DeepDyve user on 16 July 2022 Journals of Gerontology: PSYCHOLOGICAL SCIENCES, 2020, Vol. 75, No. 2 301 Baltes, M. M., & Carstensen, L. L. (1996). The process of suc- health outcome (Whitehead & Ali, 2010). A  similar cessful ageing. Ageing Society, 16, 397–422. doi:10.1017/ measure based on weights such as those presented here s0144686x00003603 might form the basis for a modified Rowe-Kahn model Blaxter, M. (1990). Health and lifestyles. New York: Routledge, (Rowe & Kahn, 2015) that better represents societal Chapman and Hall. attitudes toward successful aging and could be used to Bloom, D. E., Chatterji, S., Kowal, P., Lloyd-Sherlock, P., McKee, M., evaluate interventions and direct policy investments to Rechel, B.,…Smith, J. P. (2015). Macroeconomic implications of promote successful aging worldwide. population ageing and selected policy responses. Lancet (London, England), 385, 649–657. doi:10.1016/S0140-6736(14)61464-1 st Bowling, A. (2007). Aspirations for older age in the 21 cen- Supplementary Material tury: What is successful aging? International Journal of Supplementary data is available at The Journals of Aging & Human Development, 64, 263–297. doi:10.2190/ Gerontology, Series B: Psychological Sciences and Social L0K1-87W4-9R01-7127 Sciences online. Bowling, A., & Iliffe, S. (2011). Psychological approach to successful ageing predicts future quality of life in older adults. Health and Quality of Life Outcomes, 9, 13. doi:10.1186/1477-7525-9-13 Funding Charbonneau-Lyons, D. L., Mosher-Ashley, P. M., & Stanford- E. Whitley and F.  Popham are funded by the Medical Research Pollock, M. (2002). Opinions of college students and inde- Council (MC_UU_12017/13) and the Chief Scientist Office, Scottish pendent-living adults regarding successful aging. Educational Government (SPHSU 13). M. Benzeval is funded by the UK Economic Gerontology, 28, 823–833. doi:10.1080/03601270290099822 and Social Research Council (ES/K005146/1). The funders had no Christensen, K., Doblhammer, G., Rau, R., & Vaupel, J. W. (2009). role in study design, data collection and analysis, decision to publish, Ageing populations: the challenges ahead. Lancet, 374(9696), or preparation of the manuscript. 1196–1208. doi:10.1016/S0140-6736(09)61460-4 Christensen, K., Thinggaard, M., Oksuzyan, A., Steenstrup, T., Andersen-Ranberg, K., Jeune, B.,…Vaupel, J. W. (2013). Physical Acknowledgments and cognitive functioning of people older than 90  years: A  comparison of two Danish cohorts born 10  years apart. This paper makes use of data from the Understanding Society Innovation Panel. Understanding Society is an initiative funded by Lancet (London, England), 382, 1507–1513. doi:10.1016/ the Economic and Social Research Council and various Government S0140-6736(13)60777-1 Departments, with scientific leadership by the Institute for Social Collings, P. (2001). “If you got everything, it’s good enough”: and Economic Research, University of Essex, and survey delivery Perspectives on successful aging in a Canadian Inuit community. by NatCen Social Research and Kantar Public. The research data Journal of Cross-Cultural Gerontology, 16, 127–155. doi:10.10 are distributed by the UK Data Service (SN 6849). The authors are 23/A:1010698200870 grateful to Vittal Katikireddi for advice on common diseases and Commission of the European Communities. (2009). Dealing with limitations in older people, Tarek Al Baghal for advice on experi- the impact of an ageing population in the EU (2009 Ageing mental design and question wording, and John Gilchrist for graphic Report). Brussels: Commission of the European Communities. production. doi:10.2765/80301 Cosco, T. D., Lemsalu, L., Brehme, D. F., Grigoruta, N., Kaufmann, L. K., Meex, R.,…Brayne, C. (2015). Younger Europeans’ con- Author Contributions ceptualizations of successful aging. Journal of the American FP had the original idea for the study, which was developed with EW Geriatrics Society, 63, 609–611. doi:10.1111/jgs.13307 and MB. EW, with FP and MB, developed the vignettes and study Cosco, T. D., Prina, A. M., Perales, J., Stephan, B. C. M., & Brayne, design. EW analyzed the data and wrote the first draft of the paper. C. (2013). Lay perspectives of successful ageing: A  system- All authors have reviewed and approved the final version. EW is the atic review and meta-ethnography. BMJ Open, 3:e002710. guarantor. doi:10.1136/bmjopen-2013–002710 Cosco, T. D., Prina, A. M., Perales, J., Stephan, B. C., & Brayne, C. (2014a). Operational definitions of successful aging: A  sys- Conflict of Interest tematic review. International Psychogeriatrics, 26, 373–381. None reported. doi:10.1017/S1041610213002287 Cosco, T. D., Prina, A. M., Perales, J., Stephan, B. C. M., & Brayne, C. (2014b). Whose “successful ageing”? Lay- and researcher- References driven conceptualisations of ageing well. Eur J Psychiat, 28, 124–130. doi:10.4321/S0213-61632014000200006 Araújo, L., Ribeiro, O., Teixeira, L., & Paúl, C. (2016). Predicting Depp, C. A., Glatt, S. J., & Jeste, D. V. (2007). Recent advances in successful aging at one hundred years of age. Research on Aging, research on successful or healthy aging. Current Psychiatry 38, 689–709. doi:10.1177/0164027515603771 Reports, 9, 7–13. doi:10.1007/s11920-007-0003-0 Auspurg, K., & Hinz, T. (2015). Factorial survey experiments. Los Depp, C. A., & Jeste, D. V. (2006). Definitions and predictors of Angeles: SAGE Publications Inc. doi:10.4135/9781483398075 successful aging: A comprehensive review of larger quantitative Baltes, P. B., & Baltes, M. M. (1990). Successful aging: Perspectives studies. The American Journal of Geriatric Psychiatry, 14, 6–20. from the behavioral sciences. New York: Cambridge University doi:10.1097/01.JGP.0000192501.03069.bc Press. doi:10.1017/CBO9780511665684 Downloaded from https://academic.oup.com/psychsocgerontology/article/75/2/293/5033526 by DeepDyve user on 16 July 2022 302 Journals of Gerontology: PSYCHOLOGICAL SCIENCES, 2020, Vol. 75, No. 2 Ferri, C., James, I., & Pruchno, R. (2009). Successful aging: Definitions Morrow-Howell, N. (2010). Volunteering in later life: Research frontiers. The Journals of Gerontology, Series B: Psychological and subjective assessment according to older adults. Clinical Gerontologist, 32, 379–388. doi:10.1080/07317110802677302 Sciences and Social Sciences, 65, 461–469. doi:10.1093/geronb/ gbq024 Gu, D., Feng, Q., Sautter, J. M., Yang, F., Ma, L., & Zhen, Z. (2017). Concordance and discordance of self-rated and researcher- Nimrod, G., & Ben-Shem, I. (2015). Successful aging as a lifelong process. Educational Gerontology, 41, 814–824. doi:10.1080/0 measured successful aging: Subtypes and associated factors. The Journals of Gerontology, Series B: Psychological Sciences and 3601277.2015.1050904 Phelan, E. A., Anderson, L. A., LaCroix, A. Z., & Larson, E. B. (2004). Social Sciences, 72, 214–227. doi:10.1093/geronb/gbw143 Huber, M., Knottnerus, J. A., Green, L., van der Horst, H., Jadad, Older adults’ views of “successful aging”–how do they compare with researchers’ definitions? Journal of the American Geriatrics A. R., Kromhout, D.,…Smid, H. (2011). How should we define health? BMJ (Clinical research ed.), 343, d4163. doi:10.1136/ Society, 52, 211–216. doi:10.1111/j.1532-5415.2004.52056.x Pruchno, R. A. (2015). Successful aging: Contentious past, productive bmj.d4163 Jäckle, A., Gaia, A., Al Baghal, T., Burton, J., & Lynn, P. (2017). future. Gerontologist, 55, 1–4. doi:10.1093/geront/gnv002 Pruchno, R., & Carr, D. (2017). Successful aging 2.0: Resilience and Understanding Society – The UK Household Longitudinal Study, Innovation Panel, Waves 1–9, User Manual. Colchester: beyond. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 72, 201–203. doi:10.1093/geronb/ University of Essex. Jopp, D. S., Wozniak, D., Damarin, A. K., De Feo, M., Jung, S., & gbw214 Rowe, J. W., & Kahn, R. L. (1997). Successful aging. The Jeswani, S. (2015). How could lay perspectives on successful aging complement scientific theory? Findings from a U.S. And Gerontologist, 37, 433–440. doi:10.1093/geront/37.4.433 Rowe, J. W., & Kahn, R. L. (2015). Successful aging 2.0: Conceptual a German life-span sample. The Gerontologist, 55, 91–106. st doi:10.1093/geront/gnu059 expansions for the 21 Century. The Journals of Gerontology, Joseph-Williams, N., Lloyd, A., Edwards, A., Stobbart, L., Tomson, Series B: Psychological Sciences and Social Sciences, 70, 593– D., Macphail, S.,…Thomson, R. (2017). Implementing shared 596. doi:10.1093/geronb/gbv025 decision making in the NHS: Lessons from the MAGIC pro- Spijker, J., & MacInnes, J. (2013). Population ageing: The timebomb gramme. BMJ (Clinical research ed.), 357, j1744. doi:10.1136/ that isn’t? BMJ (Clinical research ed.), 347, f6598. doi:10.1136/ bmj.j1744 bmj.f6598 Katz, S., & Calasanti, T. (2015). Critical perspectives on success- Stowe, J. D., & Cooney, T. M. (2015). Examining Rowe and Kahn’s ful aging: Does it “appeal more than it illuminates”? The concept of successful aging: Importance of taking a life course Gerontologist, 55, 26–33. doi:10.1093/geront/gnu027 perspective. The Gerontologist, 55, 43–50. doi:10.1093/geront/ Lupien, S. J., & Wan, N. (2004). Successful ageing: From cell to self. gnu055 Philosophical Transactions of the Royal Society of London, Strawbridge, W. J., Wallhagen, M. I., & Cohen, R. D. (2002). Series B: Biological Sciences, 359, 1413–1426. doi:10.1098/ Successful aging and well-being: Self-rated compared with rstb.2004.1516 Rowe and Kahn. The Gerontologist, 42, 727–733. doi:10.1093/ Manning, L. K., Carr, D. C., & Kail, B. L. (2016). Do higher lev- geront/42.6.727 els of resilience buffer the deleterious impact of chronic illness Tate, R. B., Swift, A. U., & Bayomi, D. J. (2013). Older men’s lay defi- on disability in later life? The Gerontologist, 56, 514–524. nitions of successful aging over time: The Manitoba follow-up doi:10.1093/geront/gnu068 study. International Journal of Aging & Human Development, Martin, P., Kelly, N., Kahana, B., Kahana, E., Willcox, B. J., Willcox, 76, 297–322. doi:10.2190/AG.76.4.b D. C., & Poon, L. W. (2015). Defining successful aging: A tangible United Nations. (2002). Report of the Second World Assembly on or elusive concept? The Gerontologist, 55, 14–25. doi:10.1093/ Ageing A/CONF.197/9. New York: United Nations. geront/gnu044 Vaupel, J. W. (2010). Biodemography of human ageing. Nature, 464, Martinson, M., & Berridge, C. (2015). Successful aging and its dis- 536–542. doi:10.1038/nature08984 contents: A  systematic review of the social gerontology litera- Whitehead, S. J., & Ali, S. (2010). Health outcomes in economic ture. The Gerontologist, 55, 58–69. doi:10.1093/geront/gnu037 evaluation: The QALY and utilities. British Medical Bulletin, 96, McDowell, I. (2006). Measuring health: A  guide to rating scales 5–21. doi:10.1093/bmb/ldq033 and questionnaires. New York: Oxford University Press. Whitley, E., Popham, F., & Benzeval, M. (2016). Comparison of the doi:10.1093/acprof:oso/9780195165678.001.0001 Rowe-Kahn model of successful aging with self-rated health and McLaughlin, S. J., Jette, A. M., & Connell, C. M. (2012). An examin- life satisfaction: The west of Scotland twenty-07 prospective ation of healthy aging across a conceptual continuum: Prevalence cohort study. The Gerontologist, 56, 1082–1092. doi:10.1093/ estimates, demographic patterns, and validity. The Journals of geront/gnv054 Gerontology, Series A: Biological Sciences and Medical Sciences, World Health Organisation. (2015). World report on ageing and 67, 783–789. doi:10.1093/gerona/glr234 health. Geneva, Switzerland: World Health Organisation. Montross, L. P., Depp, C., Daly, J., Reichstadt, J., Golshan, S., Moore, Young, Y., Frick, K. D., & Phelan, E. A. (2009). Can successful aging D.,…Jeste, D. V. (2006). Correlates of self-rated successful and chronic illness coexist in the same individual? A  multidi- aging among community-dwelling older adults. The American mensional concept of successful aging. Journal of the American Journal of Geriatric Psychiatry, 14, 43–51. doi:10.1097/01. Medical Directors Association, 10, 87–92. doi:10.1016/j. JGP.0000192489.43179.31 jamda.2008.11.003

Journal

"The Journals of Gerontology - Series B: Psychological Sciences and Social Sciences" – Oxford University Press

Published: Jan 14, 2020

Keywords: productive aging; older adult; aging; disability; mental processes; cognitive ability; physical function; attitude

There are no references for this article.