Ageism and Its Impact on Quality of Life Among Older Persons Experiencing Difficulties with Activities of Daily Living

Ethics

For the collection of the primary data, Ethical Approval for ELSA Wave 1 was granted from the Multicentre Research and Ethics Committee (reference number: MREC/01/2/91) approved on 7th February 2002.²⁰ The participants were freely able to choose to participate or not to participate in the research without any repercussions. The participants’ informed consent could be withdrawn at any time without questioning the reason for the withdrawal. Additionally, approval from the Ethics Committee of the University of Tsukuba, Japan, was obtained for the secondary analyses of the data (approval number: 1817; R4/10/14). Our study was conducted in compliance with the ethical guidelines of the Helsinki Declaration and it complied with the Ethical Guidelines for Medical and Biological Research Involving Human Subjects.

Study sample and source

We used information from the English Longitudinal Study of Ageing (ELSA), a recurring panel survey that gathers information from a representative sample of English people aged 50 and over, collected biennially across a variety of fields. Interviewers collected data by conducting Computer-Assisted Personal Interviewing (CAPI) and utilising self-completion questionnaires. At the participants’ homes, face-to-face first interviews were done. Telephone or in-person visits might be used to perform follow-up interviews.²¹ We used the 2010–2011 period (wave 5) for the original data set because, in contrast to other iterations of this panel survey, it contained explicit questions about perceived age discrimination. To see the effect on quality-of-life scores in subsequent periods, waves 7 (period 2016-17) and wave 9 (period 2018-2019) were also obtained.

Number of participants

The number of participants initially consisted of 10,274 people. Inclusion criteria were older adults who answered the questions themselves and with valid responses to the questions on perceived age discrimination and difficulties with ADL/IADL. A total of 537 participants (5.2%) were excluded due to their responses being provided by a proxy respondent, leaving 9,737 participants. The decision to exclude proxy respondents from the analysis was guided by the objective to understand and explore the direct experiences and perceptions of older individuals with respect to age discrimination and quality of life. While proxy respondents can provide valuable information, particularly when the older person is unable to respond due to cognitive or physical limitations, they may not fully capture the personal perceptions and experiences of the older individuals themselves. Given the subjective nature of perceived age discrimination and quality of life assessments, it was crucial to utilise first-hand responses to ensure the authenticity and accuracy of the data. Furthermore, studies have shown that proxy responses can sometimes be systematically different from self-reports, potentially introducing bias into the analysis.²²

Out of these respondents, data for perceived age discrimination (either positive or negative responses) was unavailable for 4,870 individuals (50%). The remaining 4,867 respondents had valid responses. Within this group, we further excluded 28 respondents (0.6%) for not having a quality-of-life measure, for a final sample of 4,839 participants. Among these, 3,531 respondents (73.0%) reported no difficulties with ADL or IADL, while 1,308 respondents (27.0%) reported having difficulties with either ADL or IADL. Details can be found in Figure 1.

Figure 1. Inclusion and exclusion criteria for the sample of participants.

Measurements

Dependency

Dependency was defined as having difficulties with at least one ADL or IADL. The activities included were:

Activities of Daily Living (ADLs):

Instrumental Activities of Daily Living (IADLs):

Given that most participants reported experiencing no difficulties, responses were dichotomised into categories delineating those without difficulties and those with one or more difficulties. This variable offers significant insights into the level of participant independence. To delve further into the compound effects of these difficulties, bivariate analyses were conducted, segregating the sample into three distinct groups: the first with a single ADL/IADL difficulty, the second with two such difficulties, and the third presenting three or more difficulties.

Perceived age discrimination

The English Longitudinal Study of Ageing (ELSA) incorporated questions concerning experiences of discrimination for a singular instance in its fifth wave. These queries have also been deployed in other longitudinal examinations, notably within the Health and Retirement Survey (HRS)²³ and the Midlife in the United States (MIDUS) survey.²⁴

Participants were asked about the perceived frequency of five discriminatory circumstances as follows: ‘How often do any of the following events occur in your daily life?’

The answers could be from 1, meaning ‘almost daily’, to 6, meaning ‘never’. We split the answers into two groups to see if participants faced discrimination in the past year (more than once a year vs. less than once a year or never).

Subsequently, participants were asked a follow-up question about their perceptions of the reasons for the discriminatory experiences. The possible reasons were age, gender, race, weight, physical disability, and other. Participants could give multiple attributing reasons. For our analytical purposes, we focused exclusively on responses that attributed the experienced discrimination to age.

Quality of life

The CASP-19 scale, a well-known tool designed particularly for assessing quality of life in older individuals, was used to evaluate overall quality of life. It covers four domains: control, autonomy, self-realization, and pleasure.

Covariates

In our study, we considered covariates found in previous studies, which were categorized as follows: Age was divided into four groups, less than 60 years old, 60-69 years old, 70-79 years old, and 80 years old or more. The participants were identified as either male or female for the variable sex. Excluding pensions, household wealth was segmented into five groups, or quintiles, with the 1st quintile representing the least wealthy and the 5th quintile symbolizing the most affluent households. The education level of the participants was classified into low, middle, or high. Marital status was categorized as either in a partnership, which includes those who are married or cohabiting, or not in a partnership, which includes those who are divorced, separated, or widowed. For ethnicity, participants were identified as either white or non-white. Physical health was assessed through the participants’ self-perception of their health, which was categorized as either excellent/very good/good or fair/poor, providing insight into the participants’ own evaluation of their physical health status. Finally, the Center for Epidemiologic Studies Depression (CES-D) scale²⁵ was employed to assess mental health. A score of 4 or higher on this scale signifies the existence of depressive symptoms.

Analytical approach

The study adopted both cross-sectional and longitudinal approaches for analysis. The cross-sectional analysis examined the data collected in ELSA wave 5 (2010-2011) to examine the relationship between perceived age discrimination, dependency, and quality of life at a single time point. Subsequently, the longitudinal analysis followed the same cohort across the subsequent waves in 2016-2017 (wave 7) and 2018-2019 (wave 9). These analyses enabled the tracking of changes in quality of life over time.

Statistical analysis

The analyses of the sociodemographic characteristics were carried out by analysing the data using bivariate statistics to summarize and organize the characteristics of the study population and to assess the distribution within the groups with difficulties with ADL/IADL characteristics and those without difficulties. At this stage, categorical variables were presented as frequencies and percentages.

To investigate the relationship between perceived discrimination and overall quality of life, we ran linear regression analyses. We considered a number of variables in these studies that may have an impact on the findings, including age, gender, ethnicity, income level, education level, marital status, physical and mental health issues, and employment status. We used the Hosmer Lemes how test²⁶ to assess how well our models fit the data.

On the relationship between perceived age discrimination and quality of life, cross-sectional (wave 5) and longitudinal (wave 7 and 9) studies were carried out. Three cross-sectional logistic regression analyses were conducted utilising baseline data as part of the cross-analysis. Regardless of their ADL/IADL status, all individuals were included in the initial cross-sectional study. The second and third cross-sectional analyses were performed on people without ADL or IADL difficulties and those with difficulties, respectively. Finally, the prospective analyses to assess changes in quality of life were carried out in two subsequent periods: wave 7 (2016–2017) and 9 (2018–2019). For the prospective analyses the sample was restricted to those who reported having trouble performing ADL or IADL.

All models were adjusted for age, sex, wealth, education level, partnership status, ethnicity, perceived health, depressive symptoms, and current working status. The prospective models for waves 7 and 9 were also adjusted for baseline CASP-19 scores (wave 5).

Before the analyses, we verified the assumptions of normality, linearity, and homoscedasticity. Additionally, we examined multicollinearity among the variables as it could potentially impact the outcomes of the logistic regression. All tests were performed using a two-sided approach and any p value, below 0.05 was considered significant. Confidence intervals were set at the 95% level of certainty. We utilized IBM SPSS Statistics version 29.0.1.0 for all our analyses.

Sociodemographic characteristics and ADL/IADL difficulties

In our sample, 61% (2,966 participants) said they faced age discrimination. Table 1 shows that participants with ADL/IADL difficulties were more likely to experience age discrimination. Specifically, 29% (867 participants) of those with difficulties reported age discrimination, compared to 23.5% (441 participants) of those who did not.

Table 1 additionally indicates a correlation between age and ADL/IADL challenges, with an age-related rising trend. It is observed that women encounter these difficulties more frequently than do males. Socioeconomic characteristics are also connected, with households in lower quintiles of wealth reporting greater difficulty than their counterparts in higher quintiles. Similarly, those with lower education levels report higher prevalence. This table also shows a connection with marital status, as individuals not in a partnership, including those divorced, separated or widowed, face these difficulties more often than those in partnerships. Ethnicity does not have a significant value in this context. Self-perception of physical health is linked to ADL/IADL difficulties, with individuals assessing their health as fair or poor reporting more difficulties. Mental health, assessed by the CES-D scale, shows a higher prevalence of ADL/IADL difficulties among individuals exhibiting depressive symptoms. Employment status shows that those who are not currently working face these difficulties more frequently.

ADL/IADL difficulties and discrimination settings

Those reporting one difficulty with ADL or IADL are 1.475 times more likely to experience age discrimination compared to the reference group with no difficulties (OR=1.475, 95% CI: 1.213-1.795). The odds rise to 1.674 for individuals with two difficulties (OR=1.674, 95% CI: 1.269-2.207). However, the increase in the odds of experiencing age discrimination for those encountering three or more difficulties is not statistically significant (OR=1.105, 95% CI: 0.915-1.335).

Table 2 also analyses the different scenarios where discrimination was perceived. For those experiencing three or more difficulties, a significant correlation was observed in situations where they were treated with less courtesy than others (OR=1.335, 95% CI: 1.109-1.608).

The likelihood of receiving poorer service in restaurants increased significantly across all difficulty categories, with the highest odds ratio for individuals with three or more difficulties (OR=1.403, 95% CI: 1.136-1.732).

Also, the odds of respondents feeling that they were regarded as less intelligent increased significantly across all categories of difficulties, peaking for those with three or more difficulties (OR=2.289, 95% CI: 1.892-2.769) when compared to those without difficulties.

A similar trend was observed for reporting that they were being threatened or harassed, with the highest odds ratio belonging to those with three or more difficulties (OR=1.788, 95% CI: 1.408-2.272).

Finally, the odds of receiving inferior service from hospitals were significantly higher, particularly for those with two difficulties (OR=1.763, 95% CI: 1.352-2.299) and those with three or more difficulties (OR=1.715, 95% CI: 1.412-2.084).

Dependency, discrimination, and quality of life

Table 3 displays the variance in the CASP 19 score, which measures quality of life, on a scale of 0 to 57 points between adults who do not have any difficulties with Activities of Daily Living (ADL) or Instrumental Activities of Daily Living (IADL) and those who face such challenges. The mean CASP 19 score for individuals without ADL/IADL difficulties (n=3531) is 41.5 points whereas the average score for individuals with these difficulties (n=1308) is significantly lower at 33.6 points. The difference between the two groups means is 7.9 points with an error of 0.29077. The t value of 27.207 indicates significance at p<0.001. The standard deviation of CASP 19 scores is slightly higher in the group facing ADL/IADL difficulties (9.3) compared to those without difficulties (8.1). This suggests a range of variability among individuals dealing with these challenges. Additionally, there is an error of the mean (SEM) in the group facing difficulties compared to those without them (0.257 vs 0.137) indicating greater variability in their scores. In summary these findings demonstrate that individuals experiencing ADL/IADL difficulties tend to have poorer scores than those without such challenges.

Table 4 displays the mean CASP-19 score difference between those that and those who did not experience age discrimination. Participants who did not perceive age discrimination had an average CASP-19 score of 40.4 (n=1873). The mean score was slightly lower at 38.7 for those who reported age discrimination (n=2966), though. This shows a mean difference of 1.7 points between the two groups that is statistically significant.

Quality of life and ADL/IADL difficulties

Table 5 shows the results from a cross-sectional linear regression analysis in wave 5. This analysis looked at the relationship between quality of life (measured using the CASP-19 score points) and perceived age discrimination in three distinct groups: all participants who perceived age discrimination, regardless of difficulties; participants without any ADL/IADL difficulties who perceived age discrimination; and participants with ADL/IADL difficulties who perceived age discrimination.

Our main outcome variable is Quality-of-Life Scale, measured by CASP-19. The results indicate that participants who perceived discrimination tended to have a lower quality of life score. This was significant in the overall participant group (Unstandardized Beta Coefficients (B): -1.106; CI:-1.577~-0.634) with a p-value of <0.001, and among participants without difficulties (B: -1.401; CI: -1.934~-0.868) with a p-value of <0.001. However, there was no significant association for participants with difficulties (B: -0.401; CI -1.353~0.552) with a p-value of 0.409. A more detailed breakdown of all variables by group, including covariates, can be found above.

All participants who Perceived Age Discrimination (n=4246)

Age was not a significant determinant of perceived age discrimination. Being female was significantly associated with increasing CASP-19 scores (B: 0.738; p=0.001). As wealth increases by quintile there is a significant increase in quality-of-life scores compared to the lower quintiles (B: 1.003; p≤0.001). Those who were married, or cohabiting had higher CASP-19 scores (B: 0.690; p=0.007). Being white had no significant association. Regarding health status, participants who rated their physical health as “fair/poor” had lower CASP-19 scores (B: -5.801; p=0.033). Participants in the group over four symptoms of depression had lower CASP-19 scores (B: -8.314; p≤0.001). The current working status showed no statistical significance.

Participants without ADL/IADL Difficulties who Perceived Age Discrimination (n=3056)

Age was not a significant determinant of perceived age discrimination. Being female was significantly associated with better CASP-19 scores (B: 0.647; p=0.014). As wealth increases by quintile, there is a significant increase in quality-of-life scores (B: 0.992; p≤0.001). Those who were married, or cohabiting had higher CASP-19 scores (B: 0.756; p=0.010). Being white had no significant association. Regarding health status, participants who rated their physical health as “fair/poor” had lower CASP-19 scores (B: - 4.714; p≤0.001). Participants in the group over four symptoms of depression had lower CASP-19 scores (B: -8.299; p≤0.001). The current working status showed no statistical significance.

Participants with ADL/IADL Difficulties who Perceived Age Discrimination (n=1308)

The likelihood of increasing quality-of-life scores were higher for women compared to men (B:1.162: 0.872; p=0.011). Also, the higher the quintile of wealth the better the CASP-19 scores (B=0.714; p≤0.001). Regarding health status, participants who rated their physical health as “fair/poor” had lower CASP-19 scores (B: -4.749; p≤0.001). Participants in the group with 4 or more symptoms of depression had lower CASP-19 scores (B: -7.628; p≤0.001). Unlike other groups, being in a partnership had no positive effect on quality-of-life scores. Also, as mentioned before, the association of perceived age discrimination and CASP-19 score was not significant in this group (B: -0.401; p=0.409), suggesting that the effect of quality of life on perceived age discrimination may be less pronounced among individuals with ADL/IADL difficulties.

Table 6 shows that there is no significant association of perceived age discrimination and quality-of-life scores among those with ADL/IADL difficulties. The lack of significance prevails on all models; cross-sectional linear regression analysis for wave 5, and prospective linear regressions for waves 7 and 9. The numbers analysed were 1,190 participants in Wave 5, 757 in Wave 7, and 530 participants in Wave 9.

Limitations

Several limitations within our study should be acknowledged when interpreting our findings. Firstly, the measure for age discrimination is subjective, meaning it is based on individuals’ perceptions rather than objective, external assessments. While subjective experiences are critical in understanding how age discrimination affects older adults, it does open the door for potential bias. Individuals’ responses could be influenced by personality characteristics, mood states, cognitive biases, or cultural norms. Therefore, these responses might not reflect the full extent or absence of actual discriminatory behaviours or attitudes experienced by the respondents.

Secondly, it is important to note that the assessment of quality of life in our study is subjective. Although this approach provides insights into experiences and perspectives it can also be influenced by variations in optimism, resilience and other character traits. Consequently, there might exist a disparity, between individuals self-reported quality of life and objective indicators of health and overall well-being.

Thirdly, the measure for difficulties with ADL/IADL in our study is binary, meaning it does not fully explore the range of needs among participants. We categorized participants into those with one or more difficulties and those without, thus potentially grouping together individuals with diverse levels of need. For most of the analysis, someone experiencing potentially minor difficulties with only one activity was treated the same as someone with severe difficulties across multiple activities. This could obscure more nuanced associations between the extent of difficulties with ADL/IADL, age discrimination, and quality of life.

An additional limitation pertains to the temporal nature of our age discrimination measure. The question on discrimination asked participants to recall past instances of perceived discriminatory situations, ranging from daily occurrences to instances less than once a year, as well as an option for never experiencing such situations. This reliance on retrospective self-reporting could introduce recall bias, particularly for those participants who were asked to remember events that happened infrequently or long ago. This bias might distort the true association between age discrimination and quality of life.

Also, our analysis skipped the most immediate wave of data collection and used data every other wave, which took place every two years. This approach could potentially distort the strength of the observed association between perceived age discrimination and quality of life. Given the biennial nature of the data collection, our study might not capture changes in perceived discrimination and quality of life that occurred in the intervening years. It is plausible that the experiences and perceptions of participants may change within shorter time intervals, and this fluctuation may not be accurately captured by our measurements.

Furthermore, our measure for difficulties did not discriminate between ADL, which includes basic self-care tasks, and IADL, which encompasses more complex activities related to independent living. This distinction could be important because difficulties with more simple but critical activities could potentially have a larger impact on quality of life and may be more closely tied to experiences of discrimination.

Lastly, while ADL/IADL difficulties can be an indicator of disability, it does not equate to a disability or dependency diagnosis. In our study, it is merely an indicator of a certain level of difficulty with daily activities that is not clearly defined. As such, this measure might not accurately capture the full range of functional capabilities or impairments among our participants. As mentioned by Gill et al.,²⁹ there is no consensus on ADL/IADL as a measure of dependence. Most epidemiological studies agree with us in considering difficulties with ADL/IADL as dependence, but other argue that it should be defined as degree of difficulty.²⁹

Future research should take these limitations into account. The integration of more objective measures, a distinction between types of difficulties and degrees of impairment, and a precise definition of difficulty levels could yield a more holistic understanding of the effects of age discrimination on the quality of life among older adults.