Study Design and Participants
We used cross-sectional data collected through the ongoing PROTECT (https://www.protectstudy.org.uk) study in 2019. Individuals were eligible to participate in the PROTECT study if they were UK residents, English speakers, aged 50 years or over, had access to internet, and did not have a clinical diagnosis of dementia at the point of recruitment. In PROTECT, participants were recruited through national publicity and via existing cohorts of older adults (Brains for Dementia Research; Exeter 10000; Join Dementia Research). Potential participants enrolled through the PROTECT study website and provided informed consent. As part of the 2019 annual assessment, PROTECT participants were invited to complete additional measures on self-perceptions of ageing and health. Analyses for this study are based only on those participants who completed the additional measures between 1st January 2019 and 31st March 2019. The PROTECT study obtained ethical approval from the London Bridge NHS Research Ethics Committee and Health Research Authority (Ref: 13/LO/1578). Ethical approval for the data analysis was sought through the ethics committee at the University of Exeter, School of Psychology (Application ID: eCLESPsy000603 v1.0).
The study sample comprised 6,192 participants. The mean (standard deviation (SD); range) age was 66.1 years (7.0 years; 51 to 95 years) and 76.0% were women. Further demographic characteristics are reported in Table 1. Only 0.37% of participants reported having mild cognitive impairment (MCI). Based on the cognitive task scores, 5.6% of participants might experience age-associated cognitive decline (as they scored more than 1 SD below the mean sample score in two or more cognitive tasks) and 1.7% might have MCI (as they scored more than 1.5 SD below the mean sample score in two or more cognitive tasks).
Table 1. Descriptive statistics of demographic variables and main study variables
Variables
|
|
Age in years, Mean (SD; Range)
|
66.1 (7; 51.4 to 95.9)
|
Female, %
|
75.9
|
Marital status, %
Married
Civil partnership
Co-habiting
Widowed
Separated
Divorced
Single
|
67.4
0.5
5.9
7.4
1.7
11.0
6.2
|
Education level, %
Secondary education
Post-secondary education
Vocational qualification
Undergraduate degree
Post-graduate degree
Doctorate
|
13.7
11.3
20.1
33.8
17.3
3.9
|
Currently employed, %
Employed full-time
Employed part-time
Self-employed
Retired
Unemployed
|
15.8
15.7
9.5
56.7
2.4
|
AARC gains, %
Not at all
A little bit
Moderately
Quite a bit
Very much
|
0.1
5.1
22.2
47.6
25.0
|
AARC losses, %
Not at all
A little bit
Moderately
Quite a bit
Very much
|
3.8
61.0
28.7
5.7
0.8
|
Digit span, Mean (SD)
|
7.6 (1.5)
|
Paired associate learning, Mean (SD)
|
4.7 (0.9)
|
Grammatical reasoning, Mean (SD)
|
37.4 (10.6)
|
Self-ordered search, Mean (SD)
|
7.7 (2.6)
|
Self-rated health, %
Poor
Fair
Good
Excellent
|
2.0
12.9
54.5
30.6
|
Functional ability, Mean (SD)
|
0.2 (0.8)
|
Depressive symptoms, Mean (SD)
|
11.4 (2.9)
|
Anxiety symptoms, Mean (SD)
|
8.4 (2.4)
|
Sample size ranged from 5811 to 6192. Secondary education – GCSE/O-Levels. Post-Secondary education – College, A-Levels, NVQ3 or below, or similar. Vocational qualification – Diploma, Certificate, BTEC, NVQ 4 and above, or similar. Undergraduate Degree – BA, BSc, or similar. Post-graduate Degree – MA, MSc, or similar. Doctorate – PhD. AARC gains – Total score on the AARC gains subscale from the AARC-10 SF. AARC losses – Total score on the AARC losses subscale from the AARC-10 SF. Digit span – Computerised cognitive task assessing verbal working memory. Paired associate learning – Computerised cognitive task assessing visual episodic memory. Grammatical reasoning – Computerised cognitive task assessing verbal reasoning. Self-ordered search – Computerised cognitive task assessing spatial working memory. IADL – Instrumental activities of daily living. Depressive symptoms – Score on the Patient Health questionnaire (PHQ-9). Anxiety symptoms – Score on the Generalized Anxiety Disorder questionnaire (GAD-7).
Measures
Socio-demographic variables.
Participants reported their age, sex, marital status (married, in a civil partnership, cohabiting, unmarried, divorced, separated, and widowed), education level (secondary education, post-secondary education, vocational qualifications, undergraduate degrees, post-graduate degrees, doctorates), and current employment (employed full-time, employed part-time, self-employed, retired, unemployed).
Awareness of Age-Related Change (AARC).
The AARC-10 SF (Kaspar et al., 2019) contains ten items (reported in Supplementary Table 1); five assessing AARC gains and five assessing AARC losses. Each item per gain and loss dimension represents one out of five life domains (health and physical functioning, cognitive functioning, interpersonal relationships, socio-cognitive and socio-emotional functioning, and lifestyle/engagement; see again Miche et al., 2014). All items start with the same stem: “With my increasing age, I realise that…”. Respondents rate how much each item applies to them on a five-point Likert scale (from “not at all” (1) to “very much” (5)). Scores can be obtained for the AARC gains and AARC losses subscales by summing items that fall into the respective scales. Scale scores range from a minimum of five to a maximum of 25; higher scores indicate higher levels of AARC. In the current study sample Cronbach’s α values (internal consistency) for the AARC gains and the AARC losses scales are 0.76 and 0.79, respectively (Sabatini et al., 2020b).
Indicators of physical and functional health.
Lawton’s Instrumental Activities of Daily Living Scale (IADL; Lawton and Brody, 1969) is a seven-item instrument assessing everyday functional status. Each item describes an activity (e.g. preparing meals). Respondents have to rate how difficult they find performing the activity (0= “no difficulty”, 1= “some difficulty”, and 2= “great difficulty”). The total score ranges from a possible 0 to 14, with higher scores indicating greater functional difficulty. Inter-rater reliability for the IADL scale resulted in a Pearson’s correlation of 0.85 (Lawton and Brody, 1969). In the current study sample Cronbach’s α value for the IADL scale is 0.79.
Self-rated health was assessed with a single-item question (taken from the SF-36; Ware and Sherbourne, 1992) asking participants to rate their own health on a four-point scale (Excellent, good, fair, poor).
Mental health.
The Patient Health Questionnaire-9 (PHQ-9; Kroenke et al., 2001) is a nine-item scale capturing depressive symptoms over the previous two weeks. Respondents are asked to indicate how frequently they experience each symptom on a 4-point scale (from “not at all” (1) to “nearly every day” (4)). The total score is the sum of the item scores and can range from a possible 9 to 36; higher scores indicate the presence of more depressive symptoms. The PHQ-9 has excellent internal consistency with Cronbach’s α coefficient of 0.84 in the normative sample (Kroenke et al., 2001) and of 0.76 in the current study sample.
The Generalized Anxiety Disorder-7 (GAD-7; Spitzer et al., 2006) is a seven-item measure asking respondents to indicate the frequency of occurrence of a list of symptoms of generalized anxiety disorder on a 4-point scale (from “not at all” (1) to “nearly every day” (4)). The scale score is the sum of the item scores and ranges from a possible 7 to 28; higher scores indicate greater presence of anxiety symptoms. The GAD-7 is an internally consistent measure with a Cronbach’s a of 0.92 in the normative sample (Spitzer et al., 2006) and of 0.86 in the sample of the current study.
Cognitive Functioning.
Cognitive function was measured with the PROTECT Cognitive Test Battery (Corbett et al., 2015; Hampshire et al., 2012; Huntley et al., 2018) which includes four tasks: Digit Span; Paired Associate Learning; Grammatical Reasoning; and Self-ordered Search. For each task a summary score can be obtained by subtracting the number of errors from the number of correct answers; a higher score indicates better performance. For Digit Span the summary score can range from 0 to 20. For Paired Associate Learning the summary score can range from 0 to 16. For Grammatical Reasoning the summary score is also obtained by subtracting the number of errors from the number of correct answers, but the score has no upper or lower limit due to the fact that respondents can make attempts on as may trials as they can in the available time (three minutes). Finally, the summary score for the Self-ordered Search Task can range from 0 to 20.
Analyses
To explore whether the population can be divided into classes of individuals characterised by different profiles of levels of AARC gains and losses we conducted latent profile analysis, using Mplus software (Muthén and Muthén, 2017). We fit the latent profile models based on manifest variables representing responses to the 10 items of the AARC-10 SF (5 gain items and 5 loss items). The 10 items assessing AARC were treated as continuous manifest variables. To identify the model with the optimal number of classes, we fit a two-class model and systematically increased the number of classes by 1 until adding more classes no longer resulted in an improvement in model fit and did not compromise the parsimony of the model. The main criteria used to identify the best fitting model were the Vuong-Lo-Mendell-Rubin and Lo-Mendell-Rubin adjusted likelihood ratio tests; we also report Akaike’s information criterion (AIC), the Bayesian information criterion (BIC) and the entropy statistic (Nylund et al., 2007). The best fitting model is one which has a low value on the information criterion while being parsimonious in the number of identified classes. Having identified the best fitting model, we reported estimates of the percentage of the population falling in each class and estimates of the mean and standard deviation for each of the 10 AARC items for each class. The latter estimates were the basis for assigning names that characterise the profile of responses in each class.
Analysis of variance and chi-square tests were conducted to compare physical, functional, mental (depressive and anxiety symptoms), and cognitive health (scores on tasks assessing digit span, paired associate learning, grammatical reasoning, and self-ordered search) across the classes identified in the latent profile analysis. For the analysis of variance comparing physical, functional, and mental health and cognitive functioning, we fitted two models: one unadjusted and one adjusted for the effects of age, sex, marital status, education level, and current employment. We adjusted for these demographic variables as they may all be related to levels of mental, physical, and cognitive health in older age (Alavinia and Burdorf, 2008; Alexopoulos, 2005; Anstey et al., 2017; Banazak, 1997; Herd et al., 2007; Hughes and Waite, 2009; Weyerer et al., 2013). For these analyses study participants were allocated to the class for which they had the greatest probability of membership. For the results from the analyses of variance, the effect size was calculated using eta squared (η2). We interpreted effect sizes for the eta squared between 0.01 and 0.05 as small, between 0.06 and 0.13 as moderate and of 0.14 or above as large (Cohen, 1988). The analyses were carried out using STATA version 16 (StataCorp, 2017).