Visual and hearing impairments are associated with cognitive decline in older

Americans, Britons and Europeans

Abstract

Introduction: Highly prevalent hearing and vision sensory impairments among older people

may contribute to the risk of cognitive decline and pathological impairments including

dementia.

This study aims to determine whether single and dual sensory impairment (hearing and/or

vision) are independently associated with cognitive decline among older adults and to

describe cognitive trajectories according to their impairment pattern.

Material and methods: We used data from totals of 13123, 11417 and 21265 respondents

aged 50+ at baseline from the Health and Retirement Study (HRS), the English Longitudinal

Study of Ageing (ELSA) and the Survey of Health, Ageing and Retirement in Europe

(SHARE), respectively. We performed growth curve analysis to identify cognitive

trajectories and a joint model was used to deal with attrition problems in longitudinal ageing

surveys.

Results: Respondents with a single sensory impairment had lower episodic memory score

than those without sensory impairment in HRS (β=-0.15, p<0.001), ELSA (β=-0.14,

p<0.001), and SHARE (β=-0.26, p<0.001). The analysis further shows that older adults with

dual sensory impairment in HRS (β=-0.25, p<0.001), ELSA (β=-0.35, p<0.001), and SHARE

(β=-0.68, p<0.001) remembered fewer words compared to those with no sensory impairment.

The stronger associations between sensory impairment and lower episodic memory levels

were found in the joint model which accounted for attrition.

Conclusions: Hearing and/or vision impairment is a marker for the risk of cognitive decline

that could inform preventative interventions to maximise cognitive health and longevity.

Further studies are needed to investigate how sensory markers could inform strategies to

improve cognitive ageing.

Keywords: sensory impairment, cognitive ageing, longitudinal analysis, older people

Keypoints:

Older adults with single or dual sensory impairment (hearing and/or vision) were able

to recall fewer words than those without sensory impairment.

The association between sensory impairment and lower episodic memory levels were

stronger after the attrition considered in the joint model.

The cognitive trajectories of older adults with no sensory impairment followed

curvilinear shapes, while those of older adults with single and dual sensory

impairments showed more linear shapes.

Introduction

Maintaining cognitive function in later life has become a public health priority as the burden

imposed by dementia in the ageing population has increased more rapidly than that of most

other diseases [1]. In the past ten years, the disability-adjusted life-years (DALYs) for

Alzheimer’s disease and other dementias have grown by one-third, from 17,905 DALYs per

1000 population in 2005 to 23,779 DALYs per 1000 population in 2015 [2]. Understanding

cognitive change in later life is thus crucial, as cognitive decline is a hallmark of dementia

[3].

Hearing and vision sensory impairments among older people may contribute to the risk of

cognitive decline and pathological impairments including dementia. Numerous cross-

sectional studies, but only a limited number of longitudinal studies have demonstrated an

association between sensory impairment and cognitive decline [4-6]. A study using six waves

of the Berlin Aging Studies reported moderate size correlations between visual/hearing acuity

and sensory decline [7]. An Australian longitudinal study found that the significant

association between sensory impairment and cognitive declines diminished after adjusting for

age and other potential confounding factors [8].

In this study, our aim is to identify how sensory impairment relates to trajectories of cognitive

decline among older adults, accounting for attrition in our models. Longitudinal studies,

especially in the field of ageing, suffer from attrition as their respondents tend to selectively

dropout because of either death or worsening health function [9]. Unless those dropouts can

be assumed to be ‘missing at random’, ignoring them can result in a bias in the analysis.

Describing and understanding trajectories of cognitive decline and how these trajectories

relate to sensory impairment may offer insight into the dynamics of cognitive decline and

identify opportunities for intervention to maximise cognitive function and longevity in older

age.

Data and measures

Data

This study used three international surveys of ageing: the Health and Retirement Study

(HRS), the English Longitudinal Study of Ageing (ELSA), and the Survey of Health, Ageing

and Retirement in Europe (SHARE). These three surveys provide information on the

personal, socio-economic and health circumstances of individuals aged 50+ [10-12]. The

baseline interviews of HRS were conducted with community dwelling adults in the US in

1992. Respondents who entered a nursing home after the baseline interview are retained in

the sample and were interviewed if possible in the following waves. The first wave of ELSA

data was collected in 2002, while the SHARE study was started in 2004. So far there are

twelve, seven, and five waves available for HRS, ELSA, and SHARE, respectively. In this

study, we use data from a similar time range: 2002-2014 for HRS and ELSA and 2004-2014

for SHARE. We restrict our attention to the nine countries with complete data in SHARE:

Austria, Belgium, Denmark, France, Germany, Italy, Spain, Sweden, and Switzerland.

Dependent variable: episodic memory score

We used the measure of cognitive function available in all the three surveys, namely, episodic

memory. Episodic memory represents general cognition [13] and is more age-sensitive than

other episodic measures [14]. Salthouse et al. [13] concluded that memory and cognitive

control variables appear to have a common mechanism. In all surveys, the interviewer reads

a list of ten common nouns to the respondent and then asks the respondent to recall as many

words as possible from the list in any order twice: immediately after the respondent heard the

complete list (immediate recall) and at the end of the cognitive function module (delayed

recall). The immediate recall represents the ability of respondents to learn or store new

information, whereas delayed recall is the ability to recall that information after a period of

distraction from that information. Including only one test does not represent the memory

function particularly well. Prior study showed evidence of a high correlation between

immediate and delayed recall (r=0.70-0.75) in the first wave of ELSA [15]. We calculated the

raw scores as the total of the number of correct words of immediate recall and delayed recall,

with a maximum score of 20 [16, 17].

Main independent variable: sensory function

Sensory impairment was measured using self-reported hearing and vision quality. In all three

surveys, hearing data was collected using the following self-reported measure of overall

hearing function: ‘Is your hearing [using a hearing aid as usual] excellent (1), very good (2),

good (3), fair (4) or poor (5)?’. In HRS and ELSA, self-reported vision quality was collected

in all seven waves using the question: ‘Is your eyesight [using glasses or corrective lens as

usual] excellent (1), very good (2), good (3), fair (4) or poor (5)?’. In SHARE, we used the

two self-reported measures of visual function that are present in all waves: distance eyesight

and reading eyesight. For each question, responses were recoded into two categories by

combining the responses excellent, very good and good into one and collapsing fair and poor

vision into a second category. We defined sensory impairment as having fair and poor

hearing and/or vision [18]. As SHARE has two questions to measure vision quality, we

classified its respondents as having poor vision if they had poor distance eyesight and/or

reading eyesight. We then categorised the results as follows: no impairment, single (vision or

hearing) sensory impairment, and dual sensory impairment (impairment in both senses).

Covariates

We included demographic (age and gender), socio-economic information, health behaviour

and the presence of chronic diseases as determinants of cognitive function. As with other

health functions, cognitive abilities are shaped by social determinants of health [19],

including education, marital status, and wealth. We categorised respondents’ education into

less than high school as reference, high school, and college. Marital status was classified as

married and not married. We used quintiles of income by country each year to measure

wealth, using the poorest quintile as the reference [16]. The social capital index is obtained in

two steps. Firstly, we sum the activities in the month leading up to the day that the respondent

was interviewed, such as performing voluntary work; helping friends, neighbours and

relatives; and taking part in a community organisation. Secondly, we standardise those

summation scores.

Cognitive function is affected by health behaviour (smoking status, drinking behaviour, and

physical activities) and health status [20]. In terms of smoking status, we categorised

respondents as current smokers, past smokers and non-smokers. We classified respondents as

drinking regularly if they consumed alcohol ≥ 5 days/week [16] and as having vigorous

physical activities if they did those activities at least once a week. Functional status was

measured using the Activity Daily Living (ADL) scale. The ADL scale used five items of

self-performance in HRS (dressing, walking across the room, bathing, getting in or out of

bed, and eating). The self-performance of toilet use was added in SHARE and ELSA. We

used the measures of the presence of depression (measured by Center for Epidemiologic

Studies Depression (CES-D) in HRS and ELSA and Euro-D in SHARE) and chronic diseases

(the sum of several chronic diseases: diabetes, stroke, lung diseases and cancer) to capture the

health status of the respondents.

Statistical analysis

We compared the characteristics of respondents in the first wave of HRS, ELSA and SHARE

separately according to the presence of sensory impairment using Kruskal-Wallis one-way

analysis of variance for numerical variables and ordinal chi-square tests for categorical

variables. We then performed multilevel growth curve models separately for each survey to

predict the level of cognition in Wave 1 and subsequent changes in cognition over further

waves, all dependent on the Wave 1 age cohort. The multilevel models in this report consist

of repeated observations nested within individuals. We included demographic and socio-

economic determinants, lifestyle factors and the presence of chronic diseases as the

covariates.

To deal with biases due to drop-out, we used a joint model where the random effects

influence both episodic memory and attrition, and given these, episodic memory and attrition

are independent. The joint model in this study had two parts: the growth curve model and the

survival model (with sex, age polynomial of degree three and the random intercepts from the

growth curve model) [16]. We compared the results of sensory impairment model with that of

the joint model to assess the robustness of the sensory impairment model to attrition [21]. The

statistical analyses were conducted using Stata 14.0 and Latent Gold 5.1.

Results

(Table 1 is about here)

The characteristics of respondents in the first wave of the HRS (2002), ELSA (2002) and

SHARE (2004) studies are presented in Table 1. On average, respondents in the first wave of

HRS, ELSA and SHARE were able to memorise 10, 9.4, and 7.9, respectively. The

descriptive analysis of respondents by the presence of sensory impairment was presented in

Appendix 1-3. The proportion of respondents with dual sensory impairment was highest

among SHARE respondents (7.7%), followed by HRS (7%) and ELSA (5.8%). Respondents

with dual sensory impairment were likely to perform less well on the cognitive tests, to have

lower income, to be older, less educated, have more ADL dependencies, and be less socially

engaged, than those with no or single impairment.

To avoid confounding relationships and to arrive at net associations, results from the growth

curve and joint models separately for each survey are presented in Table 2. It shows that

sensory impairment has a significant negative relationship to cognitive ability. Focusing on

the growth curve model, respondents with single and dual impairments performed less well

than those with none. This negative association was considerably greater after attrition was

taken into account in the joint model. For example, the episodic memory levels of HRS

respondents with single and dual sensory impairments were lower by -0.15 and -0.25 words

before accounting for attrition and those associations were larger in the joint model (by -0.56

and -1.14 words). The same findings are confirmed in ELSA and SHARE.

(Table 2 is about here)

Apart from age and sensory impairment coefficients, several socio-demographic

characteristics and other confounders showed stable significant associations with cognitive

function in all three surveys. Being female, having attained a higher level of education, being

employed, and being relatively wealthy were associated with better cognitive abilities. The

social capital index and physical activities showed a positive and significant association with

higher cognitive function. Functional status as measured by ADL and depression had a

significant negative association with cognitive function in all surveys.

(Figure 1 is about here)

Figure 1 illustrates the predicted baseline episodic memory score and trajectory over time for

respondents with different levels of sensory functions (hearing and visual). After controlling

for the covariates, in general, the cognitive trajectories in all surveys took on a curvilinear

shape. Respondents with better hearing function were able to recall more words in all

surveys. Similarly, the predicted value of episodic memory scores of respondents with better

vision function is higher than those with poor function. The presence of sensory impairment

had a negative correlation with cognitive trajectories. The cognitive trajectories of older

adults with no sensory impairment followed curvilinear shapes, while those of older adults

with dual sensory impairment showed a consistent trajectory of cognitive decline after the

age of 50.

Discussion

Our findings, from three longitudinal surveys of ageing, showed that both single and dual

sensory impairments in older adults were independently associated with accelerated rates of

decline in cognitive abilities. The association is stronger among those with dual sensory

impairment. Our findings extend the discussion in the literature on the relationships between

sensory impairment and cognitive function by estimating the trajectories of summary

cognitive scores for older adults with different levels of sensory function using a large

longitudinal multinational sample. We found that, in general, the trajectory of cognitive

function took on a curvilinear shape. However, when we separated the cognitive trajectories

according to the presence of sensory impairment, the cognition of respondents with sensory

impairment declined faster than that of respondents with no sensory impairment. Crucially,

the trajectories of respondents with dual sensory impairment took the shape of a more linear

decline after the age of 50.

The patterns of association between hearing and visual sensory impairment with cognitive

decline described in the present study partially support previous longitudinal studies [22, 23].

However, studies in Australia [24] and the Netherlands [25] reported that decline in hearing

function was not associated with cognitive ability. A key limitation across these prior studies

is that those studies did not account for attrition and lead to bias. The strengths of our study

include the fact that it performed a joint model to deal with that limitation. Our analysis using

the joint model shows stronger negative relationships between sensory impairment and levels

of episodic memory.

Additionally, the 10-year follow-up duration in the ELSA data included in our study

facilitates a fuller examination of trajectories of cognitive ageing. The literature has proposed

several possible relationships: cognitive decline precedes sensory impairment through the

reduction of the cognitive resources that are available for sensory perception (the ‘cognitive

load on perception’ hypothesis), sensory impairment causes cognitive loss, possibly through

the effects of sensory impairment on social isolation (the ‘sensory deprivation’ hypothesis),

and the presence of third factors causes both declines (the ‘common cause’ hypothesis) [26].

Our analysis showed that respondents with dual sensory impairment joined fewer social

activities than those with single impairment and no impairment. It is possible that sensory

impairments may impact cognitive trajectories via facilitating or limiting social activity as the

magnitude of the association between sensory impairment and cognitive function in our study

declines after we include the social capital index and other demographic covariates. The

common-cause hypothesis should also be considered. Our findings do not allow for a

conclusive distinction between hypotheses, and the hypotheses are not mutually exclusive. It

may be the case that all the possibilities described are valid to some extent. Further studies

are needed to disentangle them. For instance, hypotheses that suggest an impact of sensory

function on cognition may be tested by identifying changes in cognitive trajectories following

sensory remediation (e.g. cataract surgery and provision of a hearing aid).

The use of self-reported measures of sensory function is a limitation of this study because

self-report measures may under-estimate rates of impairment and do not provide estimation

of the severity of the impairment. However, self-reported measures of sensory function are

commonly used in epidemiological studies [27], and previous studies support the validity of

self-reported measures of both vision impairment [18] and hearing impairment [28]. The

second limitation of this study is that the episodic memory score does not define all the

cognitive abilities of older adults and other abilities have different rates of decline with

advancing age. However, this measure has been known to have a good validity and to relate

to the every-day activities of older people [29]. The last limitation is the observational design

of the study, which means that the relationship between sensory impairment and cognition

may be affected by unmeasured predictors of cognitive ageing, such as social network,

employment status and dietary intakes and the findings should be interpreted with caution.

In conclusion, those with sensory impairment are at a greater risk of developing cognitive

impairment and may show a faster trajectory of cognitive decline that those without sensory

impairment. A recent publication using ELSA found that respondents with the most

advantaged trajectory of episodic memory had an odds ratio more than five times less than

those with the most disadvantaged trajectory after allowing for established risk factors for

dementia [30]. Further studies are needed to investigate how sensory markers could inform

strategies to prevent cognitive decline. Strategies may include hearing and vision

rehabilitative intervention in combination with healthy ageing interventions to promote social

engagement, physical activity and positive health behaviours.

Funding source:

This work was supported by SENSE-Cog project. This project has received funding from the

European Union’s Horizon 2020 research and innovation programme under grant agreement

No 668648.

Conflict of interest:

The authors have no financial or any other kind of personal conflicts with this paper.

References

1. Prince M, Wimo A, Guerchet M, Ali GC, Wu YT, Prina M. World Alzheimer Report

2015. The global impact of dementia. An analysis of prevalence, incidence, cost & trends,

London: Alzheimer’s Disease International, 2015.

2. Kassebaum NJ, Arora M, Barber RM, Bhutta ZA, Brown J, Carter A et al. Global,

regional, and national disability-adjusted life-years (DALYs) for 315 diseases and injuries

and healthy life expectancy (HALE), 1990-2015: a systematic analysis for the Global Burden

of Disease Study 2015. The Lancet 2016; 388(10053): 1603-1658.

3. Borson S, Lori F, Bayley PJ, Boustani M, Dean M, Lin P et al. Improving dementia

care: the role of screening and detection of cognitive impairment. Alzheimer’s & Dementia

2013; 9(2): 151-159.

4. Lindenberger U, Baltes PB. Sensory functioning and intelligence in old age: a strong

connection. Psychology and Aging 1994; 9(3): 339-355.

5. Baltes PB, Lindenberger U. Emergence of a powerful connection between sensory

and cognitive functions across the adult life span: a new window to the study of cognitive

aging? Psychology and Aging 1997; 12(1): 12-21.

6. Tay T, Wang JJ, Kifley A, Lindley R, Newall P, Mitchell P. Sensory and cognitive

association in older persons: findings from an older Australian population. Gerontology 2006;

52(6): 386-394.

7. Lindenberger U, Ghisletta P. Cognitive and sensory declines in old age: gauging the

evidence for a common cause. Psychology and Aging 2009; 24(1): 1-16.

8. Hong T, Mitchell P, Burlutsky G, Liew G, Wang JJ. Visual impairment, hearing loss

and cognitive function in an older population: Longitudinal findings from the Blue Mountains

Eye Study. PloS ONE 2016; 11(1): e0147646.

9. Chatfield MD, Brayne CE, Matthews FE. A systematic literature review of attrition

between waves in longitudinal studies in the elderly shows a consistent pattern of dropout

between differing studies. Journal of Clinical Epidemiology 2005; 58(1): 13-19.

10. Sonnega A, Weir D. The Health and Retirement Study: A public data resource for

research on aging. Open Health Data 2014; 2(1).

11. Marmot M, Oldfield Z, Clemens S, Blake M, Phelps A, Nazroo J et al. English

Longitudinal Study of Ageing: Waves 0–6, 1998–2013. 2013.

12. Börsch-Supan A, Brandt M, Hunkler C, Kneip T, Korbmacher J, Malter F et al. Data

resource profile: the Survey of Health, Ageing and Retirement in Europe (SHARE).

International Journal of Epidemiology 2013; 42(4): 992-1001.

13. Salthouse TA, Pink JE, Tucker-Drob EM. Contextual analysis of fluid intelligence.

Intelligence 2008; 36(5): 464-486.

14. Dere E, Easton A, Nadel L, Huston JP. Handbook of Episodic Memory. Vol. 18.

2008: Elsevier.

15. Allerhand M, Gale CR, Deary IJ. The dynamic relationship between cognitive

function and positive well-being in older people: A prospective study using the English

Longitudinal Study of Aging. Psychology and Aging 2014; 29(2): 306.

16. Tampubolon G. Cognitive ageing in Great Britain in the new century: Cohort

differences in episodic memory. PloS ONE 2015; 10(12): e0144907.

17. Weir D, Faul J, Langa K. Proxy interviews and bias in the distribution of cognitive

abilities due to non-response in longitudinal studies: a comparison of HRS and ELSA.

Longitudinal and Life Course Studies 2011; 2(2): 170-184.

18. Whillans J, Nazroo J. Assessment of visual impairment: the relationship between self-

reported vision and ‘gold-standard’ measured visual acuity. British Journal of Visual

Impairment 2014; 32(3): 236-248.

19. Marmot M, Allen J, Goldblatt P, Boyce T, McNeish D, Grady M et al. The Marmot

review: Fair society, healthy lives. The Strategic Review of Health Inequalities in England

Post-2010, 2010.

20. Yaffe K. Chronic medical disease and cognitive aging: Toward a healthy body and

brain. Oxford: Oxford University Press, 2013.

21. Rabbitt P, Diggle P, Smith D, Holland F, Mc Innes L. Identifying and separating the

effects of practice and of cognitive ageing during a large longitudinal study of elderly

community residents. Neuropsychologia 2001; 39(5): 532-543.

22. Yamada Y, Denkinger MD, Onder G, Henrard J, Van der Roest HG, Finne-Soveri H,

et al. Dual sensory impairment and cognitive decline: The results from the Shelter Study.

Journals of Gerontology Series A: Biomedical Sciences and Medical Sciences 2015; 71(1):

117-123.

23. Anstey KJ, Hofer SM, Luszcz MA. A latent growth curve analysis of late-life sensory

and cognitive function over 8 years: Evidence for specific and common factors underlying

change. Psychology and Aging 2003; 18(4): 714-726.

24. Anstey KJ, Luszcz MA, Sanchez L. Two-year decline in vision but not hearing is

associated with memory decline in very old adults in a population-based sample. Gerontology

2001; 47(5): 289-293.

25. Gussekloo J, de Craen AJM, Oduber C, Van Boxtel MPJ, Westendorp RGJ. Sensory

impairment and cognitive functioning in oldest-old subjects: the Leiden 85+ Study. The

American Journal of Geriatric Psychiatry 2005; 13(9): 781-786.

26. Wayne RV, Johnsrude IS. A review of causal mechanisms underlying the link

between age-related hearing loss and cognitive decline. Ageing Research Reviews 2015; 23:

154-166.

27. Liljas AEM, Wannamethee SG, Whincup PH, Papacosta O, Walters K, Iliffe S et al.

Socio-demographic characteristics, lifestyle factors and burden of morbidity associated with

self-reported hearing and vision impairments in older British community-dwelling men: a

cross-sectional study. Journal of Public Health 2015; 38(2): e21-e28.

28. Gibson WK, Cronin H, Kenny RA, Setti A. Validation of the self-reported hearing

questions in the Irish Longitudinal Study on Ageing against the Whispered Voice Test. BMC

Research Notes 2014; 7(1): 361.

29. McArdle JJ, Smith JP, Willis R. Cognition and economic outcomes in the Health and

Retirement Survey, The National Bureau of Economic Research, 2009.

30. Tampubolon G, Nazroo J, Pendleton N. Trajectories of general cognition and

dementia in English older population: An exploration. European Geriatric Medicine

2017.

List of tables

Table 1: Sample characteristics at baseline by survey

HRS ELSA SHAREN=13,123 N=11,417 N=21,265

Mean episodic memory 10 (3.4) 9.4 (3.5) 7.9 (3.6)Mean age 67.8 (10.5) 64.8 (10) 64.8 (10.1)Female, % 58.2 54.4 54.5Married/cohab, % 68.1 67 72.2EducationLess than high school, % 19.4 34.5 50.8High school, % 33.9 17.8 30.6Some college, % 46.6 47.5 18.4Employed, % 38.1 35.9 27.2Smoking behaviourNon-smoker, % 41.9 33.7 28.3Past smoker, % 44.2 48.5 17.8Current smoker, % 13.7 17.7 53.7Drink daily or almost daily, % 12.3 28.2 26.3Mean depression score* 1.3 (1.8) 1.5 (1.9) 0.2 (0.4)Mean ADL (SD) 0.24 (0.73) 0.4 (0.9) 0.2 (0.7)Vigorous exercise, % 24.6 27.5 47.2Number of comorbidities 1.1 (1.1) 0.6 (0.8) 0.6 (0.8)Note: * Depression scores are CESD in HRS and ELSA and Euro-D in SHARE

Table 2: Growth curve and joint models predicting episodic memory scores

HRS ELSA SHAREGrowth curve Joint Growth curve Joint Growth curve Joint

Age 0.4 (0.01)* -0.13 (0.00)* 0.42 (0.01)* -0.12 (0.00)* 0.33 (0.01)* -0.13 (0.00)*Age2 -0.00 (0.00)* 0.00 (0.00)* -0.00 (0.00)* 0.00 (0.00)* -0.3 (0.00)* 0.02 (0.00)*Sensory function, ref: No impairmentSingle impairment -0.15 (0.02)* -0.56 (0.01)* -0.14(0.02)* -0.55 (0.02)* -0.26 (0.01)* -0.58 (0.01)*Dual impairment -0.25 (0.04)* -1.14 (0.02)* -0.35 (0.05)* -1.3 (0.03)* -0.68 (0.03)* -1.61 (0.02)*Female 1.21 (0.03)* 1.05 (0.01)* 0.79 (0.04)* 0.66 (0.01)* 0.97 (0.02)* 0.74 (0.01)*Married -0.02 (0.03) -0.07 (0.02)* -0.01 (0.03) -0.03 (0.02) 0.08 (0.02)* 0.05 (0.01)*Education, ref: Less than high schoolHigh school 0.78 (0.04)* 0.65 (0.03)* 1.58 (0.04)* 1.27 (0.03)* 1.67 (0.02)* 1.52 (0.02)*Some college 1.58 (0.04)* 1.3 (0.03)* 1.94 (0.05)* 1.54 (0.03)* 2.5 (0.02)* 2.26 (0.02)*Wealth, ref: 1st quartile (poorest)2nd quartile 0.22 (0.03)* 0.28 (0.03)* 0.07 (0.04) 0.24 (0.04)* 0.16 (0.02)* 0.21 (0.02)*3rd quartile 0.35 (0.03)* 0.44 (0.03)* 0.23 (0.04)* 0.49 (0.04)* 0.2 (0.02)* 0.32 (0.02)*4th quartile 0.46 (0.03)* 0.59 (0.03)* 0.33 (0.04)* 0.7 (0.04)* 0.27 (0.02)* 0.41 (0.02)*5th quartile (richest) 0.59 (0.03)* 0.76 (0.03)* 0.4 (0.04)* 0.99 (0.04)* 0.35 (0.02)* 0.51 (0.02)*Employed 0.28 (0.02)* 0.22 (0.02)* 0.13 (0.03)* -0.02 (0.03) 0.35 (0.02)* 0.4 (0.02)*Social capital 0.2 (0.01)* 0.32 (0.01)* 0.28 (0.01)* 0.24 (0.01)* 0.37 (0.02)* 0.47 (0.00)*Smoking, ref: Non-smokerCurrent smoker -0.12 (0.04)* -0.46 (0.02)* -0.06 (0.05) -0.5 (0.02)* 0.39 (0.02)* 0.2 (0.01)*Past smoker -0.01 (0.03) 0.00 (0.01) 0.00 (0.04) 0.02 (0.01) 0.34 (0.02)* 0.45 (0.01)*Drink daily or almost daily 0.12 (0.03)* 0.21 (0.03)* 0.07 (0.03) 0.2 (0.03)* -0.2 (0.02)* -0.19 (0.02)*Vigorous physical activity 0.15 (0.02)* 0.1 (0.02)* 0.14 (0.02)* 0.23 (0.02)* 0.31 (0.01)* 0.39 (0.01)*ADL -0.23 (0.01)* -0.26 (0.01)* -0.08 (0.01)* -0.12 (0.01)* -0.25 (0.01)* -0.38 (0.01)*Depression score -0.05 (0.00)* -0.07 (0.00)* -0.06 (0.00)* -0.08 (0.00)* -0.44 (0.01)* -0.49 (0.02)*Number of comorbidities -0.11 (0.01)* -0.01 (0.01)* -0.07 (0.01)* -0.03 (0.01) -0.06 (0.01)* -0.02 (0.01)*Constant -1.58 (0.49)* 17.07 (0.07)* -2.47 (0.6)* 17 (0.08)* -1.68 (0.4)* 15.2 (0.05)*Note: Reported are coefficients (standard errors). Sig.: *: significant at 1%.

List of figures

Figure 1: The predicted trajectories of summary cognitive scores of the HRS, ELSA and

SHARE participants by the presence of sensory impairment

16