key: cord-0977059-jt3vjsqe
authors: Kivi, Marie; Hansson, Isabelle; Bjälkebring, Pär
title: Up and about: Older adults’ wellbeing during the COVID-19 pandemic in a Swedish longitudinal study
date: 2020-06-30
journal: J Gerontol B Psychol Sci Soc Sci
DOI: 10.1093/geronb/gbaa084
sha: 6ad823dcbeb0afdb85e383cd414df9c75d0a0e4b
doc_id: 977059
cord_uid: jt3vjsqe

OBJECTIVES: To investigate early effects of the COVID-19 pandemic related to 1) levels of worry, risk perception, and social distancing; 2) longitudinal effects on wellbeing; and 3) effects of worry, risk perception, and social distancing on wellbeing. METHODS: We analyzed annual changes in four aspects of wellbeing over five years (2015–2020): life satisfaction, financial satisfaction, self-rated health, and loneliness in a subsample (n=1071, aged 60–71) from a larger survey of Swedish older adults. The 2020 wave, collected March 26–April 2, included measures of worry, risk perception, and social distancing in response to COVID-19. RESULTS: 1) In relation to COVID-19: 44.9% worried about health, 69.5% about societal consequences, 25.1% about financial consequences; 86.4% perceived a high societal risk, 42.3% a high risk of infection, and 71.2% reported high levels of social distancing. 2) Wellbeing remained stable (life satisfaction and loneliness) or even increased (self-rated health and financial satisfaction) in 2020 compared to previous years. 3) More worry about health and financial consequences was related to lower scores in all four wellbeing measures. Higher societal worry and more social distancing were related to higher wellbeing. DISCUSSION: In the early stage of the pandemic, Swedish older adults on average rated their wellbeing as high as, or even higher than, previous years. However, those who worried more reported lower wellbeing. Our findings speak to the resilience, but also heterogeneity, among older adults during the pandemic. Further research, on a broad range of health factors and long-term psychological consequences, is needed.

M a n u s c r i p t

The COVID-19 pandemic is a major threat to public health, and older adults are particularly vulnerable for severe health consequences (Le Couteur et al., 2020 , Verity et al., 2020 Zhou et al., 2020) . As illustrated by the COVID-19 Government Response Stringency Index (Hale, et al., 2020) ), responses to COVID-19 have varied along a continuum, with some imposing lockdowns, while others rely on guidelines and recommendations to slow the spread. Close to one end of this continuum, Sweden has largely remained open, relying on voluntary measures and imposing comparably few mandatory restrictions. The impact on daily life has nevertheless been huge, especially for those over 70. The stated aim of many government responses to COVID-19 has been to protect older adults and other risk groups. However, lockdowns and arbitrary age restrictions may also put additional strain on older adults (Ayalon et al., 2020) . The psychological consequences related to the pandemic itself, but also to the varying governmental responses to the pandemic, are still relatively unknown. Due to the heterogeneity of responses, a broad range of studies, from a variety of places, will be needed to fully understand the consequences for older adults.

The present study investigates wellbeing in a Swedish sample during the early days of the COVID-19 pandemic. The first case of COVID-19 in Sweden was reported February 1 (World Health Organization (WHO), 2020a). On March 25, Swedish authorities issued statements recommending social distancing, but did not enact a general lockdown. An additional, yet voluntary, recommendation to "shelter in place" was also directed to adults aged 70 and older (Public health agency of Sweden, 2020a Sweden, , 2020b Sweden, , 2020c . At the time of data collection, Sweden had 28 confirmed deaths per million (12 th highest death rate) and the number of infected doubled every 3 days (WHO, 2020b).

In times of crisis, protecting lives and physical health is paramount; nevertheless, accurate assessment and understanding of psychological health should guide governmental response (Drury et al., 2013) . WHO has warned of lower wellbeing during the pandemic, especially among older adults (WHO, 2020c) . In the U.S., worry has risen and wellbeing is at M a n u s c r i p t a 12 year low (Witters & Harter, 2020) . At the same time, one recent study suggests that older adults are less vulnerable to psychological distress and loneliness during the pandemic (Losada-Baltar et al., 2020) . In order to assess the true impact of longitudinal studies are needed to compare wellbeing before, throughout, and after the pandemic has ended.

Our study uses data from the longitudinal HEalth, Aging and Retirement Transitions in Sweden (HEARTS) project to investigate the early psychological effects of COVID-19 in a sample of older adults. The HEARTS data provides a unique opportunity to investigate longitudinal effects on wellbeing over a period of five years (2015-2020). We specifically aim to: 1) determine levels of worry, risk perception, and social distancing in relation to ; 2) investigate longitudinal effects on life satisfaction, financial satisfaction, selfrated health, and loneliness; and 3) quantify the effects of worry, risk perception, and social distancing on wellbeing.

Since 2015 (from March to June each year), the HEARTS study has conducted an annual survey in a population-based sample (N=5913) of older adults born 1949-1955 (age 60-66 at baseline; for more info on HEARTS please see Lindwall et al, 2017) .

In this study we used a subsample of HEARTS participants who responded during the first seven days of data collection in 2020 (March 26, 10 am, until April 2, 10 am; n=1071).

The 2020 questionnaire included an additional set of questions related to the COVID-19 pandemic. These questions assessed levels of worry in relation to health, financial, and societal consequences, perceived risk of societal consequences, likelihood of being infected, and social distancing. Further, participants indicated if they or anyone in their immediate surroundings had been diagnosed with COVID-19. Longitudinal data on life satisfaction (Diener et al., 1985) , financial satisfaction, self-rated health, and loneliness (Russell et al., A c c e p t e d M a n u s c r i p t 1980) were included to account for changes in relation to previous measurements. Age, gender, education, and retirement status were included to control for sociodemographic differences. For measurement details on study items, see Supplementary Table S1 .

Data was analyzed with linear mixed-effects models using the lme4 (Bates et al., 2015) package in R (R Core Team, 2019). Deviations in life satisfaction, financial satisfaction, self-rated health, and loneliness were analyzed by comparing the scores of the 2020 assessment to the scores across previous waves (2015-2019). The model included a linear slope (measurement year) centered on the 2020 wave, as well as a dummy variable separating the 2020 measurement (1) from previous years (0). Effects of worry, risk perception, and social distancing were evaluated in a second model by regressing the four wellbeing measures on each of the five predictor variables. The four wellbeing measures were analyzed separately; age, gender, education, and retirement status were included as control variables in all models. The alpha level was restricted to .0125 using the Bonferroni correction to account for multiple testing.

Respondents to the 2020 survey had a mean age of 68.1 years (SD=2.0; range 65-71), 47.3% were women, 52.4% had tertiary education. The sample comprises more men and a larger proportion of individuals with tertiary education compared to the baseline sample of HEARTS participants (45.4% and 40.2%, respectively). About 2% (n=22) of participants reported that they or someone in their immediate surrounding had (or previously had) a confirmed COVID-19 infection. An additional 6.6% (n=71) said they had reason to believe that they or someone in their immediate surroundings had been infected. Descriptive statistics on the study variables can be found in Table 1 ; bivariate correlations among the four wellbeing measures (r=.24-.51) are presented in Supplementary Table S2. A c c e p t e d M a n u s c r i p t --Insert Table 1 about here-

The results showed that 44.9% (i.e., proportion of individuals with a score ≥4) worried about their own or others' health, 69.5% worried about societal consequences, and 25.1% worried about financial consequences related to COVID-19. The majority (86.4%) reported high societal risks, 42.3% perceived the risk of being infected as high (≥3), and 71.2% reported engaging in social distancing. Figure 1 shows the proportion of individuals with high scores on level of worry, risk perception, and social distancing for individuals below and above age 70 (i.e., defined as a risk group by Swedish authorities; Public health agency of Sweden, 2020c). Participants age 70 and older (n=333) reported less financial worry (<70=27.7%, ≥70=20.0%; ꭓ 2 =6.73, p=.010), more social distancing (<70=68.9%, ≥70=79.3%; ꭓ 2 =11.81, p=.001), and a lower risk of being infected (<70=45.7%, ≥70=35.1%; ꭓ 2 =10.01, p=.002) compared to those age 65-69. No significant differences were found with respect to health (<70=44.5%, ≥70=47.1%; ꭓ 2 =0.55, p=.46), societal worry (<70=71.7%, ≥70=66.7%; ꭓ 2 =2.51, p=.11), or societal risks (<70=86.6%, ≥70=87.9%; ꭓ 2 =0.24, p=.63). were related to lower scores on life satisfaction (β=-0.12/-0.17, p<.001), financial satisfaction (β=-0.10/-0.25, p≤.001), and self-rated health (β=-0.16/-0.08, p≤.008), and higher scores on loneliness (β=0.09/0.13, p≤.007). On the contrary, more worry about societal consequences was related to higher financial satisfaction (β=0.09, p=.003) and less loneliness (β=-0.08, p=.009). Higher scores on social distancing were related to higher satisfaction with life (β=0.11, p<.001) and finances (β=0.08, p=.006). Societal risk and perceived risk of being infected were not significantly related to any of the four outcome variables (all p>.0125). For all estimates from the linear mixed-effects models see Supplementary Table S3 .

This study investigated the effects of the COVID-19 pandemic on wellbeing in a Swedish sample of older adults. Our first aim was to determine the level of worry, risk perception, and social distancing in response to COVID-19. Older adults aged 65-71 perceived high societal risks related to the COVID-19 pandemic; the majority reported having reduced close physical contact with others (social distancing), and were concerned A c c e p t e d M a n u s c r i p t about adverse effects on social structures and the world economy. At the same time, the majority were not particularly worried about their personal economy, nor for themselves or loved ones being infected by the virus; they perceived the risk of infection for themselves or loved ones as rather low. However, those who were given stricter recommendations (in our sample those 70 and older) regarded their risk of becoming infected as lower than those 65-69 years old. This might be explained by a higher degree of social distancing among those 70 and older, thus reducing the risk of infection. In sum, few participants reported being personally affected, while the majority generally considered COVID-19 a major threat to health, security, and wellbeing in Sweden and worldwide (Drury et al., 2013) . The stricter recommendations to those 70 or older seem to have resulted in more social distancing in this group, but not more worry (Ayalon et al., 2020) .

Our second aim was to investigate the longitudinal effects of the COVID-19 pandemic on wellbeing. Contrary to the expected negative impact of the pandemic (WHO, 2020c), across a five-year period, we found no negative effect of COVID-19 on wellbeing. On the contrary, self-rated health was as high as five years ago, and financial satisfaction was higher than in any of the previous years. To illustrate these numbers, in 2019, 60% of participants rated their health as good or very good, while the corresponding number in 2020 was 69%. In 2019, 70% were satisfied or very satisfied with their financial situation, while in 2020 75% were satisfied or very satisfied. Life satisfaction and loneliness showed no deviation in 2020 compared to previous years. Although subjective health declines with increasing age, in 2020 that trend was broken, and subjective health was rated as high as in 2015. Based on these findings, we conclude that COVID-19, so far, has had few adverse effects on wellbeing among older adults in Sweden. Instead, most people rated their wellbeing just as high as or even higher than they did in previous years.

One possible explanation for this positive effect is a hesitance among older adults in Sweden to identify as 'older', as there have been reports of older citizens rejecting the notion A c c e p t e d M a n u s c r i p t that they belong to a risk group (Ekroth, 2020) . However, this seems to be contradicted by our data showing high levels of social distancing, increasing with age. Another possible explanation is the effect of contrasts (Tversky, 1977) . Circumstances that might have seemed less satisfying last year now seem more satisfying compared to the potential negative effects of COVID-19. Further, it is possible that the increased wellbeing in our sample is due to the contrast between the relative freedom of movement in Sweden and other countries' strict regulations (Schwarz & Strack, 1999) .

Our third aim was to quantify the effects of level of worry, risk perception, and social distancing on wellbeing. While levels of worry were generally moderate, COVID-19 may have caused some participants to worry to an extent that negatively affected their wellbeing.

Those who worried more about negative health and financial consequences reported lower wellbeing, indicating that socioeconomic dimensions might influence the effects of the pandemic (Cappelen et al, 2020) . Surprisingly, worrying more about adverse effects on social structures and the world economy were connected to higher wellbeing. Lastly, those who practiced more social distancing reported higher satisfaction with life and finances. Overall, this suggests that while many older adults handle the distress of the pandemic well, older adults are a heterogeneous group, and some older adults worry to the extent that their wellbeing suffers (Losada-Baltar et al., 2020) . Important limitations include the short time frame of data collection, the narrow agerange of the participants, and potential problems in isolating worry and behavior caused by the pandemic from other events and circumstances. The 2020 data used in this study was collected during one week in the early days of the pandemic's progression in Sweden and continued follow-ups are needed to evaluate if the effects persist over time. However, the time frame is also one of the foremost strengths of this study, as we have captured effects that later studies will miss. A major strength of the study is the HEARTS database, which presents a unique opportunity to analyze changes longitudinally.

A c c e p t e d M a n u s c r i p t

In sum, Swedish older adults were still 'up and about' during the early part of the COVID-19 pandemic; while the majority practiced social distancing, they also rated their wellbeing as high as, or even higher than they did five years prior. Generally low levels of worry and high wellbeing might be a consequence of the relatively few restrictions in Sweden, but studies from other countries will be needed for comparison. It is also important to note that older adults are a heterogeneous group; while most had high wellbeing, those who worried more about the health and financial effects of COVID-19 had lower wellbeing.

Hence, determining ways to reduce worry will be important to mitigate lower wellbeing during the pandemic. Finally, the findings in this study should not be taken as an endorsement of any particular governmental response. Although subjective wellbeing is important, it is only one component of health. Older adults' high wellbeing in this study should not be used to disregard other components of health. Hence, more research, on a broad range of health indicators, is needed to monitor and counteract the consequences of COVID-

A c c e p t e d M a n u s c r i p t 

None reported.

This data can be made available upon request and in accordance with applicable laws. For further information about accessibility of data, contact hearts@psy.gu.se

This study was not preregistered. M a n u s c r i p t A c c e p t e d M a n u s c r i p t Figure 2 

Aging in Times of the COVID-19 Pandemic: Avoiding Ageism and Fostering Intergenerational Solidarity

Fitting Linear Mixed-Effects Models Using lme4

Opinion | What Do You Owe Your Neighbor? The Pandemic Might Change Your Answer. The New York Times

The Satisfaction With Life Scale

Psychological disaster myths in the perception and management of mass emergencies

Många äldre struntar i risken att bli smittade

Variation in Government Responses to COVID-19. Version 6.0. Blavatnik School of Government Working Paper

COVID-19 Through the Lens of Gerontology

Psychological Health in the Retirement Transition: Rationale and First Findings in the HEalth, Ageing and Retirement Transitions in Sweden (HEARTS) Study

Association of Self-perceptions of Aging, Personal and Family Resources

Protect yourself and others from spread of infection

With the ongoing community transmission, how can we protect people over the age of 70?

If you are 70 or over -limit close contact with other people

/faktablad-covid-19-70-ar-engelska.pdf R Foundation for Statistical Computing

The revised UCLA Loneliness Scale: Concurrent and discriminant validity evidence

Reports of subjective well-being: Judgmental processes and their methodological implications

Features of similarity

Estimates of the severity of coronavirus disease 2019: A model-based analysis. The Lancet Infectious Diseases

Life Ratings Plummet to 12-Year Low

Mental health and psychosocial considerations during the COVID-19 outbreak

Novel Coronavirus(2019-nCoV), Situation Report -12

Clinical course and risk factors for mortality of adult inpatients with COVID-19 in

A retrospective cohort study. The Lancet

Descriptive statistics of the study variables by year for the HEARTS subsample of 1071 persons surveyed in 2020 Measurement Year

The PI for the HEARTS program is professor Boo Johansson. The research group behind We want to thank Boo Johansson for his comments on this work and Reghan Borer for editorial support.

A c c e p t e d M a n u s c r i p t