If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Send correspondence and reprint requests to Savannah Kiah Hui Siew, B.Sc. Hons., Department of Psychological Medicine, Yong Loo Lin School of Medicine, National University of Singapore, Tahir Foundation Building (MD1), Research Dry Laboratories (North), #12-01, 12 Science Drive 2, Singapore, Singapore 117549
Department of Psychological Medicine, Yong Loo Lin School of Medicine, National University of Singapore (SKHS, RM, JY), Singapore, SingaporeAcademic Development Department, Duke-NUS Medical School (RM), Singapore, Singapore
What is the primary question addressed by this study? What are some predictive factors of older adults’ anxiety symptoms during the COVID-19 pandemic lockdown?
•
What is the main finding of this study? Higher social isolation and lower quality of life at baseline significantly predicted more anxiety symptoms at follow-up and the covariates of age, gender and socioeconomic status were not significant in the model. The social isolation and anxiety symptoms relationship was unidirectional while the quality of life and anxiety symptoms relationship was bidirectional.
•
What is the meaning of the finding? Older adults who were previously socially isolated and had a lower quality of life are particularly vulnerable to the psychological impacts of the COVID-19 lockdown and resources and support should be channeled to this at-risk group.
ABSTRACT
Objective
The COVID-19 lockdown could see older adults facing increased anxiety levels due to social isolation. Additionally, the lockdown could be more difficult for those with lower Quality of Life (QoL). We aim to understand predictive factors of older adult's anxiety symptoms during the lockdown as it is a main psychological concern of COVID-19.
Methods
Four hundred eleven participants (Mage = 68.95, S.D. = 5.60) completed questionnaires at two time points — before the pandemic and during the lockdown period. Cross-lagged analysis was carried out on two structural equation models – social isolation and anxiety symptoms, and QoL and anxiety symptoms.
Results
Baseline social isolation was associated with more anxiety symptoms at follow-up. However, baseline anxiety symptoms were not associated with social isolation subsequently. For QoL and anxiety symptoms, the relationship was bidirectional.
Conclusion
Older adults who were previously socially isolated and had a lower QoL are particularly vulnerable to the negative psychological impacts of the COVID-19 lockdown.
Coronavirus disease (COVID-19) is a global health crisis that has taken many lives and pushed countries to introduce lockdowns to curb its spread. Singapore was placed under lockdown from 7th April to 1st June. All non-essential workplaces, schools, and food establishments were closed while social gatherings were prohibited. From 2nd to 19th June, some workplaces reopened and restricted visitations to parent's or grandparent's houses were allowed. The present study was conducted from 11th May to 19th June.
Most restrictive measures were to protect older adults since they were more vulnerable to becoming severely ill from contracting COVID-19. Additionally, the pandemic can adversely affect their mental health. Anxiety levels are a main concern, with rates three times higher during the pandemic,
as uncertainty and fear regarding transmissibility can cause one to worry and face increased symptoms. This could stem from a fear of mortality or economic worries.
Lockdowns could worsen the consequences of increased anxiety levels. Social disconnectedness and isolation can worsen one's mental health, contributing to increased anxiety symptoms.
Feelings of loneliness or the lack of perceived support, since the presence of social support relationships is crucial to having good mental health, could be responsible for these consequences.
Another potential predictor of increased anxiety level is Quality of Life (QoL). We define QoL as a subjective, holistic measure of experiences, relating to health and functioning, both directly and indirectly.
Lower QoL reflects less satisfaction with different areas of life, and fewer resources to weather the lockdown. Thus, those with lower QoL initially could be more vulnerable to negative psychological effects. The correlation between QoL and anxiety symptoms could also act in both directions.
We sought to understand the predictors of anxiety levels among older adults during the COVID-19 lockdown. This is timely and helps us understand the consequences of this pandemic beyond its direct health risk for the older adult population. To date, most research in this period has been narrative or cross-sectional in nature,
making it difficult to establish the direction of influence between a pair of correlated outcomes. The ‘predictors’ could plausibly manifest as consequences instead of antecedents. Hence, we used cross-lagged longitudinal structural equation models (SEM) to study the directional effects between predictors and anxiety levels.
We hypothesize that older adults who were more socially isolated before the pandemic will experience higher anxiety levels during the lockdown. Secondly, we postulate that QoL and anxiety symptoms are bidirectionally related; those who reported poorer QoL before the pandemic will experience higher anxiety levels during the lockdown, and vice versa.
METHODOLOGY
Detailed methodology and statistical parameters used are presented in the supplementary material (see Text, Supplemental Digital Content 1).
Participants and Procedures
411 participants completed two waves of data collection, before COVID-19 and during the lockdown. The average length of follow-up was 15.4 months (SD=6.2). Informed consent was taken before participation. Ethical approval was granted from the National University of Singapore's Institutional Review Board (Ref No. S-20-118E).
Measures
The 5-item short form of the Geriatric Anxiety Inventory (GAI) was used to measure levels of anxiety. Each question has a yes or no response, corresponding to scores of 0 or 1, respectively. Higher scores represent higher levels of anxiety symptoms.
Participants’ perceived social isolation was determined by the 6-item Friendship Scale (FS). It is scored on a 5-point Likert Scale, and items 1, 3, and 4 were reverse coded such that lower scores indicate higher levels of social isolation.
The 13-item World Health Organization QoL assessment for older adults was used to determine the participant's QoL. It is scored on a 5-point Likert scale, and higher scores indicate a higher QoL.
Statistical Analysis
The SEM was carried out in R 4.0.1. Individual items of each questionnaire were fitted into their respective latent factor models. Cross-lagged models involving GAI and FS, and GAI and QOL were analyzed separately while controlling for age, sex and socioeconomic status as covariates.
RESULTS
Descriptive Statistics and Pair-wise Correlations
Participants were mainly female (65%), Chinese (95%), with a mean age of 69 years and 13 years of education. Other descriptive statistics are presented in the supplementary material (see Table, Supplemental Digital Content 2). Figure 1A. displays the Pearson correlations among all variables used in this study.
FIGURE 1a) Pearson correlation heatmap of all variables used in the study. Robust maximum-likelihood estimation of the b) social isolation and anxiety, and c) quality of life and anxiety cross-lagged models with age, sex, and socioeconomic status included as covariates. Straight lines represent regression paths and figures represent standardized regression coefficients. Figures in parentheses represent standard errors of the estimates. The statistical significance of the regression coefficients were determined via its t-statistic. None of the covariates were significant in both models. The final longitudinally invariant model had satisfactory fit indices in b) where Comparative Fit Index = .903, Root Mean Square Error of Approximation = .044, Standardized Root Mean square Residual = .054, Degree of Freedom = 259 and in c) where Comparative Fit Index = .903, Root Mean Square Error of Approximation = .042, Standardized Root Mean square Residual = .056, Degree of Freedom = 688. For model fit criteria used in our study, see Data, Supplemental Digital Content 1. For other model fit indices, see Table, Supplemental Digital Content 3. GAI, Geriatric Anxiety Inventory; FS, Friendship Scale; QOL; Quality of Life Scale. * ρ < 0.05, ** ρ < 0.01, *** ρ < 0.001.
All subsequent model's fit indices are presented in the supplementary material (see Table, Supplemental Digital Content 3).
Anxiety and social isolation
The results suggested that both the unconstrained and longitudinal invariant model, where unstandardized item factor loadings are held constant across time, had unsatisfactory fit indices. However, the difference between both models was not significant (∆χ2=4.84, ∆df=9, p = 0.848). The longitudinal invariant model was preferred and used subsequently. Correlating items 3 (“someone to share my feelings with”) and 4 (“found it easy to make contact with people”) of the FS significantly improved model fit to satisfactory indices (∆χ2=46.81, ∆df=2, p <0.001). Henceforth, the longitudinal-invariant model with correlated errors was used for subsequent cross-lagged analyses.
The cross-lagged effect was significant; higher FS scores at baseline were associated with lower GAI scores at follow-up after controlling for baseline GAI scores (Fig. 1B). However, the reverse effect was not statistically significant; baseline GAI scores did not significantly predict lower FS scores at follow-up after controlling for baseline FS scores. As mentioned, higher FS scores reflect less social isolation.
Anxiety and quality of life
The unconstrained and longitudinal invariant models both had good fit indices. The χ2 test of the difference between both models was not significant (∆χ2=11.24, ∆df=16, p = 0.794). The preferred longitudinal-invariant model was used for subsequent analyses.
Higher QoL scores at baseline were associated with lower GAI scores at follow-up after controlling for baseline GAI scores (Fig. 1c); higher GAI scores at baseline were also associated with lower QoL scores at follow-up after controlling for baseline QoL.
DISCUSSION
We hoped to understand the directional influences of QoL and social isolation on anxiety levels among older adults during the COVID-19 lockdown. Overall, our findings revealed a unidirectional effect of social isolation on anxiety symptoms and a bidirectional effect of QoL and anxiety symptoms. These directional effects are small-to-medium in magnitude but are of potential interest as it teases out the sequential nature between social isolation and quality of life, and anxiety symptoms.
The unidirectional effect of social isolation on anxiety symptoms can be due to the perceived lack of social support which is essential to mental well-being.
The interpersonal model of emotion also posits that as emotional regulation develops socially in early attachment relationships, it evolves such that social isolation impedes emotional regulation and often associates with negative affect among adults.
Cognitive mechanisms can also be implicated, where social isolation leads to more negative cognitions and heightened sensitivity to threats which can increase anxiety symptoms.
Higher QoL before the pandemic could be indicative of various protective factors that could result in lower levels of anxiety during the lockdown. This includes being more satisfied with their living condition and personal relationships, which made the lockdown experience more tolerable. On the contrary, for those with higher baselines of anxiety symptoms, the lockdown could have exacerbated their fear and worries, resulting in diminished QoL. Hence, these could explain our findings.
The low correlation between anxiety levels at baseline and follow-up may seem unusual. However, this is expected due to the floor effect as most participants scored 0 at baseline since this is a mainly healthy community-dwelling cohort of older adults.
Our results imply that anxiety level is an important mental health factor to target in older adults. It has shown to affect subsequent overall well-being in a fear-inducing pandemic setting. The availability of psychological services is thus essential, to help cater to those with higher anxiety levels. Moreover, our findings show that this can be partially attributed to social isolation and low QoL, both important factors to be addressed.
Besides, social distancing measures can result in unwanted consequences as they increase social isolation. Many among the older demographic are new to using technology to maintain social connections. Those who are unable to adapt to this new form of interaction may emerge from the lockdowns with poorer mental health. Resources should be channeled to ensure they are able to maintain social connections during this period. This can be done by teaching them the use of infocommunication applications such as video calls. Regular check-ins regarding their QoL can help to pinpoint areas of dissatisfaction early. This way, changes can be made and assistance rendered.
Some limitations include the self-report scales used which could subject our results to social desirability and retrospective recall biases. Secondly, our sample's racial makeup is overwhelmingly from the majority racial group – Chinese, which may not accurately account for the experiences faced by other racial groups here in Singapore. Lastly, our lockdown might not be as severe as those in other countries which may limit the generalizability.
AUTHOR CONTRIBUTIONS
JY is the Principal Investigator of the study and conceptualized and designed the study and made substantial contributions to the data analysis, interpretation and preparation of the manuscript. RM made substantial contributions to the design of this study. SKHS made substantial contributions to the design of the study, acquired the data and carried out data analysis and interpretation. The first draft of the manuscript was prepared by SKHS. JY, RM and SKHS revised subsequent drafts and approved the final version to be published. All listed authors have contributed significantly to the manuscript and consent to their names on the manuscript.
DISCLOSURE
The authors would like to thank the research team consisting of research assistants (Ms. Lim Xin Ying, Ms. Madeline Han, Ms. Yap Ai Che) and research nurse (Ms. Ng Siew Yee).
This work was supported by Research Donations from Kwan Im Thong Hood Cho Temple and Lee Kim Tah Holdings Pte Ltd, under the Mind Science Centre, Department of Psychological Medicine, National University of Singapore. The authors report no conflicts with any product mentioned or concept discussed in this article.
Previous presentation: Society of Behavioral Health, Singapore's Virtual Scientific Meeting 2020; 7 November 2020.
00287-6&id=gr1.jpg){kind=link}