Impact of Social Isolation on Mental Health Amid COVID-19 Pandemic: A Nationwide Survey

Background: Although social isolation is known to limit the spread of a pandemic, the impact of mental health for such measures is yet unknown. In this cross-sectional study, we investigated the impact on mental health among different age groups due to social isolation during the ongoing COVID-19 pandemic in Dhaka, Bangladesh. Methods: We conducted a carefully designed cross sectional survey on mental health that was disseminated widely by way of email, personal contact and social media to subjects aged between 11 and >70 years. For our analysis we stratied data into three distinct groups: children/young adults (11-40), middle age (40-60) and older adult age (> 60) groups. 3214 respondents answered the survey. Bonferroni corrected Chi-square tests were used to nd signicant relationships between the demographic groups and mental health related variables. Results: We observed a high percentage of insomnia (79%) in old age respondants compared to children/young adults (61%) and middle age (66%) groups, suggesting that ‘age’ is signicantly associated (p= 3.8 X 10 -06 ; odds ratio (OR) = 2.34) with ‘insomnia’. Respondents who were retired also reported a higher prevalence (73%) of insomnia (p = 2.79 X 10 -8 ) compare to employed individuals. A higher level of mental stress (84%) was observed in middle aged respondents followed by old adult (71%) respondents (p=0.001). Signicantly higher rates (p = 5.08 X 10 -27 ; OR = 2.06) of mental stress were detected in people with preexisting comorbidities compared with the healthy group. The old age participants were less familiar with the concept of social isolation and 54% of old age participants had a negative perception towards social isolation compared to children/young adults (12%) and middle-aged (7%) respondents. Conclusion: Our results indicate an association between age and mental stress concomitant on the COVID-19 social isolation policy in Bangladesh. Social isolation increased insomnia and mental stress, particularly in old age and middle age group. Moreover, these older age groups also tended to have a negative perception of the COVID-19


Introduction
The outbreak of SARS-CoV-2 from the Hubei province, Wuhan, China, has now affected the entire world with over 135 million con rmed cases and approximately 3 million deaths (1,2). Following the outbreak, The World Health Organization (WHO) characterized this spread as a pandemic on the 11th March 2020 (3). At the beginning of the pandemic, several containment measures had been taken by many countries to contain viral shedding and to slow down the spread of this transmissible disease. Since no preventive treatment or established vaccine was available at that time, a social isolation strategy was adopted by almost all countries across the globe to stop the transmission. The term isolation refers to a full or partial closure of public/private institutions, to minimize close proximity human interactions.
Bangladesh reported the rst con rmed COVID-19 case on 8th March 2020(4). The country observed a slow increase in COVID-19 cases and death toll. As a mitigation policy, the Govt. of Bangladesh restricted the regular movement or gathering of people to minimize viral shedding. While mitigating transmission of the virus and COVID-19 was the principal concern, mental health aspects of the pandemic have received relatively little attention. According to the WHO, anxiety and distress are expected to increase among people as a natural psychological consequence of the challenging condition brought about by the pandemic (5). Studies have shown that low mood, irritability and insomnia can also be precipitated by isolation(6). As expected, therefore, highlevels of anxiety have been reported globally during the COVID-19 outbreak (7). Moreover, the increase in psychological distress is highly correlated with the rapid spread of SARS-CoV-2 infection and its associated mortality(8).
Although it was reported that people in isolation may experience mental symptoms (9), it may be compounded by uncertainty, clearly another characteristic of the pandemic on a global level. (10). which can vary according to the age, sex or medical comorbidities. Studies have also shown that older people are susceptible to frequent symptoms of insomnia (11), anxiety or depression (12) during any type of isolation. These are of particular concern among older people, who, as a group, experience disproportionatephysical health problems such as heart disease, diabetes, and hypertension.(13-16) COVID pateints who are above 55 years have three times the mortality rate (17) and 20 percent of deaths occurred in patients above 60 years of age (18). In addition, compare to other age groups, the mortality is 6 times greater for patients above 80 years of age (19). Likewise in Bangladesh 44 percent of fatality cases are above 60 years of age(20) although the young people were more infected (21).
Several studies have demonstrated the impactof lockdown related isolation on mental health during COVID-19 outbreak in different countries (22,23). These studies have helped national-level policy development, such as undertaking the necessary steps to minimize mental health complications such as depression and suicide. A recent study has altered to the real possibility of COVID-19 related suicide in Bangladesh (24). However, a systematic study of the mental health impact of COVID-19 on a large representative sample of individuals across different socio-demographic groups has not been investigated in Bangladesh or, indeed, in other developing countries. The aim of this study is to identify the impact of social isolation on mental health among different socio-demographic groups in Bangladesh.

Study Design
The study was approved by the Institutional Review Board (IRB) of North South University (2020/OR-NSU/IRB-No.0701). We conducted an online survey from 18th April to 20th May 2020 among the Bangladeshi population. An anonymous online questionnaire form was sent through social media (Facebook) to people of all groups population (aged 11 to > 70 years) inside and outside Dhaka in Bangladesh. Based on COVID severity, mortality and considering life expetancy, these people were further categorized in three different classes: below 40, 40-60 and above 60 years of age. Partially completed forms were excluded such that only people who answered all questions were included in the analysis.

Construction of the Questionnaire
We developed a questionnaire to assess mental wellbeing during social isolation due to COVID-19. A structured questionnaire comprising 21 questions was used in this study (Supplementary Table 1). For privacy and anonymity, during the construction of the questions, we have not used any variables (name, phone number, ID, address etc.) that can be used to identify the participants. The questionnaire included a wide variety of demographic variables such as age, gender, occupation, income, urbanversus rural location, comorbidity and questions related to activity, quarantine knowledge, viewpoints, and the perceived impact of the quarantine on mental health. To ensure face and content validity, the questionnaire was sent to a group of expert researchers who reviewed the survey for content accuracy, clarity, and comprehensiveness. The reviewers include expert from neuropsychiatry and public health background provided feedback on the questionnaire related to mental health symptoms. In this study, 9,400 people received the Google forms and 3,214 (34.19%) respondents returned completed forms. We included the participants in the rage of 11 to > 70 years of age, and requested that children who are between the age of 11 and 17 complete the questionnaire with their parent's assistance. Participants who partially lled the forms were excluded from this study. The participants were categorized into three age groups: 'children/young adults' (11-40 years), 'middle age' (40-60 years) and 'older adult' (above 60 years). These age groups are based upon known severity of sCOVID-19 symptoms.

Statistical Analysis
Summary statistics, plots and association tests were performed using R packages. For each sociodemographic group, we quanti ed the baseline summary statistics. Association test was performed between participant response and participant groups for the same question. We have not conducted association tests between variables obtained from two independent questions. We analyzed the data by using the Chi-square (χ2) test to explore the association between participant demographic and mental wellbeing related variables,with appropriate Bonferroni correction applied to the signi cance threshold.

Social Isolation and Insomnia
Out of three groups, insomnia was reported by 61% (1487), 66% (345) and 79% (172) of children/young adults, middle age and older adult groups, respectively (Fig. 1A). The more frequent reporting insomnia is signi cantly correlated with older age (χ2 = 29.7; p = 3.88 X 10 − 6 ; odds ratio (OR) = 2.34) (Table 1b) who suffered insomnia 2.3 times higher compared to the children/young adult group. Due to theknown association between insomnia and mental symptoms(25, 26), we studied the prevalence of mental symptoms in the three age groups and observed 77% (1882/2440), 84% (434/517) and 70% (154/218) of children/young adults, middle age and older adult groups, respectively, reported experiencing mental wellbeing due to social isolation ( Fig. 1A-B). A signi cant (χ2 = 18.4; p = 0.001) increase (1.5 times) of mental wellbeing was observed in middle age participants compared to the children/young age group. A similar age-related i trend was observed between the groups in experiencing di culties in decision making in their daily activities (Fig. 1C). Table 1b Age-dependent and gender-dependent association study data. We next categorized the self-reported mental distress into three symptom clusters, a) anger and irritability, b) anxiety and depression and c) fear and panic. Children/young age group reported (29%) experiencing the most anger and irritability compared to other age groups (18% and 7% for middle age and older adult groups, respectively) (Fig. 2). The older age group reported (62%) experiencing the highest rates of anxiety and depression compared to the other age groups (51% and 50% for children/young adult and middle age group, respectively). Finally, the middle age group reported (30%) experiencing the highest rate of fear and panic compared to the other age groups ( 17% and 29% for children/young and older adult group, respectively) (Fig. 2).
We next examined the occurrence of insomnia in relation to vocation. We found that retired individuals reported higher rates of insomnia (72%) compared to students or those in active employment ( Supplementary Fig. 3). (χ2 = 36.17; p = 2.79 X 10 − 8 ). The prevalence of mental health symptoms was higher among the participants who ran businesses (χ2 = 11.67; p = 0.005) compared with other groups (Supplementary Fig. 3 Fig. 5-6). We also observed participants who live in an urban setting reported a signi cantly (χ2 = 21.623; p = 0.000006) higher incidence of insomnia compared to their rural counterparts. Additionally, mental symptoms were 1.30 times higher among those living in Dhaka. Next, we analyzed the sleeping time duration among in different aged groups. Consistent with our insomnia association, old adult group showed a reducedsleep duration (3-4 hours) compared to other groups (χ2 = 392.7; p = 3.7016 X 10 − 77 ) ( Supplementary Fig. 9).

Correlation of Age with the COVID-19 Social Isolation
We asked respondants whether they felt that social isolation was a helpful strategy to protect against COVID-19 transmission ( Fig. 3 and Table 1b). 93% (2,277/2,443) of children/young adults, 88% (456/520) of middle aged and 46% (100/219) old adult respondents agreed with the social isolation strategy (Fig. 3A). In comparison between the three groups the older adult participants showed (16.3 times) a more negative perception regarding social isolation strategy. The same group of participants were less favorable of social isolation and more fearful of coming back to normal life. In contrast children/young adults were the most enthusiastic (63%) regarding going back to normal life as well as being equally comfortable extending (72%) the social isolation if required (Fig. 3D-E). Finally, the children/young aged group were also the most knowledgeable regarding the COVID-19 policies.

Gender association with COVID-19 related social isolation
On the whole there were no gender differences across all questionnaire items, with the exception of daily living impact (χ2 = 25.095; p = 5.4 X 10 − 07 ) ( Fig. 4 and Supplementary Fig. 10) indicating that perception about daily life along with quarantine might gender wise.

Discussion
To our knowledge, this is the rst study conducted in Bangladesh to examine the impact of the COVID-19 pandemic on the mental health of a large group of individuals. In particular, it examined the effect of social isolation on mental distress and insomnia. Social isolation is known to be associated with insomnia (29) and other psychological symptoms, including anger and irritability,, anxiety and depression (14,30,31) Our ndings of high rates of insomnia are consistent with similar studies conducted during other infectious outbreaks(6, 32). Our ndings also reported a more severe impact on the mental health and well-being of older respondents. Similarly, the older age group reported a higher prevalence of decision making di culties.. Such cognitive di culties are known to be related to insomnia due to the impact of sleep deprivation on cognitive acuity (27).
The consistency of our results in relation to the association of both insomnia and mental health symptoms in older age is perhaps unsurprising. In general terms, the rates of physical and mental health symptoms are known to be signi cantly higher in this group. The social isolation, therefore, will have merely compounded this existing tendency to mental health symptoms, including insomnia. Anxiety is a key risk factor for insomnia, with the relationship between the two also echoed in our own results. Indeed, the older age group also reported greater di culties with decision making, another key indicator of raised levels of anxiety. It is interesting that the older age group also had a more negative attitude towards lockdown, presumably re ecting, in part at least, that the younger age groups were better informed about policies, and better able to understand the rationale. Of course, a negative attitude towards social isolation may also re ect their insight into the impact of lockdown on their mental wellbeing. We also found that people living in urban Dhaka reported higher rates of insomnia compare to those living in other areas. This is very likely secondary to anxiety about risk of spread in the more densely populated areas.
Clearly our results have policy implications, both in terms of the need to more closely monitor mental health symptoms among the older population during the current and possible future lockdowns, as well as the need to more fully inform this group about the reasons for lockdown and time scale to allay their anxiety. There may be a need, for example, for a public health initiative that targets the older sector of society so that they understand the policy and, importantly, have the opportunity to ask questions about it.
There are inherent limitations in cross sectional based survey data. For example, we were unable to cover the full gamut of mental health symptoms to avoid a burdonsome questionnaire. Moreover, the wording of questions may have been misunderstood, and ideally the responses would be corroborated by interview or follow up. Clearly, with the numbers included in this study such an undertaking would be very challenging. We did seek consultation from experts in the design of our questionnaire, however, and therefore believe that we have captured the most important symptoms in a reliable and valid manner.
Another key issue is the generalizability of our results, given that our contact methods did not necessarily capture all sectors of society. We do believe that urban Dhaka was well-represented among respondents, for example, but we are uncertain how generalizable our results are to the rest of the country. A survey capturing data from rural groups in a country such as Bangladesh is very challenging, however. In summary, therefore, our analysis indicates that older adults in Bangladesh experience greater mental health issues due to social isolation.. Policymakers should take into account how to best provide a comprehensive strategy to monitor mental health and offer counseling and education.

Availability of data and materials
The datasets analysed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that the research was conducted in the absence of any commercial or nancial relationships that could be construed as a potential con ict of interest. Mosaic plot illustrating the effect of COVID-19 quarantine on the mental health of the respondents with respect to their age. (A) Relation between COVID-19 quarantine and sleep disorder among children and young adults, middle age and old population. (B) Impact of quarantine on mental stress among the respondents. (C) Effect of social isolation on the decision making and thinking capabilities of the children and young adults, middle age and old respondents. The plot was built using R where area of boxes of the plot are proportional to the number of respondents belonging to each group. 'Age' is de ned on the x-axis and 'response' on the y-axis, where 'no' is color-coded in red and 'yes' is color-coded in yellow.

Figure 2
Impact of COVID-19 quarantine on the type of psychological problem faced by children and young adults, middle age and old population. The mosaic plot was built using R where area of boxes of the plot are proportional to the number of respondents belonging to each group. 'Age' is de ned on the x-axis and 'type of associated problem' on the y-axis. Anger or Mood Swing, Fear or Panic, Anxiety and Depression, None are color coded in red, yellow, blue and black, respectively.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. SupplementaryMaterial.docx