key: cord-1050051-dtltempn
authors: Osterrieder, A.; Cuman, G.; Pan-ngum, W.; Cheah, P. K.; Cheah, P.-K.; Peerawaranun, P.; Silan, M.; Orazem, M.; Perkovic, K.; Groselj, U.; Schneiders, M. L.; Poomchaichote, T.; Waithira, N.; Asarath, S.-a.; Naemiratch, B.; Ruangkajorn, S.; Skof, L.; Kulpijit, N.; Mackworth-Young, C. R. S.; Ongkili, D.; Chanviriyavuth, R.; Mukaka, M.; Cheah, P. Y.
title: Economic and social impacts of COVID-19 and public health measures: results from an anonymous online survey in Thailand, Malaysia, the United Kingdom, Italy and Slovenia
date: 2020-10-27
journal: nan
DOI: 10.1101/2020.10.26.20209361
sha: 789d795bc0a3349ee09d3aa1083f8ce4a07b3166
doc_id: 1050051
cord_uid: dtltempn

In the absence of a vaccine and widely available treatments for COVID-19, governments have relied primarily on non-pharmaceutical interventions to curb the pandemic. To aid understanding of the impact of these public health measures on different social groups we conducted a mixed-methods study in five countries ('SEBCOV' - Social, ethical and behavioural aspects of COVID-19). Here we report the results of the SEBCOV anonymous online survey of adults. Overall, 5,058 respondents from Thailand, Malaysia, the United Kingdom, Italy and Slovenia completed the self-administered survey between May and June 2020. Post-stratification weighting was applied, and associations between categorical variables assessed. Among the five countries, Thai respondents appeared to have been most, and Slovenian respondents least, affected economically. Overall, lower education levels, larger households, having children under 18 in the household, being 65 years or older and having flexible/no income were associated with worse economic impact. Regarding social impact, respondents expressed most concern about their social life, physical health, and mental health and wellbeing. There were large differences between countries in terms of voluntary behavioural change, and in compliance and agreement with COVID-19 restrictions. Overall, self-reported compliance was higher among respondents reporting a high understanding of COVID-19. UK respondents felt able to cope the longest and Thai respondents the shortest with only going out for essential needs or work, with 60% and 26% respectively able to cope with 29 days or longer. Many respondents reported seeing news that seemed fake to them, the proportion varying between countries, and with education level and self-reported levels of understanding of COVID-19. Our data showed that COVID-19 public health measures have uneven economic and social impacts on people from different countries and social groups. Understanding the factors associated with these impacts can help to inform future public health interventions and mitigate their negative consequences on people's lives.

country languages. The self-administered online survey was set up using the 'JISC Online surveys' 129 platform 21 . 130

The survey questions were pilot-tested with 25 people from participating countries, and revised 132 accordingly based on feedback. In addition, the Bangkok Health Research Ethics Interest Group, a 133 public involvement group set up by the Mahidol Oxford Tropical Medicine Research Unit (MORU) 22 , 134 discussed the study and the survey questions in a dedicated virtual meeting. Selected questions were 135 tested using an adapted cognitive testing technique using the "thinking out loud" approach 23 , and the 136 collaborative virtual sticky notes board 'Padlet' 24 . 137

Adults of any age residing in Thailand, Italy, Malaysia, United Kingdom (UK) or Slovenia at the time 139 of the study were eligible to take part. Participants needed to be able to use a computer or smart phone 140 to access the survey and provide online consent to participate. 141

The survey was open from 1 st May to 30 th June 2020 (1 st -30 th June for Slovenia due to late start). 142

Participants were recruited using a combination of approaches: snowball sampling through personal 143 and professional networks (via email, social media and messenger apps, mailing lists, and 144 organisations such as the Medical Chamber 25 in Slovenia); a polling company 26 in Thailand; and 145 through promoted posts on Facebook. Facebook allows users to 'boost' posts to selected demographic 146 audiences for a small fee, so that the post appears on their Facebook newsfeed 27 . To achieve more 147 balanced responses in the categories of gender, education level and geographic distribution, promoted 148 Facebook posts were targeted at people with primary or lower/secondary education in UK and 149

Malaysia; potential participants in Wales, Scotland and Northern Ireland in the UK; and at men in the 150 UK and Italy. 151

Each country aimed to recruit a minimum sample of 600 respondents, exceeding the 40-200 153 respondents recommended for a mixed-methods study 28 . A minimum sample size of 600 respondents 154 is adequate to estimate the prevalence of a response assuming a 50% prevalence rate, with 95% 155 confidence and with a precision of 4%. The 50% prevalence is the standard assumption for precision 156 sample size calculations when the true prevalence is not available, as this gives the highest sample 157 size for a binomial distribution for a desired level of precision. 158 . CC-BY 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20209361 doi: medRxiv preprint

To simplify analysis, answers in the following categories were combined as follows: "slightly 160 agree/highly agree" were combined into one "agree", category, and "slightly/strongly disagree" 161 responses into one "disagree" category (Suppl . Tables 23-27) . To understand the distribution of the 162 basic demographic variables in the respondent sample, the observed frequencies and sample 163 characteristics are reported using unweighted percentages (Suppl. Table 1 ). The characteristics for the 164 rest of the variables are presented using the observed survey frequency counts followed by weighted 165 percentages (Suppl. Tables 2-37). Post-stratification weighting was used to align the composition of 166 the respondents' sample with the known distribution of the whole population's characteristics, 167 reducing sampling error. Weights were computed considering three stratifying variables that were 168 available from population census data from each country 29 , namely, gender, age and education level. 169

Weights were obtained as the ratio between the proportion of each possible combination of the three 170 variables in the whole country population and the correspondent proportion in the respondent sample. 171

Survey data was analysed using Stata 15.0 software 30 . Frequency counts and percentages were used to 172 summarise categorical data. Associations between categorical variables were assessed using Pearson's 173

Chi-squared test. P-values have been provided in the tables and considered statistically significant 174 below the two-sided alpha=0.05 level. All p-values presented in the tables are for global tests of 175 significance. Practical significance was taken into account when interpreting differences in the results. 176

At the time of the inception of this study, governments in Thailand, Malaysia, Italy, the UK and 178

Slovenia had initiated public health measures, using varying degrees of "lockdowns" to curb the 179 pandemic. Figure 1 shows is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint In order to understand the economic impacts of COVID-19, respondents who had been working 204 before the pandemic (paid or unpaid work) were asked whether COVID-19 had created any work-205 related inconvenience for them. Overall, 56% of respondents said that they experienced loss of 206 earnings, 44% reduction of working hours, 36% closure of workplace and 14% job loss (Fig. 2 , Suppl. 207 Table 2 ). Seventy five percent reported that they continued to work during COVID-19. Of all 208 respondents, 53% expressed financial concerns, and 32% worried about professional/career 209 progression. Our results indicated that the most affected country was Thailand, with 85% of 210 respondents reporting loss of earnings, 23% loss of job, and 86% expressing financial concerns 211 (Suppl. Table 2 ). Slovenian respondents reported the least severe economic impacts e.g. 30% reported 212 loss of earnings, 3% reported loss of job, and 28% had financial concerns. 213

To investigate the impact of public health measures on different social groups, we analyzed responses 214 based on gender, level of education, age group, household size, whether respondents lived with 215 children under 18 years old, and income type. 216

Overall, there were no significant differences between male, female and respondents who identified as 217 'other/prefer not to say', and who had been working before COVID-19, in terms of loss of earnings, 218 loss of job, reduction of working hours and closure of workplace (Fig. 2 , Suppl. Table 3 ). Overall, 219 fewer women continued to work during COVID-19 (71% women vs 78% men; p=0.010). The trend 220 was similar at country level, except for Malaysia (73% women versus 67% men; Suppl. Table 3) . 221

Overall, 65% of respondents with primary and secondary education who had been working before 222 COVID-19 reported a loss of earnings, compared to 38% of respondents with tertiary education 223 (p<0.001; Fig. 2 is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint Overall, larger households and having children under 18 in the household appeared to be associated 242 with worse economic impacts (Fig. 2 , Suppl. Tables 6 and 7). Overall, 67% of respondents whose 243 household included 6 people or more reported loss of earnings (compared to 54% of households with 244 1-5 people; p=0.013), and 23% reported loss of job (compared to 13%; p=0.009; Suppl. Table 6 ). 245

Respondents with children reported a higher loss of earnings compared to respondents without 246 children (62% versus 53%; p=0.005), and higher job loss (18% versus 12%; p=0.008; Suppl. Table 7) . 247

Analysing by country, respondents living with children appeared to be particularly affected in 248

Thailand and Malaysia. 249

We also analysed responses according to three types of income: fixed income (e.g. fixed salary, 250 benefits or pension), flexible income (e.g. contract, freelance), and other/no income ( Fig. 2 ; Suppl. 251 Table 8 ). We did not ask for amount of income. Overall, respondents with fixed income were less 252 affected economically than those with flexible or other/no income. Of the latter only 38% reported 253 loss of earnings, compared to 81% of respondents with flexible income and 69% of respondents with 254 other/no income (p<0.001). Only 8% of people with fixed income had lost their job, compared to 22% 255 with flexible income and 27% with other/no income (p<0.001). At country level, the trends were 256 similar (Suppl. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20209361 doi: medRxiv preprint

We asked respondents if they were concerned about the following areas of life if advised no physical 262 contact/not allowed to go out/allowed to go out only for essential needs: caring responsibilities, 263 physical health, recreational pursuits, sports, mental health and wellbeing, living arrangements, 264 infrastructure (e.g. access to transport, internet), social, and religious and spiritual needs/aspects 265 (Suppl . Table 9 ). Overall, respondents expressed most concern about their social life (64%), their 266 physical health (59%), and their mental health and wellbeing (58%). This trend was largely similar in 267 individual countries, except for Thailand, where caring responsibilities attracted the most concern 268 (62%); Malaysia, where 58% were concerned about religion and spirituality; and Slovenia, where 269 65% of people worried about recreational aspects. In general, there were no major differences 270 between gender, age groups, education level, household size, living with children or income type 271 (Suppl . Tables 10-15 ). Overall, those who were most worried about caring responsibilities were 272 women (52%, versus 42% men, p<0.001; Suppl. We asked respondents how many days they could cope with not going out except for essential 277 needs/work, with answer options ranging from one to 59 days or more. In total, 44% of respondents 278 said that they could cope for 29 days or longer (Suppl . Table 16 ). However, coping time varied 279 significantly between countries (p<0.001): in the UK, 60% of people felt they would be able to cope 280 for 29 days or longer, whereas in Thailand, only 26% of respondents said that they could cope this 281 long. Overall, gender, age, and household size did not appear to be associated with coping time 282 (Suppl . Tables 17-19 ). Factors that appeared to be associated with lower coping times were living 283 with children under 18 years (p=0.004, Suppl. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20209361 doi: medRxiv preprint 96% and 95% of respondents indicated that they had been using personal protective equipment (PPE; 296 e.g. face masks and gloves), compared to only 33% in UK, 55% in Italy, and 67% in Slovenia 297 (p<0.001). We also asked respondents how much they agreed with quarantine/isolation/social 298 distancing measures and the statement that these are a necessary strategy to help control COVID-19 299 (Suppl . Table 23 ). There was a significant difference between countries (p<0.001): agreement with 300 public health measures was highest amongst respondents from Thailand (94%) and lowest amongst 301 those from Slovenia (around 75%). 302

Overall, fewer male than female respondents changed their social behaviour before the government 303 implemented official restrictions (65% and 70%, respectively, p=0.039; Fig. 3 , Suppl. We asked respondents to indicate their perceived level of understanding of COVID-19. Overall, 59% 327 of respondents indicated a 'high/very high' level of understanding, 36% reported 'some' 328 understanding, and only 5% reported 'very little/none' (Fig. 4 , Suppl. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20209361 doi: medRxiv preprint significant differences at country level (p<0.001): perceived levels of understanding were highest in 330 Slovenia (66% reported 'high/very high', and 30% 'some' understanding) and Thailand (63% 331 'high/very high' and 33% 'some'), and lowest in Italy, with 47% reporting 'high/very high', and 50% 332 reporting 'some' level of understanding. 333

To probe for factors associated with perceived level of understanding of COVID-19, we broke down 334 responses by gender, age, education and healthcare worker status (Fig. 4, Suppl. Table 29 ). Overall, 335 there was no significant difference between men, women and people who identified as other or 336 preferred not to say (p=0.058; Fig. 4 , Suppl. Information about COVID-19, unclear information and fake news 351

When respondents were asked how they receive/received information about COVID-19 (Suppl . Table  352 31), most reported traditional mass media (TV, radio, newspapers; 93%), followed by online methods 353 (websites, email; 83%) and social media and messenger apps (79%). When asked about their 354 preferences for receiving information, the top three responses were traditional mass media (78%), 355 government or institution's website (77%), and online (76%). There were no significant differences 356 based on gender (Suppl . Table 32 ). Fewer respondents over 65 years said that they had used online 357 channels or social media and messenger apps, and they expressed significantly lower preference for 358 these channels too. For example, only 66% of over 65 year olds wanted to receive information online, 359 compared to 78%/79% of the other age groups (p<0.001), and only 52% of over 65 year olds 360 expressed preference for social media and messenger apps, compared to 64%/64% (p=0.005; Suppl. 361 Table 33 ). Overall, most respondents with primary/secondary education and those with tertiary 362 education had received information through traditional mass media, and social media/messenger apps 363 . CC-BY 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20209361 doi: medRxiv preprint (Suppl . Table 34 ). Fewer respondents with primary/secondary education had used online channels in 364 the form of websites and emails (79% versus 92%, p<0.001), and more had received face-to-face 365 information compared to those with tertiary education (43% versus 35%, p<0.001; Suppl. Table 34) . 366

However, both education level groups indicated that their preferred methods of communication were 367 mass media channels, online methods and government/institutions' websites. 368

We asked respondents if they had seen unclear or conflicting information about COVID-19 in nine 369 categories relating to infection, symptoms and various public health measures. Overall, between 36-370 54% of respondents indicated that they had seen such information. Ways to avoid the infection (54%), 371 government support schemes (52%) and testing (51%) were identified as the most unclear areas 372 (Suppl . Table 35 ). Thailand reported the lowest levels of seeing unclear or conflicting information in 373 most categories (around 35-40%), while respondents in the UK reported the highest levels in most 374 categories (around 55-70%). Overall, those with tertiary education reported significantly higher levels 375 of seeing unclear information than those with primary/secondary education in almost all categories 376 (p<0.001) except government support schemes (Suppl . Table 36) . 377

When asked "Have you come across news about the following COVID-19 topics that seemed fake to 378 you?", overall 63% of respondents had encountered news on "Coronavirus as an engineered modified 379 virus", 60% reported seeing "general spread of fear", and 51% had come across seemingly fake news 380 about "numbers of infected/deceased people", "home-made recipes to make sanitizer products" and 381 "alternative drugs/cure" (Fig. 5 , Suppl. Table 35 ). Thailand reported the lowest percentages in all 382 'fake news' categories, with a range of 27-42% (Suppl . Table 35 ). Overall, respondents with tertiary 383 education reported significantly higher levels of seeing 'fake news' in all categories compared to 384 those with primary/secondary education (p<0.001; Fig. 5 , Suppl. Table 36 ). For example, only 56% of 385 people with primary/secondary education reported coming across fake news about "coronavirus as an 386 engineered modified virus" versus 79% of those with tertiary education (p<0.001). There did not 387 appear to be an association between self-reported levels of understanding of COVID-19 and seeing 388 unclear/conflicting information or 'fake news' (Suppl. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint Our results showed that among the countries surveyed, respondents from Thailand were most 407 affected. Thailand is a middle-income country with a large informal economy, and relies heavily on 408 the tourism industry (15% GDP) 40 . Thailand also had a high government stringency index during the 409 period of the study (Fig. 1) , which included closure of borders, businesses and nighttime curfews 41 . is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20209361 doi: medRxiv preprint compliance was not higher in older people even though they might be expected to comply more due 431 to being a risk group. Similarly, our data showed that overall and in Malaysia, UK and Slovenia, far 432 fewer respondents over 65 years reported changing their behaviour voluntarily before official 433 restrictions came into place. However, overall, over 80% of respondents in all three age groups 434 expressed agreement when asked if they would comply voluntarily or with government-mandated 435 restrictions (Suppl . Table 26) . 436

Improving COVID-19 communication 437

Our findings indicated that younger age and lower education levels appeared to be associated with 438 lower self-perceived/subjective levels of understanding of COVID-19. Also, higher self-reported 439 levels of understanding of COVID-19 seemed to be associated with higher levels of understanding of 440 public health measures. A recent modelling study suggests that self-imposed public health measures 441 combined with fast spreading of disease awareness in the population can help reduce transmission of 442 the virus 11 . Our findings suggest that specific groups of people, such as those with primary/secondary 443 education levels and those 18-34 year old, may benefit most from targeted COVID-19 communication 444

initiatives. 445

In terms of channels of communications, the three most popular channels across countries were 446 traditional mass media, government or institutional websites, and online media. Similar results 447 emerged from a recent survey carried out in the Netherlands, Germany and Italy 52 . However, 448 respondents in Thailand reported that they preferred to receive information face-to-face, especially 449 those with primary/secondary education. This suggests that in order for communication strategies to 450 be effective, they need to be sensitive to population preferences and tailored to local contextual 451 factors (e.g. levels of connectivity, literacy 61 ). 452

Our survey showed that a significant proportion of the population received conflicting information 453 and news that seemed fake to them, in particular about coronavirus being an engineered modified is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20209361 doi: medRxiv preprint large sample of respondents in each country, we were able to compare population segments (e.g. men 464 versus women or younger versus older people) in our overall cohort and at country level. We did not 465 aim to obtain nationally representative samples and acknowledge that although we used weighting 466 strategies in our analysis, our results may not be fully representative of the populations in the 467 respective countries. Overall, there was a high proportion of respondents who were healthcare 468 workers (19%), and some variation in this proportion between countries. This may have influenced 469 the country level analysis, in particular in the areas of perceived understanding, 470 compliance/agreement and communication preferences. 471

Because the survey was online, only people who were literate, had internet access, and had access to 472 computers or smartphones could take part. Due to COVID-19 related restrictions, it was not possible 473 to conduct face-to-face data collection to reach groups who were illiterate in the language of the 474 survey, or who did not have access to online technology. This is likely to have biased our data 475 towards more educated and economically advantaged populations. Our study was also subject to 476 response bias and other biases arising from self-reporting and recall. Lastly, as a cross-sectional 477 survey, our data only sheds light on the prevalence of certain phenomena and opinions of respondents 478 but does not imply causality. 479

The results of the survey reported here form part of a mixed-methods study, which also includes an 480 in-depth qualitative study, the findings of which are currently being analysed and will be published 481 separately. Combined, our results may help explain some of the trends reported in this survey, as well 482 as the differences between countries and social groups. We have also conducted a preliminary is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20209361 doi: medRxiv preprint

The authors declare no conflict of interest. The funders had no role in the design of the study; in the 526 collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to 527 publish the results. 528 is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20209361 doi: medRxiv preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20209361 doi: medRxiv preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20209361 doi: medRxiv preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20209361 doi: medRxiv preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20209361 doi: medRxiv preprint

New insights of emerging SARS-CoV-2: epidemiology, 545 etiology, clinical features, clinical treatment, and prevention

Review of the 2019 novel coronavirus (SARS-CoV-2) based 548 on current evidence

The first 75 days of novel coronavirus (SARS-CoV-2) 551 outbreak: recent advances, prevention, and treatment

Centers for Disease Control and Prevention. Non-pharmaceutical interventions (NPIs). 2020. 558

Non-pharmaceutical public health interventions 561 for pandemic influenza: an evaluation of the evidence base

Design of non-pharmaceutical intervention strategies for pandemic 564 influenza outbreaks

Impact of non-pharmaceutical 567 interventions (NPIs) to reduce COVID19 mortality and healthcare demand

Interventions to mitigate early spread of SARS-CoV-2 571 in Singapore: a modelling study

Estimating the effects of non-pharmaceutical 574 interventions on COVID-19 in Europe

Impact of self-imposed prevention measures and 577 short-term government-imposed social distancing on mitigating and delaying a COVID-19 578 epidemic: a modelling study

Case-control study of use 581 of personal protective measures and risk for severe acute respiratory syndrome coronavirus 2 582 Infection

Scientific and ethical basis for social-distancing interventions against 585 COVID-19

What is Inconvenient for You is Life-saving for Me': How Health Inequities are 588 playing out during the COVID-19 Pandemic

Using social and behavioural science to support 591 COVID-19 pandemic response

Improving the impact of non-pharmaceutical 594 interventions during COVID-19: examining the factors that influence engagement and the 595 impact on individuals

World Health Organisation. A coordinated global research roadmap: 2019 novel coronavirus

Social, ethical and behavioural aspects of 601 COVID-19

The remaining unknowns: a 604 mixed methods study of the current and global health research priorities for COVID-19

Online survey questions: Social, ethical 607 and behavioural aspects of COVID-19 (Version Version 2

Online surveys (formerly BOS)

National Research Council. Cognitive aspects of survey methodology: building a bridge 614 between disciplines

The Medical Chamber of Slovenia

Super Poll Thailand

About boosted posts

Mixed Methods Research and Data Analyses

Demographic and human capital scenarios for the 21st 628 century: 2018 assessment for 201 countries

Statacorp. Stata Statistical Software: Release 15. College Station, TX: StataCorp LLC

Oxford COVID-19 Government Response Tracker

Blavatnik School of Government

Coronavirus' devastating economic impact on workers age 50-plus

the Community. COVID-19: economic impact on age in the workplace. 2020. 638

The risk older workers face in the wake of COVID-19

Valuing older people: time for a global campaign 644 to combat ageism

First findings

The indirect impact of COVID-19 on women

COVID-19 and gender equality: Countering the 652 regressive effects

Thailand Economic Monitor: Thailand in the Time of COVID-19 659 (English). 2020

Thailand's experience in the COVID-19 response

The psychological impact of quarantine and 666 how to reduce it: rapid review of the evidence

Mental health before and during the COVID-19 pandemic: 669 a longitudinal probability sample survey of the UK population

Public knowledge, attitudes and practices 675 towards COVID-19: A cross-sectional study in Malaysia

Knowledge, attitudes, impact, and anxiety regarding COVID-19 678 infection among the public in China

Study of knowledge, attitude, anxiety & perceived 681 mental healthcare need in Indian population during COVID-19 pandemic

Use of rapid online surveys to assess people's perceptions during infectious 684 disease outbreaks: A cross-sectional survey on COVID-19

Knowledge, attitudes, and practices towards COVID-19 687 among Chinese residents during the rapid rise period of the COVID-19 outbreak: a quick 688 online cross-sectional survey

Economic and social consequences of human 690 mobility restrictions under COVID-19

Why people comply with COVID-19 social 693 distancing restrictions: Self-interest or duty?

Public perspectives on protective measures during the 696 COVID-19 pandemic in the Netherlands, Germany and Italy: A survey study

Factors associated with people's behavior in 699 social isolation during the COVID-19 pandemic

Oxford Supertracker: The Global Directory for 702 COVID Policy Trackers and Surveys, Department of Social Policy and Intervention

Elderly people and responses to COVID-19 in 27 Countries

Misinformation of COVID-707 19 on the internet: infodemiology study

Family Life in Lockdown

Immediate impact of stay-at-home orders to 713 control COVID-19 transmission on socioeconomic conditions, food insecurity, mental health, 714 and intimate partner violence in Bangladeshi women and their families: an interrupted time 715 series

Community responses during early phase of 719 COVID-19 epidemic

Effective health risk communication about pandemic influenza for 722 vulnerable populations

Perspectives on public health 730 interventions in the management of the COVID-19 pandemic in Thailand