Some lessons learned from the COVID-19 pandemic: Subjective well-being before and during the pandemic among Brazilian adults

Graphical abstract Image, graphical abstract


Introduction
The world is facing an unprecedented health crisis. The World Health Organization (WHO) declared on January 30, 2020 a public health emergency of international concern due to COVID-19, a disease caused by the severe acute respiratory syndrome coronavirus (SARS-CoV-2) ( Phelan et al., 2020 ). The disease spread across countries and continents. It reached Brazil in February 2020, where it produced a rapidly growing epidemic with high casualty rates, and the country became one of the worst affected countries of the world ( Gelfand et al., 2021 ). The first death record occurred on March 12, 2020. At the time of the first data collection during the COVID-19 pandemic in the present study, in June 2020, Brazil had reached 29,937 deaths, and at the time of the second data collection during the pandemic, in April 2021, the country had reached 325,284 deaths.
The pandemic has dramatically impacted livelihoods, and people have had to reinvent themselves to adapt to the new reality. The effects of the pandemic vary greatly among people, both in the unevenness of the economic and health impacts, as well as how they experience and adapt to the challenges. Recent studies have investigated how this experience has affected overall psychological health and well-being. The results are quite diverse, with studies showing an increase in suicidal thoughts, and others even showing a positive impact on daily life ( O'Connor et al., 2021 ;Evans et al., 2020 ). A recent review of 19 studies reported an overall increase in symptoms of depression and anxiety taking into account prevalence before and after the pandemic ( Xiong et al., 2020 ). Some risk factors were identified for developing mental disorders: being female, being under 40, having comorbidities, poor economic status, lower education level, unemployment and frequent exposure to news/information concerning COVID-19.
The World Happiness Report, using the Cantril Self-Anchoring Scale (a ladder with steps numbered from zero at the bottom, representing the worst possible life, to 10 at the top, representing the best possible life), compared 95 countries around the globe ( Helliwell et al., 2021 ). The report shows that COVID-19 has led to relatively modest changes in the overall rankings, reflecting a generally shared resilience in the face of the pandemic. But it is notable that Brazil fell from rank 29 in 2017-2019 (Mean score = 6.38 95% CI 6.30-6.46) to rank 41 in 2020 (Mean score = 6.11 95% CI 5.89-6.33). The country with the highest rank was Finland both in 2017-2019 (Mean score = 7.81 95% CI 7.75-7.87) and in 2020 (Mean score = 7.89 95% CI 7.78-8.00), and the country with the lowest rank (95) was Zimbabwe both in 2017-2019 (Mean score = 3.30 95% CI 3.18-3.41) and in 2020 (Mean score = 3.16 95% CI 2. 95-3.36).
Despite this result on happiness, a study conducted in São Paulo (Brazil) did not find higher levels of depression and anxiety in the population studied in 2020 compared to pre-pandemic levels ( Brunoni et al., 2021 ). However, they identified risk factors representing socioeconomic disadvantages. Self-evaluated good health and good quality of relationships were associated with decreased odds for psychological symptoms, whereas younger age, female gender, and lower educational level were associated with increased risk.
Indeed, several studies demonstrate that women and younger people experienced more mental health problems than men and older people ( Banks and Xu, 2020 ;Oreffice and Quintana-Domeque, 2021 ), and these variables accounted for the largest share in explaining changes in the General Health Questionnaire 12-item scale (GHQ-12) scores in the COVID-19 period ( Davillas and Jones, 2020 ).
In a study conducted in Spain, at the outbreak of COVID-19 (March and April 2020), Sandín et al. (2020)  making persons more aware of life aspects they had not noticed before, and discovering new interests and abilities. Among protective factors they highlighted higher age, higher income, having a garden at home.
The pandemic raises the question of the coping strategies used by people to relieve stress. Positive coping strategies function as protective factors against the negative impacts of the pandemic regarding mental and physical health, whereas maladjusted coping strategies function as risk factors. In a qualitative study conducted in the UK between April and May 2020 with a small sample of 15 males and 35 females, Ogueji et al. (2021) identified 11 themes in the following order: "socializing through video calls ", "engaging in exercise ", "being occupied with jobs ", "being occupied with studies ", "avoiding negative news on COVID-19 ″ , "consumption of alcohol ", "healthy eating ", "engaging in meditation activities ", "gaming activities ", "hope ", and "self-care and self-appreciation ". Of the mentioned strategies, alcohol consumption is maladjusted because among other negative outcomes it increases the risk of acquiring viral infections ( Simou et al., 2018 ;Lassen et al., 2021 ).
Given that Brazil was one of the countries most impacted by the pandemic in number of deaths from COVID-19 and that the pandemic has further underscored the economic crisis that was already present in the country before the pandemic ( Neves et al., 2021 ), we were interested in understanding how the pandemic affected the way people evaluate happiness, subjective well-being and life satisfaction in three times: before (T1 = 2018) and during the pandemic (T2 = 2020 and T3 = 2021) in a Brazilian sample.
Our aims were (i) to investigate the exposure of the respondents and their network to COVID-19 during pandemic in two time points (2020 and 2021). Our hypothesis was that more people would report having acquaintances, friends or relatives infected with COVID-19; (ii) to examine the impact of COVID-19 pandemic on feelings of happiness. Our hypothesis was that lower happiness would be found during the pandemic (2020 and 2021) in comparison to the levels found before (2018); (iii) to assess well-being (MSWS) as a function of time (2018, 2020 and 2021), family income, housing status, sex, and age. Our hypothesis is that well-being would be lower in the years 2020 and 2021 (during the pandemic), especially among those with lower family incomes, younger people, and females; (iv) to assess the sub-dimensions that compose the MSWS scale: negative states; positive states, purpose in life, satisfaction with the environment, appearance linked to self-esteem, and spirituality as a function of time. Our hypothesis is that the pandemic would increase negative states and spirituality, and decrease positive states and purpose in life; (v) to assess behavioral coping strategies that respondents report to manage the psychological stress associated with the COVID-19 pandemic; (vi) to identify coping behaviors that were predictive of higher well-being and to verify whether knowing more people infected with COVID-19 was a risk factor predictive of lower well-being. Our hypotheses were lower well-being among those who knew more people infected with COVID-19 or among those who reported more severe cases of COVID-19; being in contact with family through video call, doing physical exercises, working and listening to music would be coping behaviors linked to higher well-being, while drinking alcohol and watching television would be behaviors associated with lower wellbeing. Analyses regarding other coping behaviors were exploratory; and finally (vii) to investigate whether coping behaviors and severe contact with COVID-19 had influence on the levels of compliance to protective measures. Our hypothesis was more use of protective measures among those more impacted by COVID-19 (COVID-19 severity) and to explore which of the behaviors used to cope with the pandemic were associated with greater use of protective measures.

Methods
This was a cross-sectional historical design which includes data from three time points: 2018 (before pandemic) and during pandemic in 2020, and 2021, when Brazil had reached 29,937 and 325,284 deaths, respectively.

Participants
A total of 1840 Brazilian adults (51.5% women and 48.5% men) in the age range 18 to 50 years participated in the study: 616 ( M = 34.46 years, SD = 8.55) in 2018, 379 ( M = 33.86 years, SD = 8.19) in 2020 and 845 ( M = 35.73 years, SD = 8.30) in 2021. They came from the five macro-regions of Brazil. See Appendix 1 for details. The inclusion criteria were: Brazilian, over 18 years old and access/familiarity with computer/tablet for internet use. The respondents were recruited among panelists from the Netquest Research Institute using quota sampling defined by sex, age, social class, and the Brazilian regions. Those who agreed to participate completed an online questionnaire. The study was approved (CAAE 80,833,817.6.0000.5561) by the National Council of Ethics in Research (CONEP) of Brazil and all participants consented to participate in the study.

Assessment instruments
Personal retrospective: Self-comparison to one year ago . Respondents compared how they felt about their lives now with a year ago answering the question "Do you feel happier or less happy nowadays than you did a year ago? " (1 = I feel less happy than last year; 2 = I feel the same, neither more nor less happy, than last year; 3 = I feel happier than last year, 4 = I do not know). See Supplementary Material, Appendix 2. These questions were answered in 2018, 2019, and 2020.
Multidimensional Subjective Well-Being Scale (MSWS). It includes 29 items related to negative states (e.g., stress, nervousness), positive states (e.g., good mood, happiness), purpose in life (e.g., goals to achieve, selfconfidence), satisfaction with the environment (e.g., residence, district), appearance linked self-esteem (e.g., feeling attractive, looking in the mirror and feeling good), spirituality (e.g., close relationship with God, habit of praying). Each item had an 11-point response scale (0-10), with a verbal anchor at each pole. We previously conducted an exploratory factor analysis of the scale, retaining factor loadings higher than 0.40 and the reliability coefficients were as follows: Socio-demographic information . A self-report sociodemographic questionnaire was used to gather information about gender, age, education, marital status, employment status, household monthly income, and region of residence. These questions were answered in 2018, 2019, and 2020.
Exposure of respondents to COVID-19 : We asked participants if someone on their social network (family members, friends, acquaintances) had experienced symptoms, been diagnosed, been hospitalized or died from COVID-19. They also provided information about themselves. Each of the items received one point if answered positively. These questions were answered in 2020 and 2021, during the pandemic period. See Supplementary Material, Appendix 4.
Coping strategies: Respondents were asked "which activities during the pandemic brought you a sense of well-being? The question could be answered by checking one or more of the fourteen options: physical activities, reading, writing, watching TV, playing, drinking alcohol, taking illicit drugs, taking medicine, browsing social networks, playing or listening to music, dancing, talking to people they live with, talking to friends or family members on the phone or video calls, working, doing housework. Questions about coping behaviors were asked only in 2021, during the peak of the pandemic in Brazil. See Supplementary Material, Appendix 5.
Compliance behavior: we asked the respondents how well they were following the social isolation restrictions on a scale of 0 to 10 (0 = no adherence to isolation measures to 10 = full adherence to social isolation). The question about compliance behavior was asked only in 2021, during the peak of the pandemic in Brazil. See Supplementary Material, Appendix 4.
COVID-19 severity: this variable is the sum of the responses to the five questions that we considered to be most severe: whether the respondent had been hospitalized (1), had any close friends who were hospitalized (2) or died as a result of COVID-19 (3), or had any family members been hospitalized (4) or who died as a result of COVID-19 (5). The score was from 1 to 5 where 1 means that the person experienced only one of these situations and 5 means that all of these situations were experienced by the respondent. This variable was calculated with samples that responded during the pandemic, i.e. in 2020 e 2021.
Health and medicine use: Regarding health, we asked three questions: (a) Are you using or did you need to use medicines to control anxiety or depression during the pandemic period (from March 2020 to today)?; (b) Have you used similar medications before the pandemic? and (c) Comparing the period before and after the pandemic, how much do you suffer from headaches, high blood pressure, or insomnia? These data were collected only in 2021. See Supplementary Material, Appendix 5.

Statistical analysis of the data
Frequency analyses were computed using contingency chi-square. When the chi square contingency test values were significant and the levels of the variables examined were greater than two, we examined the values of the adjusted residuals by cell. The standardized residual is found by dividing the difference of the observed and expected values by the square root of the expected value. If the standardized residual is > + 2 then that cell can be considered to be a major contributor to the overall chi-square value. The adjusted standardized residuals are standardized residuals that are adjusted for the row and column totals.
In order to investigate how the coping strategies and the severe contact with COVID-19 predicted the variance of the subjective well-being and social isolation measures, two Regression Analysis were conducted.

Exposure of respondents and their social network to COVID-19
Analyses of the exposure of respondents and their social network to COVID-19 (see Supplementary Material, Appendix 6) revealed significant increases over time in the proportion of respondents who gave affirmative answers to the questions asked. All questions except one revealed significant differences. Regarding death of COVID-19 in the respondent's social network, odds ratios (OR) were 2.96 (acquaintances) to 4.49 (family members) times higher in 2021 than in 2020. Regarding hospitalization, ORs were 2.06 (acquaintances) to 11.05 (own person) times higher in 2021 than in 2020. With respect to symptoms (1.39-3.31) and diagnosis of COVID-19 (1.91-5.32), ORs were also higher in 2021 than in 2020. Respondents reported high social isolation compliance with no difference between the 2020 (Mean = 7.77 95% CI 7.52-8.02) and 2021 (Mean = 7.81 95% CI 7.65-7.97) samples. These results were in line with our expectation that respondents would have known more people infected with COVID-19 in 2021 compared to early 2020, when the pandemic was starting in Brazil. We compared the frequency of behavioral coping strategies among people who had at least one family member or close friend who was hospitalized or died from COVID-19 or self-reported experience of hospitalization (cases we call severe) with those who had no severe cases among friends or in the family nor self-reported experience of hospitalization. The behavioral coping strategies that were significantly more frequent among severely affected respondents were social interaction at home ( X 2 = 12,129; p < 0.001), socializing with family and friends ( X 2 = 3120; p = 0.045), working ( X 2 = 2987; p = 0.049), and doing housework ( X 2 = 4.160; p = 0.024).

Personal retrospective on happiness: comparison with the previous year
Examining respondents' affective retrospective, we found a statistically significant association with time, X 2 (4, N = 1765) = 23.650, p = 0.001). In line with our initial hypothesis (i), their perspective was more positive before the COVID-19 pandemic than during the pandemic ( Table 1 ). High positive adjusted residuals ( > 2) indicated that before the COVID-19 outbreak (2018 sample) there were more respondents who perceived themselves happier in comparison with the last year than expected, adjusted for sample size. At the peak of the COVID-19 pandemic (2021 sample) there were more respondents who perceived themselves less happy in comparison with the last year than expected, adjusted for sample size.
In addition, two by two comparisons of the years considering a dichotomous personal retrospective (feeling happier than in the previous year vs. the same or less happy) revealed no difference between 2018 and 2020, X 2 (1, N = 953) = 0.761, p = 0.383. However, respondents were significantly happier in 2018 than in 2021, X 2 (1, N = 1395,) = 17.850, p = 0.001, and happier in 2020 than in 2021, X 2 (1, N = 1182) = 7.528, p = 0.006. Our initial hypothesis (ii) was that lower happiness would be found during the pandemic (2020 and 2021) in comparison to the levels found before (2018). However, the difference was between the 2018 and 2021 sample, i.e. before the pandemic and during the peak of the pandemic. At the start of the pandemic in 2020, although fewer people responded that they felt happier than the previous year and more people responded that they were less happy compared to the pre-pandemic period, this difference was not significant.

Multidimensional subjective well-being scale (MSWS)
We  p = 0.001, and Gender, Wilks' lambda = 0.975, F (6,1408) = 6.035, p = 0.001, and a significant effect of the covariate Age, Wilks' lambda = 0.945, F (6,1408) = 13.648, p = 0.001. Well being was lower for women, younger respondents and for those with lower family income in comparison with men, older participants and those with higher family income. However, no significant interaction effects were found. Our initial hypothesis (iii) was that compared to pre-pandemic (2018), during the pandemic (2020 and 2021) people with lower family income, younger people, and women would be more negatively affected than people with higher family income, older people and men. However, contrary to our initial hypothesis, we did not find interaction effects between each of these variables and time. Subsequent univariate analyses revealed significant main effects of time on Negative States, F (2,1413) = 2.949, p = 0.05, and Spirituality, F (2,1413) = 3.526, p = 0.03 Table 2 . Pairwise post-hoc comparisons ( p < 0.05) showed that negative states scores were higher (2021 > 2018 = 2020) and spirituality scores were lower (2018 > 2020 = 2021) during the COVID-19 pandemic than before.
Our initial hypothesis (iv) was that the pandemic would increase negative states, and decrease feelings of positive states and purpose in life. However, the pandemic seems to have had an effect by increasing negative states, as we had predicted, but did not decrease positive feelings and purpose in life. Our results also indicated that the pandemic decreased feelings of spirituality, contrary to what we had predicted.

Behavioral coping strategies and well-being
In this section our objective (v) was to assess behavioral coping strategies that respondents report to manage the psychological stress associated with the COVID-19 pandemic and (vi) to identify coping behaviors that were predictive of higher well-being.
Twenty-eight percent of respondents of the 2021 sample reported increased physical symptoms (headache, high blood pressure or insomnia) during the COVID-19 pandemic. Twenty-one percent reported using medications to control these symptoms. However, this proportion was not higher than before the pandemic. Alcohol consumption as part of the response to COVID-19 was reported by 22% of the respondents and use of illegal drugs by 4%. The most reported stress relievers were: watching TV (59%), surfing the Internet (50%), engaging in exercise (45%).
Linear models were created in order to investigate whether contact with COVID-19 and coping behaviors (predictor variables) influence the MSWS or the social isolation compliance scores (outcome variables). First, full models were built with all predictor variables and only significant predictor variables ( p < 0.05) were kept. For both outcome variables, the selected models presented better likelihood values than the null models and showed no difference from the full model. By the law of parsimony, the selected model was adopted with fewer parameters. The residual errors of the selected models showed a distribution assumed to be Gaussian by quantile-quantile plots and by the Shapiro-Wilk test after the Yeo-Johnson transformation.
The results of the regression analysis showed that the greater the number of people the participants knew who had been infected by the virus, the lower their well-being. Some behaviors used to cope with the stress of the pandemic predicted higher well-being: physical exercise, socializing at home, socializing through video calls, working and housework ( Table 3 ). Use of alcohol and illicit drugs, as well as watch TV on the other hand, were behaviors that predicted lower well-being.
Our results support our initial hypothesis (vii) that the number of people known to have been infected by COVID-19 would be a predictor of lower well-being. Again in line with what we had predicted (vii), our results show that physical exercise, interaction with family, socializing via social networks and working were associated with greater well-being. Alcohol and illicit drug use and watching television were behaviors associated with lower well-being, as predicted. Contrary to what was predicted, listening or playing music was not associated with greater well-being. Housework was associated with higher well-being, something we had not predicted.

Behavioral strategies and social isolation
Our last analysis aimed to (vii) assess whether the number of known people infected with COVID-19 would influence adherence to social isolation recommendations in the year 2021, at the peak of the pandemic. We also aimed to investigate which coping strategies that participants mentioned predicted greater adherence to social isolation measures.
Regression analysis revealed that some behaviors used to cope with stress of the pandemic predicted greater adherence to social isolation measures: reading or writing, watching television, and socializing with family and friends by videoconference ( Table 4 ). Working, on the other hand, was a behavior that predicted lower adherence to social isolation measures. Contrary to our initial hypothesis, the number of people the participants knew who had been infected with the virus was not associated with greater adherence to social isolation. However, the experience of having been hospitalized or having close family members and friends who were hospitalized and/or who died from COVID-19, i.e., only the most severe cases, predicted greater adherence to social isolation (Level 4 of severity of exposure to COVID-19 on a scale from 1 to 5).

Discussion
In the present study, we found evidence of psychological impact experienced by Brazilian adults related to the COVID-19 pandemic. At the time of data collection in June 2020 the number of cases of COVID-19 in Brazil was 526,447 and the number of deaths was 29,937. By April 2021 these numbers had increased to 12,839,844 and 325,284 respectively. Brazil began administration of COVID-19 vaccines on January 17, 2021, when the country had 210 thousand deaths ( da Fonseca et al., 2021 ). On April 13, 2021, the Brazilian Senate opened the COVID Parliamentary Commission of Inquiry (CPI) to investigate authorities' responsibilities in dealing with the pandemic, including lack of central coordination associated with the country's reduced success in limiting cases and deaths. In the current study, as reported by the respondents, odds ratios (OR) of death in their social network were 2.96 (acquaintances) to 4.49 (family members) times higher in 2021 than in 2020. Regarding hospitalization, ORs were 2.06 (acquaintances) to 11.05 (own person) times higher in 2021 than in 2020. It is in this context that our data should be understood. The results regarding the psychological impact of COVID-19 on Brazilian respondents are in line with the World Happiness Report, which showed that among 95 countries Brazil fell from rank 29 in 2017-2019 to rank 41 in 2020 ( Helliwell et al., 2021 ).
The analysis of personal retrospective data in the current study showed that happiness decreased significantly during the pandemic. In the 2018 sample there were more respondents who perceived themselves happier in comparison with the last year than expected, whereas in the 2021 sample there were more respondents who perceived themselves less happy in comparison with the last year than expected. There was no difference between the 2018 and 2020 samples. In the current study, our findings regarding the personal retrospective data are in line Table 3 Findings from a linear model using coping behaviors and severe contact with COVID-19 as the predictor variables to explain sense of well-being. 835 829 p-value < 0.001 < 0.001 * The referential class for coping behaviors was the answer 'No' and for the severe contact with COVID-19 it was 'level 0 ′ ; AIC is Akaike criterion; BIC is the Bayesian information criterion; SE is standard error; p < 0.05 are highlighted in bold. (0.2347) * The referential class for coping behaviors was the answer 'No' and for the severe contact with COVID-19 it was 'level 0 ′ ; AIC is Akaike criterion; BIC is the Bayesian information criterion; SE is standard error; p < 0.05 are highlighted in bold.
with the analyses of the scores on the MSWS dimensions. The scores of negative states were higher in the 2021 sample than in the 2018 sample. No difference was found between 2020 and 2018. These results are in agreement with other studies that found increased negative affect in association with COVID-19. Pérez-Fuentes et al. (2020) points to a circular relationship, in which threat perception triggers negative mood, and negative mood, in turn, lowers the threshold for experiencing negative emotions from present situations, stimulating the feeling of threat.
During the third data collection in 2021 the number of deaths due to the COVID-19 virus had increased by more than 10 times in comparison to the period when the data were collected in 2020. In 2021, the population had also been exposed for a longer time to social restrictions and to protective measures against the coronavirus. People might have not fully realized how the pandemic had changed their lives in 2020. We believe it was for this reason that there was no difference comparing data collected in 2018 and 2020 regarding negative states and perceived happiness compared to the previous year. Our hypothe-sis is that the effects of the pandemic became increasingly stressful from a psychological, social, and probably economic point of view.
When facing life-threatening experiences spirituality may be a protective factor ( Norris and Inglehart, 2011 ). Feelings of vulnerability are the underlying motivational mechanism. The literature recognizes two main dimensions of vulnerability: exposure to financial and lifethreatening risks ( Immerzeel and van Tubergen, 2011 ;Molteni et al., 2021 ). During the COVID-19 pandemic, when negative states scores were higher, the spirituality scores were lower than before the pandemic and not higher as we expected. This finding may indicate that people are struggling with their faith in this time of crisis of increase in negative feelings ( Goodman, 2020 ). Respondents may be experiencing feelings of anger, abandonment or punishment by a higher power. In line with previous studies (e.g., Brunoni et al. 2021 ) in the current study, irrespective of time, well-being was lower for women, for younger respondents and for those with lower family income in comparison, respectively, with men, older respondents and those with higher family income. When we planned our study, we expected to find some buffering effects. We anticipated that being men, older, with higher family income would reduce the impact of pandemic on well-being. Specifically, we expected that MANCOVA would reveal interaction effects, i.e., effect of time (2018 × 2020 × 2021) dependent on the value of the other variables (gender, age, and family income). However, the statistical analysis did not reveal two-way interactions nor higher order interactions. Only main effects were found. Consequently, we are led to conclude that the COVID-19 pandemic harmed people's well-being in a similar way, irrespective of their gender, age, or family income.
Regarding activities that contribute to well-being during the pandemic, in our study respondents of the 2021 sample reported engagement in social interaction at home (42%), socializing through video calls with family and friends (32%), and playing with kids and pets (40%). These are recommended coping strategies for reducing psychological distress ( Kim et al., 2008 ;Gee et al., 2021 ). They also reported engaging in exercise (45%) and dancing (13%) which are also recommended behavioral strategies considering the available evidence that exercise is a protective factor against viral infections ( Ranasinghe et al., 2020 ).
In the current study respondents also mentioned surfing the Internet (50%) and watching TV (59%) as coping behaviors. Bland et al. (2014) reported that website navigation does not relieve stress, but rather might exacerbate it. Welcome sources of entertainment may result in excessive involvement (e.g., "binge-watching ", see Boursier et al. 2021 ). TV plays a critical informational role during the COVID-19 pandemic, but viewers may be overwhelmed by the news ( Solomon et al., 2021 ).
Our results showed that knowing more people infected with COVID-19 was predictive of lower well-being as well as the use of alcoholic and illicit drugs and watching television. On the other hand, physical exercise, social interaction at home, socializing with friends and family via video calls, working, and doing housework were predictive of higher well-being. Regarding preventive measures, watching television, socializing via video calls with family and friends, reading or writing and having relatives and close friends who were hospitalized or died as a result of COVID-19 predicted greater adherence to social isolation measures during the pandemic. We interpret the findings of our study as an indication that the loss of friends and family over COVID-19 and even the fear of losing loved ones had a profound impact on the well-being.
Participants in the 2021 sample reported using medication to control anxiety/depression (21%), alcohol (22%), or illicit drugs (4%) to feel better. Twenty-two percent of those who used medication also reported alcohol consumption and 8% reported illicit drug consumption. Ongoing substance use may increase the risk for infection with COVID-19 ( Melamed et al., 2020 ;Kumar et al., 2021 ).
Proper psychosocial support programs are needed to reinforce adaptive coping strategies and to reduce maladaptive ones, helping change people into active survivors of the coronavirus crisis rather than passive victims.

Conclusion
This exploratory study examined the perceived well-being levels and behavioral strategies related to COVID-19 among Brazilian adults living in one of the hardest hit countries in the world. Even people who were not directly affected by the virus, had acquaintances, relatives, and friends who have been affected. During the COVID-19 pandemic, scores of negative states were higher and spirituality scores were lower than before the pandemic. Greater number of people known to have COVID-19 was associated with lower well-being, as well as the use of alcohol and illegal drugs and watching TV. On the other hand, physical exercise, socializing at home, and working were associated with higher well-being. People who had family members or close friends who was hospitalized or died as a result of COVID-19 reported higher levels of compliance to isolation measures. Furthermore, some behavioral strategies were associated with high levels of compliance with the COVID-19 social isolation prevention measures, such as reading, writing, watching television, and socializing with family and friends through video calls.

Limitations
The current study examined data collected from a Brazilian sample and is subject to influences from the local context. Some sociodemographic subpopulations, such as individuals who were elderly or with primary education, were underrepresented. We do not have data on the frequency of behaviors that respondents used to cope with stressful situations before and after the pandemic, and therefore the coping strategies refer only to coping with the stress of the pandemic during peak infection of COVID-19 in Brazil in 2021.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.