Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID- 19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. Environment International 154 (2021) 106664 Contents lists available at ScienceDirect Environment International journal homepage: www.elsevier.com/locate/envint Exposure to nature and mental health outcomes during COVID-19 lockdown. A comparison between Portugal and Spain Ana Isabel Ribeiro a, b, c, *, 1, Margarita Triguero-Mas d, e, f, g, 1, Cláudia Jardim Santos a, b, c, Alicia Gómez-Nieto h, Helen Cole d, e, f, Isabelle Anguelovski d, e, f, i, Filipa Martins Silva j, Francesc Baró d, f, k, l a EPIUnit – Instituto de Saúde Pública, Universidade do Porto, Rua das Taipas, n◦ 135, 4050-600 Porto, Portugal b Departamento de Ciências da Saúde Pública e Forenses e Educação Médica, Faculdade de Medicina, Universidade do Porto, Porto, Portugal c Laboratório para a Investigação Integrativa e Translacional em Saúde Populacional (ITR), Rua das Taipas 135, 4050-600 Porto, Portugal d Institute of Environmental Science and Technology (ICTA), Universitat Autònoma de Barcelona (UAB), Edifici Z (ICTA-ICP), Carrer de les Columnes s/n, Campus de la UAB, 08193 Cerdanyola del Vallès, Spain e Hospital del Mar Medical Research Institute (IMIM), Carrer Doctor Aiguader 88, 08003 Barcelona, Spain f Barcelona Lab for Urban Environmental Justice and Sustainability (BCNUEJ), Carrer Doctor Aiguader 88, 08003 Barcelona, Spain g Mariana Arcaya’s research lab, Department of Urban Studies and Planning, Massachusetts Institute of Technology, Cambridge, MA, USA h Departamento de Arquitectura, Universidad de Alcalá de Henares, Spain i Institució Catalana de Recerca i Estudis Avançats (ICREA), Passeig Lluís Companys 23, Barcelona, Spain j Departamento Pedopsiquiatria e Saúde Mental da Infância e da Adolescência do Centro Hospitalar e Universitário do Porto, Largo Prof. Abel Salazar, 4099-001 Porto, Portugal k Vrije Universiteit Brussel (VUB), Geography Department, Pleinlaan 2, B-1050 Brussels, Belgium l Vrije Universiteit Brussel (VUB), Sociology Department, Pleinlaan 2, B-1050 Brussels, Belgium A R T I C L E I N F O A B S T R A C T Handling Editor: Zorana Jovanovic Andersen Background: To control the spread of the SARS-CoV-2 virus, countries around the world implemented lockdowns with varying intensities. Lockdowns, however, have been associated with a deterioration of mental health, Keywords: including post-traumatic stress symptoms, anger and anxiety. Exposure to nature might reduce stress and provide Green space relaxation opportunities. Nature Objective: Firstly, we aimed to determine which sociodemographic, housing and lockdown-related characteristics Confinement were associated with changes in exposure to nature during the COVID-19 lockdown in Portugal and Spain. Pandemic Mental health Secondly, we sought to estimate the associations of these changes with mental health, and test whether these SARS-CoV-2 associations differed according to sociodemographic characteristics and between the two countries, which experienced different restrictions and epidemiological situations. Methods: A cross-sectional study was conducted between March 27 and May 6, 2020, using an online ques­ tionnaire to measure changes in exposure to nature (including private green space and other greenery, views of nature from home and public natural spaces); sociodemographic, housing and lockdown-related characteristics; stress levels (visual stress scale); psychological distress (General Health Questionnaire − 12 items) and soma­ tization (somatization scale). Adjusted regression models were fitted to estimate associations. Results: This study included 3157 participants (1638 from Portugal, 1519 from Spain). In Portugal, maintaining/ increasing the use of public natural spaces during the lockdown was associated with lower levels of stress (adjusted beta − 0.29; 95%CI − 0.49, − 0.08) and maintaining/increasing the frequency of viewing nature from home was associated with reduced psychological distress (0.27; − 0.51, − 0.03), somatization (− 0.79; − 1.39, − 0.20), and stress levels (− 0.48; − 0.74, − 0.23). In Spain, maintaining/increasing contact with private green space and greenery was associated with lower stress levels: for contact with indoor plants (− 0.52; − 0.96, − 0.07) and for use of private community green spaces (− 0.82; − 1.61, − 0.03). Conclusion: Exposure to nature was associated with better mental health outcomes during lockdowns, but the natural features associated with improved mental health differed between the two countries. Nature should be * Corresponding author at: EPIUnit – Instituto de Saúde Pública, Universidade do Porto, Rua das Taipas, n◦ 135, 4050-600 Porto, Portugal. E-mail address: ana.isabel.ribeiro@ispup.up.pt (A.I. Ribeiro). 1 Equal contributors. https://doi.org/10.1016/j.envint.2021.106664 Received 17 July 2020; Received in revised form 22 March 2021; Accepted 22 May 2021 Available online 28 May 2021 0160-4120/© 2021 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). A.I. Ribeiro et al. Environment International 154 (2021) 106664 incorporated into urban planning interventions and housing design and exposure to nature should be promoted during lockdowns. 1. Introduction prohibited until May 2nd. While lockdowns have helped to contain the spread of the SARS-CoV- Research has shown that nature can provide many benefits for public 2 virus, they have been associated with worsening mental health and health. Exposure to nature, particularly to public green and blue spaces, well-being, including post-traumatic stress symptoms, confusion, anger is associated with numerous health benefits including better cardio­ and anxiety (Armitage and Nellums, 2020; Brooks, 2020; Pfefferbaum vascular and respiratory health (Lane Kevin, 2017; Cavaleiro Rufo, and North, 2020; Rajkumar, 2020; Torales et al., 2020). These problems 2020), improved immune (Cavaleiro Rufo, 2020) and cognitive function seem to originate from the social isolation and disrupted daily routines, (de Keijzer, 2018), and lower risk of disease (Twohig-Bennett and Jones, but also from the health and financial concerns that emerged from 2018) and mortality (Gascon, 2016). Although most of the research has COVID-19 public health crisis (Brooks, 2020). In this context, exposure focused on the physical health benefits of nature exposure, there is to nature might reduce stress and provide relaxation opportunities to growing evidence showing that both physical and visual contact with cope with the lockdown (Berto, 2014; Tendais and Ribeiro, 2020). nature is associated with better mental health outcomes (Wendelboe- Indeed, coincidentally, mobility data from Google shows that, in coun­ Nelson, 2019) such as lower levels of stress, anxiety (Triguero-Mas, tries with more lenient lockdowns, such as Sweden, more people used 2017) and rumination (Bratman, 2015), reduced prevalence of somati­ parks during this time, possibly to escape from home confinement zation symptoms (Triguero-Mas, 2017), improved biomarkers of stress (Google, 2020). However, socioeconomic differences are decisive for (Ribeiro, 2019), and overall better psychological well-being (Pearson nature exposure during lockdown. Population with the highest incomes and Craig, 2014). Commonly proposed mechanisms that explain these usually have better access to green space and other natural spaces mental health benefits include the provision of opportunities for phys­ (Hoffimann et al., 2017), while most vulnerable populations, especially ical and social activities along with restoration from stress and mental in large cities live in small apartments with limited access to natural fatigue (Markevych, 2017). Moreover, evidence suggests that nature spaces. Precisely, the most vulnerable and socioeconomically deprived exposure may help individuals develop and enhance their resilience populations were found to be at higher risk of psychological distress (Wu, 2013); i.e., the ability to cope with stressful life events and associated with COVID-19 pandemic (Pierce, 2020). Lack of contact adversity (van den Berg, 2010; Marselle et al., 2019; Berto, 2014). For with natural environments might in part contribute for their worst instance, van den Berg and colleagues, in the Netherlands, found that mental health outcomes (Tendais and Ribeiro, 2020). individuals with a high percentage of green space within 3 km from the Yet, as far as we know, no investigation to date has been conducted residential location were less affected by experiencing a stressful life to assess to what extent regular exposure to nature alleviates the nega­ event (e.g. unemployment, financial crisis) than those with less (van den tive mental health effects of COVID-19 lockdowns. As conceptualized in Berg, 2010). Fig. 1, it is important to take into account that, for people in stricter By the beginning of 2020, societies worldwide were experiencing an lockdowns such as in Spain, exposure to private green spaces and unprecedented, disruptive event – the coronavirus disease 2019 greenery (such as indoor and balcony plants, home/courtyard gardens, (COVID-19) pandemic. The first confirmed case of COVID-19 in Europe community private gardens and green roofs) and views of nature might was identified on January 23, 2020 (Spiteri, 2020), officially reaching be greater than the exposure to public natural spaces (e.g. public parks Spain on January 31 and Portugal on March 1, 2020. According to the and gardens, woodlands, water bodies). Thus, all types of natural fea­ latest data available (from December 11, 2020), there were a total of tures and environments, indoors and outdoors, public and private, must 340,287confirmed cases of COVID-19 and 5373 deaths in Portugal, be taken into consideration. while in Spain the total number of cases reached 1,741,439 with 47,624 Bearing this in mind, the main research hypothesis underlying this related deaths (DGS, 2020). In the absence of effective pharmacological study is that individuals with greater exposure to nature would suffer solutions, the COVID-19 pandemic led many governments to implement less from the negative mental health effects of COVID-19 lockdown general residential lockdowns, where non-essential mobility, services compared to those with less exposure. Also, due to the different mobility and events were suspended, and people were requested to stay-at-home restrictions, a second hypothesis is that the beneficial effects of private and practice physical distancing. greenery and green space might be more prominent in Spain than in Due to the different epidemiological situations, the timeline and Portugal, while those of public natural spaces would be more relevant in intensity of mobility restrictions on the Iberian Peninsula were sub­ Portugal. stantially different. In Portugal, the State of Emergency was declared on Specifically, this study aimed to explore: (i) the sociodemographic, March 19, 2020 and was renewed biweekly until May 2, 2020 (Portu­ housing and lockdown-related factors associated with changes in gueseGov., 2020), while in Spain the State of Emergency was declared exposure to nature during the COVID-19 lockdown; (ii) the associations on March 14, 2020 and remained applicable until June 21, 2020 with a between changes in exposure to nature and mental health outcomes de-escalation period in three phases (SpanishGov., 2020). In both during the COVID-19 lockdown; and (iii) whether these associations countries, the State of Emergency enforced similar mobility constraints, differed according to sociodemographic characteristics and country, as including the closure of international borders and the suspension of non- Portugal and Spain were under different mobility restrictions and essential services and events. Residents could leave their homes only to epidemiological situations. shop for basic needs, to take care of vulnerable people, to walk their dogs or dispose daily residuals, and to go to work. In Portugal and Spain, 2. Material and methods travelling to work was limited to those in essential jobs (only during two weeks over the State of Emergency, in Spain), and working from home 2.1. Study design and procedures was encouraged. Public spaces such as playgrounds and fenced parks/ gardens and some beaches were closed in both countries. The main A cross-sectional study was conducted between March 27 and May 6, difference between the lockdowns in the two nations was the fact that, in 2020. The online questionnaire was launched first in Spain (on March Portugal, residents were allowed to go out for short periods to exercise 27, 2020) and, then, the Portuguese team translated, culturally adapted, outdoors (e.g. walking or running alone), making it possible to visit and launched the online questionnaire (on April 23, 2020). During this public natural spaces, while in Spain these activities were expressly period, both countries were under a strict general lockdown. 2 A.I. Ribeiro et al. Environment International 154 (2021) 106664 The questionnaire was created using Microsoft Forms and distributed than once per week, 3 = once or twice per week, 4 = three to six times by email and social networking platforms such as Twitter, WhatsApp, per week and 5 = every day. Being this a cross-sectional study, the LinkedIn, Instagram, and Facebook. We also requested contacts to questionnaire was applied once and the data on the exposure to nature circulate the survey link among their respective personal and profes­ before the lockdown was assessed retrospectively. sional networks. Moreover, the communication offices of the Public We considered private green space and greenery, views of nature and Health Institute University of Porto (ISPUP) and the Institute of Envi­ public natural spaces, summarized in Fig. 1. More precisely, the ronmental Science and Technology (ICTA-UAB) published press releases following types of natural elements and environments were included: (i) in various media outlets to reach a wider audience and university indoor plants; (ii) balcony plants; (iii) home/courtyard garden; (iv) mailing lists were used to distribute the questionnaire as well. private community green spaces (gardens, green roofs, etc.); (v) views of Because of the questionnaire design and its dissemination strategy, nature from home (i.e. green spaces and other natural environments that the response rate could not be ascertained as it was not possible to es­ the person can observe from their home); (vi) public natural spaces (i.e. timate how many people were reached by email, social media, or via outdoor natural spaces of public use, such as street greenery, public media outlets. parks and gardens, woodlands, water bodies, etc.); and (vii) others The questionnaire included 25 questions developed by the team to (open-ended question where the participant could specify the type of measure exposure to nature (private green space and greenery, views of natural environment they were referring to). Whenever answers from nature and public natural spaces), sociodemographic, housing and (vii) matched to one of the previous types of natural elements and en­ lockdown-related characteristics, stress levels, psychological distress vironments, the variable was recoded accordingly; those that did not and somatization (the English translation of the questionnaire is pro­ match were kept in this ‘others’ category. vided in Supplementary Material 1). Before launching it, we conducted a Based on this information, a variable measuring change in exposure pilot questionnaire among a convenience sample of participants to nature was created by comparing the level of exposure before and (Portugal = 8; Spain = 7) to assess its comprehension and acceptability. during the lockdown. A dichotomous variable was created to distinguish To participate in the survey, participants had to be 18 years old or older individuals who maintained or increased exposure to nature from those and reside in Portugal or Spain during the lockdown. The questionnaire whose exposure to nature decreased during the lockdown. This was used took on average 8 min to complete. as our main exposure variable. Aiming at a statistical power of 80% and using the 95% confidence interval to assess the absence/presence of associations, a sample size of 2.3. Mental health outcomes roughly 1500 individuals (for each country), gives an effect size of 0.1, i. e., detects differences between averages or proportions even if these 2.3.1. Psychological distress differences are only 0.1 of the value of the standard deviation (Cohen, Psychological distress was measured using the General Health 2013). Questionnaire (GHQ-12) (Goldberg and Williams, 2000), a self- Ethical approval was obtained from each country’s corresponding assessment screening tool to measure common mental disorders. This authority: Ethics Committee on Animal and Human Experimentation questionnaire has been successfully used in Portuguese (Laranjeira, (CEEAH), Spain (approval number 5141) and Ethics Committee of the 2008) and Spanish (del Pilar Sánchez-López and Dresch, 2008) pop­ ISPUP, Portugal (approval number 20147). The study was carried out ulations and has good psychometric properties. The questionnaire con­ according to the Helsinki Declaration and all participants provided sisted of 12 items. In the first six items, response options assumed the informed consent. following categories – much less than usual, less than usual, same, more than usual, and do not know/do not answer – while responses to the 2.2. Exposure to nature remaining six items assumed these categories – not at all, no more than usual, a bit more than usual, much more than usual, and do not know/do Exposure to nature was assessed using seven questions covering the not answer. The study used Goldberg’s original scoring method (Gold­ frequency of visits, views, care of different types of natural elements and berg and Williams, 2000). Scores for all items were then summed, environments before and during the lockdown, each with a five-category resulting in a score ranging from 0 to 12. response scale: 1 = do not have access/not applicable, 2 = never or less Fig. 1. Schematic representation of studied types of natural elements and environments. 3 A.I. Ribeiro et al. Environment International 154 (2021) 106664 2.3.2. Somatization healthcare, etc.; 2-going to work; 3-assisting vulnerable people (e.g. As in a previous study (Triguero-Mas, 2017), seven questions were elderly, disabled); 4-walk the companion animal; 5-walk and exercise; 6- used from an adaptation of the four-dimensional symptom questionnaire take out the garbage; 7-never went out); and (v) frequency with which (4DSQ) (Terluin, 2006) to measure somatization plus two additional respondents left their homes during the lockdown (1-less than once a questions (Cronbach’s α = 0.82 for Portugal and Cronbach’s α = 0.78 for week or never, 2-one to two times a week, 3-three to six times a week, 4- Spain). Specifically, participants were asked: “During lockdown, have once a day, 5-several times a day). you suffered from: (i) dizziness/light-headed, (ii) painful muscles, (iii) Note that since data on income (both monthly disposable and change back and/or shoulder pain, (iv) headache, (v) nausea, (vi) pain in the during lockdown) and house size had a large amount of missing values in abdomen or stomach area, (vii) pain in the chest. And the additional Portugal (over 15%) these variables were not included as covariates in items were (viii) ache in the back of the head, and (ix) fatigue. The the adjusted models. questionnaires were translated and adapted to Spanish and Portuguese following the WHO guidelines (WHO. Process of translation and adap­ 2.5. Statistical analyses tation of instruments., 2020). The nine questions aiming to measure somatization had five possible responses scored as: no with a 5, some­ For the descriptive analysis, we calculated the mean or median and times with a 4, regularly with a 3, often with a 2, very often with a 1. We standard deviation or interquartile range (IQR), as well as counts and constructed a sum score of all the items ranging between 9 and 45, with proportions for each variable. For group comparisons, Chi-squared tests high scores indicating no perceived somatization symptoms (i.e. higher (or Fisher’s exact tests) were used for categorical variables, while for scores of no somatization indicated better mental health). continuous variables we used T-tests, Mann-Whitney tests or Wilcoxon- tests, according to variable type and distribution. The statistical signif­ 2.3.3. Perceived stress icance was considered at p < 0.05. Perceived stress was assessed using a question applied in a previous We fitted logistic regression models with adjustment for covariates to study on the topic (Triguero-Mas, 2017): “Please, indicate how stressed understand the sociodemographic, housing and lockdown-related (in general terms) have you felt during the lockdown on this scale”. characteristics associated with change (maintenance/increase versus Responses were recorded using a visual scale from 0 (“not at all ”) to 10 decrease) in nature exposure. These associations were expressed using (“as bad as it could be”), with a mid-point labelled “usual stress level”. the odds ratio (OR) and 95%CI. Additionally, crude and adjusted linear Because this scale was composed of a single item, it was not possible to regression models were used to estimate the associations between compute measures of internal consistency, such as Cronbach’s α. How­ change in the exposure to nature during the lockdown (maintenance/ ever, convergent validity (how closely the new scale is related to other increase vs. decrease) and (i) psychological distress, (ii) stress levels, (iii) measures of the same construct, namely from well-established and somatization. We selected the variables to be included in the adjusted validated scales (Krabbe, 2016) was possible to assess by measuring the model based on our theoretical understanding of the importance of each correlation between perceived stress and the other mental health out­ available variable in contributing to mental health during lockdown. We comes using the Spearman’s rank correlation coefficient (ρ). From this first verified that all model assumptions were met for each variable and analysis, we observed – despite the limitations of our perceived stress combination of variables. Adjusted models included mental health measure – a moderate and statistically significant correlation between outcomes and change in exposure (maintenance/increase in nature perceived stress and psychological distress (ρ for Portugal 0.53 and ρ for exposure), plus the following covariates: age, gender, household size Spain 0.42) and somatization (ρ for Portugal 0.48 and ρ for Spain 0.41) and dependents in the household, education attainment, house typol­ (Supplementary material 2). ogy, country of birth, frequency of going out during the lockdown, outside activities during the lockdown, duration of the lockdown, and 2.4. Covariates type of housing during the lockdown. The correlation matrix of the included covariates is shown in Supplementary Material 2. Note that 2.4.1. Sociodemographic and housing characteristics since data on income (both monthly disposable and change during We collected information on gender (male, female, other); age group lockdown) and house size had a large amount of missing values in (≤24 years, 25–39 years, 40–64 years, 65–79 years and ≥ 80); country Portugal (over 15%) these variables were not included as covariates in of birth (native, other European country, Non-European country); in­ the adjusted models. However, income was significantly (p < 0.001) dividual monthly disposable income before the lockdown (<600, associated with education (included in adjusted models) and house size 600–1199, 1200–1999, 2000–2999, 3000–4999, ≥5000 euros, do not was significantly (p < 0.001) associated with house typology (also know/do not answer); education attainment (primary, secondary and included in adjusted models), minimizing the risk of residual con­ tertiary); household size (number of cohabitants); and dependents in the founding. . Separate models were run for each country and type of household (none, children, elderly, disabled and other dependents). natural element and environment. Associations were expressed using Because exposure to certain types of green spaces/greenery (e.g. balcony beta and 95% Confidence Intervals (95% CI). plants, home/courtyard gardens) may be influenced by housing char­ Effect modification by gender, age and socioeconomic factors was acteristics, we also collected data on housing typology (rural single- explored by assessing the statistical significance of multiplicative in­ family house, urban single-family house, apartment and top-floor teractions using likelihood ratio tests comparing models with and apartment - which often have access to rooftop outdoor space) and without interaction terms. All analyses were run in R statistical package house size in square meters (≤45, 46–75, 76–105, 106–150, 151–200, (version 3.5.1). and ≥200). 3. Results 2.4.2. Lockdown-related characteristics Several questions about the lockdown characteristics were included 3.1. Participants’ characteristics in the questionnaire, more precisely: (i) change in income during the lockdown (severely reduced, moderately reduced, remained the same, Table 1 depicts the sociodemographic characteristics of the study increased, do not know/do not answer); (ii) type of housing where they participants. This study includes a total of 3157 participants, of which stayed during the lockdown (primary home, second-home, home from 1638 were residents from Portugal and 1519 from Spain. The majority of family/friends); (iii) the duration of the lockdown (in days) at the time the participants in both countries were women and were between 40 and of the survey response; (iv) outside activities during the lockdown (1- 64 years old. Participants from Spain were significantly older than those obtain essential goods and services, such as food, medication, from Portugal. In both countries, more than 85% of participants were 4 A.I. Ribeiro et al. Environment International 154 (2021) 106664 Table 1 Table 1 (continued ) Sample characteristics according to country (Portugal and Spain, n = 3157). Portugal (n = Spain (n = p-value Portugal (n = Spain (n = p-value 1638) 1519) country 1638) 1519) country differencesa Variable n/ %/SD n/ %/SD differencesa Variable n/ %/SD n/ %/SD mean mean mean mean Walk the companion 248 15.1 163 10.7 <0.001 Sociodemographic and housing characteristics animal Gender Walk and exercise 787 48.0 – – – Female 1229 75.1 1125 74.1 0.012 Take out the garbage 1008 61.5 1020 67.1 0.001 Male 407 24.9 386 25.4 Never went out 67 4.1 119 7.8 <0.001 Other 0 0.0 8 0.5 Other 5 0.3 8 0.5 0.489 Age group Frequency of leaving the house ≤24 y 406 24.8 78 5.1 <0.001 Less than once a week or 432 26.4 102 7.9 <0.001 25–39 y 558 34.1 518 34.1 never 40–64 y 638 38.9 827 54.4 1–2 times a week 601 36.7 729 56.3 ≥65y 36 2.2 96 6.3 3–6 times a week 297 18.1 252 19.5 Education attainment Once a day 233 14.2 120 9.3 Primary or no education 23 1.4 22 1.4 <0.001 Several times a day 75 4.6 91 7.0 Secondary 379 23.1 153 10.1 Mental health outcomes Tertiary (university 1236 75.5 1344 88.5 Psychological distress 2.9 2.0 2.7 1.9 <0.001 degree) Perceived stress 5.8 2.1 5.1 2.3 <0.001 Household size (no. 4.5 2.0 3.4 1.7 <0.001 Somatization 15.0 5.2 13.6 4.2 <0.001 inhabitants) a Chi-squared test (or Fisher’s exact test) was used for categorical variables, Dependents in the household Yes, children 473 28.9 525 34.6 <0.001 while for continuous variables we used T-test, Mann-Whitney test or Wilcoxon- Yes, elderly 113 6.9 52 3.4 <0.001 test, according to variable type and distribution. Yes, person(s) with 23 1.4 18 1.2 0.699 physical/mental natives (i.e. born in Portugal or Spain). disabilities Yes, other type of 55 3.4 20 1.3 <0.001 Most participants had a university degree, but this proportion was dependent person significantly larger in Spain (88.5%) than in Portugal (75.4%). In terms No 982 60.0 850 56.0 0.023 of geographical coverage, in Portugal, roughly 50% of the participants Housing typology resided in the Northern region, while in Spain, circa 85% resided in Rural single-family house 413 25.2 135 8.9 Catalonia. Regarding housing characteristics, the majority of the par­ <0.001 Urban single-family house 314 19.2 240 15.8 Apartment 869 53.1 970 63.9 ticipants resided in apartments, but, compared with Spain, in Portugal, Top-floor apartment 33 2.0 171 11.3 there was a significantly higher proportion of individuals residing in Other 9 0.5 3 0.2 single-family homes located in rural areas. In Portugal, there was also a House area higher proportion of individuals residing in houses that were 200 square <45 sq. m 52 3.8 60 4.0 <0.001 46–75 sq. m 155 11.2 370 24.7 meters or larger and the average number of cohabitants per household 76–105 sq. m 297 21.5 489 32.6 was significantly larger than in Spain. Overall, Spanish households were 106–150 sq. m 305 22.0 285 19.0 more likely to have dependents than those from Portugal. In Portugal, 151–200 sq. m 194 14.0 129 8.6 the proportion of households with older adults was significantly higher More than 200 sq. m 381 27.5 167 11.1 than in Spain, while the opposite was observed for the proportion of Individual monthly disposable income before the lockdown <600 euros 158 9.6 95 6.3 <0.001 households with children, which was more common in Spain. Monthly 600–1199 euros 578 35.3 266 17.5 disposable income differed between countries as well: in Portugal, the 1200–1999 euros 378 23.1 535 35.2 most common income bracket was ‘600–1199 euros’, while in Spain it 2000–2999 euros 167 10.2 390 25.7 was ‘1200–1999 euros’. 3000–4999 euros 33 2.0 126 8.3 5000 euros or more 12 0.7 14 0.9 More than one-quarter of the participants, in both countries, expe­ Country of birth rienced moderate to severe reduction in income during the lockdown Native 1466 89.5 1326 87.3 0.137 period. In both countries, more than 90% of the participants remained in Other European country 75 4.6 89 5.9 their primary home during the lockdown period (particularly in Spain). Non-European country 97 5.9 104 6.8 Regarding mobility, although in Spain movement restrictions were Lockdown related variables Income change during the lockdown stricter, there was a higher proportion of Spanish participants leaving Severely reduced 200 14.2 191 13.2 0.394 the home several times a day (7.0% vs. 4.6% in Portugal) and a lower Moderately reduced 238 16.9 224 15.5 proportion saying they left the home less than once a week or never Remained the same 958 68.0 1023 70.7 (7.9% vs. 26.4% in Portugal). In both countries, the main reasons for Increased 13 0.9 9 0.6 Type of housing during the lockdown leaving the home were obtaining essential goods and services (83.8% in Main home 1506 92.0 1432 94.3 0.002 Portugal and 86.0% in Spain) and taking out the garbage (61.5% in Second-home 61 3.7 23 1.5 Portugal and 67.1% in Spain). In Portugal, 48.0% of the participants Family/friends home 66 4.0 59 3.9 referred leaving the home to walk or exercise, while in Spain those ac­ Other 4 0.2 4 0.3 tivities were not allowed. Due to the different survey starting dates, Lockdown duration 32.8 10.9 14.1 8.7 <0.001 (days) there was a large difference in the time spent in lockdown when par­ Outside activities during the lockdown ticipants answered the questionnaire between the two countries (32.8 Obtaining essential goods 1373 83.8 1306 86.0 0.101 days in Portugal vs. 14.1 days in Spain on average). There were also and services (e.g. food, small but significant differences in health outcomes between the two medicines, healthcare) Going to work 396 24.2 159 10.5 <0.001 countries. Portuguese participants presented significantly higher scores Assisting vulnerable 270 16.5 123 8.1 <0.001 of psychological distress (2.9 in Portugal vs. 2.7 in Spain) and somati­ people (e.g. elderly, zation (15.0 in Portugal vs. 13.6 in Spain), as well as higher levels of disabled) stress (5.8 in Portugal vs. 5.1 in Spain). 5 A.I. Ribeiro et al. Environment International 154 (2021) 106664 3.2. Nature exposure before and during COVID-19 lockdown lockdown. In Spain, both presence of children (OR 3.14; 95%CI 1.20, 7.95) and total absence of dependents in the household (OR 2.90; 95%CI In Fig. 2, we show the frequency of nature exposure before and 1.21, 6.53) were both associated with a greater likelihood of exposure to during the COVID-19 lockdown according to country by different nat­ indoor plants. Residing in apartments or top-floor apartments was also ural elements and environments. In both countries, as expected, there associated with a greater likelihood of increased or maintained exposure was a significant reduction in the use of public natural spaces. For to home/courtyard gardens in Portugal (OR 3.49; 95%CI 1.88, 6.68) and instance, in Spain, the proportion of people that visited this type of in Spain (OR 2.75; 95%CI 1.17, 6.03) in comparison to residing in rural spaces everyday went from 20.8% (before the lockdown) to 9.0% single-family houses. A similar trend was seen for balcony plants in (during the lockdown). Portugal (OR 2.01; 95%CI 1.22, 3.32). On the other hand, spending the Regarding other types of natural elements and spaces, changes in lockdown in second-homes or the homes of family/friend (in compari­ patterns of exposure before and during the lockdown were less pro­ son to spending it in the main home) was associated with lower likeli­ nounced or even in the opposite direction. For instance, the frequency of hood of increasing/maintaining exposure to indoor (Portugal: OR 0.41; exposure to private green space and greenery increased in both coun­ 95%CI 0.24, 0.74; Spain: OR 0.43; 95%CI 0.21, 0.95) or to balcony tries. The overall frequency of the remaining forms of nature exposure plants (Spain: OR 0.23; 95%CI 0.12, 0.47). Another detrimental (views from home and community private green spaces) was not circumstance for contact with indoor plants was living in households significantly altered, as there were equivalent proportions of people that with more cohabitants (Spain: OR 0.82; 95%CI 0.71, 0.96) or being increased and reduced their utilization after the implementation of the older (Spain: OR 0.28; 95%CI 0.07, 0.95). lockdown, in both countries. The frequency of outside activities during lockdown was positively The proportion of participants who increased or maintained the associated with increase/maintenance of exposure to house home/ exposure to nature during the lockdown were: 92.4% in Portugal and courtyard gardens, chiefly in Portugal. Leaving the house for work 91.2% in Spain for indoor plants; 94.0% and 92.6% for balcony plants; (Portugal: OR 0.34; 95%CI 0.18, 0.63) and to obtain essential goods/ 95.4% and 95.3% for home/courtyard gardens; 90.9% and 90.8% for services (Spain: OR 0.16; 95%CI 0.09, 0.77) were negatively associated community private green spaces; 80.0% and 87.2% for views of nature with maintaining/increasing the use of house gardens. from home; 46.2% and 34.0% for public natural spaces; and finally, Concerning community private green spaces, in Portugal, individuals 92.3% and 89.8% for the other natural spaces. staying in second-homes or family/friends’ home during the lockdown (OR 0.42; 95%CI 0.25, 0.74) and those who left home to work (OR 0.59; 95%CI 0.38, 0.94) were less likely to maintain/increase the use of these 3.3. Characteristics associated with changes in nature exposure natural elements and environments, although the frequency of leaving home showed a positive association with this outcome. In Spain, adults Table 2 shows the sociodemographic, housing and lockdown-related aged ≥ 65 years (OR 0.07; 95%CI 0.01, 0.28) were less likely to main­ characteristics associated with maintenance/increase in nature expo­ tain/increase the exposure to community private green spaces and the sure during the lockdown. As mentioned before, the frequency of likelihood of use also decreased with time spent in lockdown (OR 0.98, exposure to private green space and greenery increased during the Fig. 2. Exposure to nature before and during COVID-19 lockdown, according to country (Portugal and Spain, n = 3157). 6 Table 2 A.I. Ribeiro et al. Adjusted associations (odds ratio and 95% confidence intervals) between sociodemographic, housing and lockdown-related variables and maintenance/increase in nature exposure during COVID-19 lockdown (Portugal and Spain, n = 3157). Variable Indoor plants Balcony plants House/courtyard garden Community private green Views of nature Public natural spaces Other space Portugal Spain Portugal Spain Portugal Spain Portugal Spain Portugal Spain Portugal Spain Portugal Spain Sociodemographic and housing characteristics Male (ref: 1.38 0.84 1.00 0.80 0.81 1.12 0.69 1.19 1.12 1.07 1.08 1.30 1.54 1.43 female) (0.87–2.24) (0.54–1.32) (0.61–1.67) (0.49–1.32) (0.46–1.46) (0.60–2.21) (0.47–1.05) (0.75–1.95) (0.82–1.53)a (0.74–1.59) (0.85–1.38) (1.00–1.71) (0.97–2.41) (0.92–2.28) Age group (ref: ≤24 years) 25–39 years 1.29 0.83 1.54 0.83 1.66 0.48 1.52 0.38 1.38 0.53 1.49 1.06 0.78 0.14 (0.70–2.38) (0.22–2.47) (0.80–2.98) (0.18–2.80) (0.75–3.72) (0.03–2.79) (0.88–2.62) (0.06–1.44) (0.95–2.01) (0.17–1.39) (1.08–2.07) (0.55–2.08) (0.43–1.42) (0.01–0.73)* * 40–64 years 0.79 0.59 1.63 0.70 1.19 0.41 1.59 0.32 2.26 0.51 1.74 1.28 0.78 0.12 (0.43–1.44) (0.16–1.64) (0.82–3.27) (0.16–2.25) (0.54–2.63) (0.02–2.27) (0.90–2.83) (0.05–1.18) (1.50–3.41) (0.16–1.32) (1.24–2.44) (0.68–2.48) (0.42–1.44) (0.01–0.57)* * * ≥65 years 0.93 0.28 0.78 0.35 1.53 0.15 0.95 0.07 1.69 0.51 1.78 1.68 0.54 0.11 (0.29–4.18) (0.07–0.95) (0.26–2.89) (0.07–1.38) (0.27–28.95) (0.01–1.02) (0.32–3.49) (0.01–0.28) (0.70–4.72) (0.14–1.74) (0.86–3.73) (0.78–3.69) (0.18–1.98) (0.01–0.61)* * * Tertiary 1.27 1.29 1.13 1.02 1.32 0.51 0.85 1.69 1.03 1.17 1.29 0.79 1.54 1.55 education (0.79–2.00) (0.70–2.26) (0.67–1.87) (0.48–1.99) (0.74–2.33) (0.15–1.35) (0.54–1.32) (0.95–2.93) (0.75–1.41) (0.66–1.99) (0.99–1.68) (0.54–1.16) (0.97–2.41) (0.86–2.71) (ref: primary/ secondary) b Household 1.01 0.82 1.05 0.96 0.87 0.89 0.98 0.88 0.91 0.92 1.02 0.98 0.95 1.01 size (no.) (0.90–1.14) (0.71–0.96) (0.92–1.19) (0.81–1.14) (0.75–1.01) (0.72–1.11) (0.88–1.09) (0.75–1.03) (0.84–0.98) (0.81–1.04) (0.96–1.08) (0.89–1.08) (0.85–1.07) (0.88–1.18) * * Dependents in the household 7 Yes, children 1.02 3.14 1.03 2.24 1.10 1.86 0.53 1.76 1.08 0.47 0.65 0.58 0.69 1.07 (ref. no) (0.40–2.60) (1.20–7.95) (0.37–2.97) (0.71–6.62) (0.33–3.91) (0.43–7.04) (0.22–1.31) (0.66–4.57) (0.54–2.15) (0.16–1.21) (0.38–1.11) (0.32–1.07) (0.27–1.77) (0.41–2.68) * Yes, older 1.36 2.27 0.55 2.70 0.92 1.78 0.41 1.17 0.97 0.52 0.89 0.82 0.54 0.40 adults (ref. (0.53–3.91) (0.72–7.99) (0.21–1.61) (0.66–14.38) (0.31–3.18) (0.32–14.60) (0.18–0.98)* (0.38–4.07) (0.49–1.98) (0.17–1.66) (0.52–1.52) (0.37–1.79) (0.23–1.39) (0.15–1.10) no) Yes, disabled 0.63 1.43 0.78 0.67 0.80 0.97 1.90 0.70 1.71 4.30 1.23 1.53 0.34 0.80 or (0.24–1.81) (0.44–5.72) (0.25–2.87) (0.19–2.76) (0.23–3.34) (0.20–7.34) (0.57–8.83) (0.23–2.58) (0.74–4.25) (0.78–80.72) (0.66–2.31) (0.66–3.49) (0.13–0.97) (0.25–3.12) dependent * person(s) (ref. no) b No (ref. yes) 1.17 2.90 0.79 1.51 1.14 1.76 0.66 1.55 1.00 0.45 0.90 0.82 0.50 0.91 (0.47–2.81) (1.21–6.53) (0.28–2.16) (0.55–3.73) (0.32–4.11) (0.46–5.62) (0.27–1.61) (0.64–3.52) (0.51–1.94) (0.16–1.08) (0.54–1.50) (0.48–1.40) (0.19–1.27) (0.38–2.02) * Housing typology (ref: Rural single-family house) Urban single- 0.87 0.94 1.40 0.97 1.14 1.12 1.07 1.05 0.81 0.28 0.71 0.49 0.78 0.44 Environment International 154 (2021) 106664 family (0.50–1.52) (0.43–1.97) (0.79–2.55) (0.41–2.22) (0.64–2.10) (0.45–2.70) (0.63–1.86) (0.46–2.35) (0.56–1.18) (0.10–0.67) (0.52–0.96) (0.30–0.82) (0.45–1.34) (0.16–1.07) house * * * Apartment 0.98 1.17 2.01 1.23 3.49 2.75 0.78 1.14 0.84 0.30 0.67 0.54 1.03 0.44 or top- (0.61–1.57) (0.58–2.23) (1.22–3.32) (0.56–2.53) (1.88–6.68)* (1.17–6.03)* (0.50–1.20) (0.54–2.21) (0.61–1.15) (0.11–0.66) (0.52–0.87) (0.35–0.84) (0.64–1.65) (0.17–0.98)* floor * * * * apartment b Country of birth (ref: native) Other 0.96 0.60 0.99 1.73 0.35 1.49 0.54 0.62 1.01 0.72 1.20 0.64 0.73 0.82 European (0.45–2.38) (0.28–1.38) (0.41–2.96) (0.59–7.44) (0.16–0.87)* (0.42–9.53) (0.27–1.23) (0.30–1.43) (0.56–1.93) (0.39–1.40) (0.74–1.95) (0.36–1.11) (0.35–1.72) (0.41–1.85) country (continued on next page) A.I. Ribeiro et al. Table 2 (continued ) Variable Indoor plants Balcony plants House/courtyard garden Community private green Views of nature Public natural spaces Other space Portugal Spain Portugal Spain Portugal Spain Portugal Spain Portugal Spain Portugal Spain Portugal Spain Other World 1.17 1.00 0.85 0.71 0.80 3.87 0.92 0.68 0.76 0.70 1.17 0.64 0.94 1.25 country (0.55–2.88) (0.46–2.50) (0.38–2.29) (0.32–1.80) (0.32–2.48) (0.81–69.62) (0.47–1.98) (0.34–1.55) (0.46–1.30) (0.39–1.33) (0.76–1.80) (0.36–1.07) (0.46–2.22) (0.59–3.10) Lockdown-related variables Type of housing during the lockdown (ref: main home) Second- 0.41 0.43 0.79 0.23 0.66 0.79 0.42 1.24 0.81 0.31 0.77 0.64 0.95 0.77 home or (0.24–0.74) (0.21–0.95) (0.41–1.65) (0.12–0.47) (0.31–1.52) (0.28–2.87) (0.25–0.74)* (0.50–3.76) (0.53–1.28) (0.17–0.60) (0.52–1.15) (0.34–1.15) (0.50–1.96) (0.35–1.88) family/ * * * * friends’ home b Lockdown 1.00 1.02 1.01 1.02 1.01 1.02 1.01 0.98 1.00 1.02 0.99 1.00 0.98 1.03 duration (0.98–1.02) (0.99–1.04) (0.99–1.04) (0.99–1.06) (0.99–1.04) (0.98–1.06) (0.99–1.02) (0.96–1.00) (0.98–1.01) (1.00–1.04) (0.98–1.00) (0.99–1.02) (0.96–1.00) (1.01–1.06)* (days) * * * Outside activities during the lockdown Obtain 0.90 1.16 1.37 0.66 0.71 0.16 1.24 0.49 1.05 0.38 0.91 0.72 0.84 0.82 essential (0.48–1.59) (0.52–2.38) (0.73–2.46) (0.21–1.67) (0.29–1.55) (0.01–0.77)* (0.73–2.04) (0.17–1.15) (0.72–1.51) (0.14–0.87) (0.66–1.25) (0.45–1.16) (0.44–1.50) (0.33–1.79) goods/ * services (ref: no) Going to 0.94 0.59 0.62 0.60 0.34 0.84 0.59 0.63 0.95 0.61 0.72 1.41 0.82 1.47 work (ref: (0.58–1.58) (0.33–1.08) (0.36–1.08) (0.32–1.17) (0.18–0.63)* (0.36–2.22) (0.38–0.94)* (0.34–1.20) (0.68–1.35) (0.38–1.03) (0.55–0.95) (0.96–2.07) (0.51–1.36) (0.75–3.19) no) * Assisting 1.19 0.80 1.17 1.28 0.83 0.72 0.91 1.04 1.40 1.93 1.25 1.12 0.86 2.38 8 vulnerable (0.72–2.08 (0.42–1.63) (0.64–2.29) (0.57–3.43) (0.44–1.66) (0.32–1.84) (0.56–1.52) (0.53–2.26) (0.97–2.07) (0.98–4.29) (0.95–1.66) (0.73–1.70) (0.53–1.43) (1.07–6.35)* people (ref: no) Walk the 0.95 0.50 1.65 0.90 0.99 0.61 0.73 1.18 1.17 0.63 1.18 1.75 1.14 0.39 companion (0.56–1.67) (0.25–1.04) (0.84–3.57) (0.40–2.17) (0.48–2.19) (0.22–1.87) (0.44–1.26) (0.55–2.70) (0.81–1.72) (0.34–1.22) (0.87–1.60) (1.10–2.77) (0.67–2.03) (0.20–0.77)* animal * (ref: no) Walk and 1.03 na 0.99 na 1.06 na 1.21 na 1.22 na 1.49 na 0.91 na exercise (0.68–1.55) (0.62–1.58) (0.63–1.80) (0.83–1.77) (0.93–1.60) (1.20–1.86) (0.60–1.37) (ref: no) * Take out the 1.02 1.16 0.96 1.30 1.23 1.12 0.92 1.03 1.14 1.21 0.91 0.96 0.90 0.70 garbage (0.67–1.54) (0.72–1.83) (0.60–1.52) (0.76–2.16) (0.72–2.07) (0.58–2.08) (0.62–1.34) (0.63–1.65) (0.87–1.50) (0.81–1.79) (0.73–1.15) (0.72–1.29) (0.59–1.36) (0.42–1.12) (ref: no) Frequency of leaving the house (ref: Less than once a week or never) 1–2 times a 1.30 0.72 1.00 0.54 1.14 2.20 1.34 1.60 0.87 1.68 1.31 0.97 1.23 2.25 week (0.75–2.23) (0.22–1.98) (0.55–1.80) (0.13–1.73) (0.58–2.20) (0.58–6.72) (0.85–2.09) (0.70–3.46) (0.62–1.22) (0.75–3.50) (0.98–1.75) (0.57–1.69) (0.71–2.11) (0.98–4.85) Environment International 154 (2021) 106664 3–6 times a 0.86 1.21 1.78 0.79 2.06 4.96 4.23 1.95 1.15 1.20 1.39 0.73 0.88 2.12 week (0.46–1.63) (0.34–3.80) (0.79–4.19) (0.18–2.99) (0.86–5.13) (1.06–21.68) (2.14–8.85)* (0.75–4.97) (0.74–1.82) (0.50–2.71) (0.97–1.99) (0.40–1.36) (0.46–1.67) (0.85–5.14) * Once a day 0.78 0.82 0.91 0.34 2.99 2.22 3.07 1.09 0.76 2.50 1.88 1.43 1.03 2.27 (0.40–1.54) (0.22–2.69) (0.42–2.01) (0.08–1.30) (1.08–9.36)* (0.47–9.60) (1.55–6.43)* (0.39–2.98) (0.48–1.22) (0.91–6.87) (1.28–2.78) (0.73–2.82) (0.51–2.11) (0.81–6.38) * Several times 1.62 3.23 0.96 0.58 4.71 3.14 3.64 1.23 0.88 1.57 1.80 1.56 0.58 5.19 a day (0.55–6.04) (0.71–16.36) (0.32–3.34) (0.12–2.71) (1.08–33.57) (0.48–20.86) (1.36–11.75) (0.38–4.18) (0.44–1.83) (0.53–4.70) (1.02–3.20) (0.73–3.35) (0.24–1.52) (1.57–18.12) * * * * a Odds ratio (95% Confidence Interval); bcategories were merged due to small counts; bold and asterisk denotes significant association. A.I. Ribeiro et al. Environment International 154 (2021) 106664 Table 3 Crude and adjusted associations (beta and 95% Confidence Intervals) between mental health outcomes and maintenance/increase in nature exposure during COVID-19 lockdown (Portugal and Spain, n = 3157). Psychological distress Somatization Stress level Portugal Spain Portugal Spain Portugal Spain Nature Crude Adjusted Crude Adjusted Crude Adjusted Crude Adjusted Crude Adjusted Crude Adjusted exposure Indoor Plants − 0.35 − 0.36 − 0.18 − 0.35 0.04 0.13 − 0.41 − 0.70 − 0.28 − 0.29 − 0.30 ¡0.52 (− 0.72, (− 0.72, (− 0.52, (− 0.72, (− 0.89, (− 0.76, (− 1.16, (− 1.48, (− 0.67, (− 0.66, (− 0.72, (− 0.96, 0.01) 0.00) 0.16) 0.03) 0.98) 1.02) 0.34) 0.09) 0.11) 0.10) 0.11) ¡0.07)* Balcony plants − 0.25 − 0.22 0.18 0.01 0.74 0.79 0.30 0.03 − 0.13 − 0.14 0.27 − 0.07 (− 0.66, (− 0.61, (− 0.19, (− 0.41, (− 0.31, (− 0.22, (− 0.52, (− 0.85, (− 0.57, (− 0.57, (− 0.17, (− 0.57, 0.16) 0.19) 0.54) 0.42) 1.78) 1.79) 1.11) 0.92) 0.30) 0.28) 0.72) 0.43) Home/ 0.02 0.08 0.27 0.33 0.57 0.62 0.11 0.32 0.01 − 0.01 0.60 0.55 courtyard (− 0.45, (− 0.39, (− 0.18, (− 0.18, (− 0.62, (− 0.53, (− 0.89, (− 0.78, (− 0.48, (− 0.51, (0.05, (− 0.07, garden 0.48) 0.54) 0.71) 0.84) 1.76) 1.78) 1.11) 1.42) 0.51) 0.48) 1.15) 1.17) Community − 0.29 − 0.23 − 0.10 − 0.28 − 0.50 − 0.37 − 0.40 ¡0.82 − 0.31 − 0.26 0.35 0.17 private (− 0.63, (− 0.56, (− 0.43, (− 0.65, (− 1.36, (− 1.20, (− 1.13, (− 1.61, (− 0.67, (− 0.61, (− 0.05, (− 0.28, green space 0.05) 0.11) 0.22) 0.09) 0.37) 0.46) 0.34) ¡0.03)* 0.05) 0.10) 0.76) 0.61) Views of ¡0.36 ¡0.27 0.12 0.16 ¡1.15 ¡0.79 − 0.05 − 0.25 ¡0.61 ¡0.48 0.13 0.21 nature (− 0.60, (− 0.51, (− 0.17, (− 0.15, (− 1.76, (− 1.39, (− 0.69, (− 0.90, (− 0.87, (− 0.74, (− 0.22, (− 0.16, ¡0.12)* ¡0.03)* 0.41) 0.46) ¡0.53)* ¡0.20)* 0.59) 0.41) ¡0.35)* ¡0.23)* 0.48) 0.58) Public natural ¡0.23 − 0.15 − 0.09 0.02 ¡0.78 − 0.43 ¡0.79 − 0.44 ¡0.43 ¡0.29 ¡0.39 − 0.13 spaces (− 0.42, (− 0.34, (− 0.29, (− 0.20, (− 1.28, (− 0.92, (− 1.24, (− 0.91, (− 0.64, (− 0.49, (− 0.64, (− 0.40, ¡0.04)* 0.05) 0.11) 0.24) ¡0.28)* 0.06) ¡0.34)* 0.04) ¡0.23)* ¡0.08)* ¡0.15)* 0.14) Other natural − 0.17 − 0.17 0.00 − 0.08 − 0.75 − 0.58 ¡0.78 ¡1.06 − 0.13 − 0.13 0.07 0.12 space/ (− 0.54, (− 0.53, (− 0.31, (− 0.42, (− 1.68, (− 1.48, (− 1.48, (− 1.79, (− 0.52, (− 0.51, (− 0.31, (− 0.29, element 0.19) 0.19) 0.31) 0.26) 0.18) 0.31) ¡0.08)* ¡0.32)* 0.26) 0.25) 0.46) 0.54) Adjusted models included mental health outcomes and change in exposure (maintenance/increase in nature exposure), plus the following covariates: age, gender, household size and dependents in the household, education attainment, house typology, country of birth, frequency of going out during the lockdown, outside ac­ tivities during the lockdown, duration of the lockdown, and type of housing during the lockdown. Bold and asterisk denotes significant association. 95%CI 0.96, 1.00). natural spaces. In both countries and principally in Portugal, main­ Although almost the same proportion of people maintained/ taining/increasing the use of public natural spaces was positively asso­ increased and decreased the frequency of viewing nature from home ciated with frequency of leaving the house. during the lockdown, the maintenance/increase of that exposure Finally, regarding any other types of natural elements and environ­ showed association with several variables. In Portugal, individuals aged ments (like, for instance, agricultural fields and workplace greenery), we 40–64 years were more likely than other age groups to maintain/in­ observed that in Spain those living in apartments and top-floor apart­ crease views of nature (OR 2.26; 95%CI 1.50, 3.41), while those living in ments (OR 0.44, 95%CI 0.17, 0.98) and those who left the house to walk households with more cohabitants were less likely to do so (OR 0.91; with companion animals (OR 0.39, 95%CI 0.20, 0.77) were less likely to 95%CI 0.84, 0.98). In Spain, individuals residing in urban single-family maintain/increase the use of these elements/environments; the recip­ houses (OR 0.28; 95%CI 0.10, 0.67), in apartments/top-floor apart­ rocal happened with younger people and with those who went out more ments (OR 0.30; 95%CI 0.11, 0.66), those who stayed in second homes frequently (several times a day: OR 5.19, 95%CI 1.57, 18.12). In Spain, or family/friends’ homes during the lockdown (OR 0.31; 95%CI 0.17, lockdown duration was positively associated with maintaining or 0.60), and those who went out to obtain essential goods and services (OR increasing the use of these type of elements/environments, but the 0.38; 95%CI 0.14, 0.87) were less likely to maintain/increase views of opposite trend was observed in Portugal. In addition, in Portugal, those nature. Conversely, time spent in lockdown was positively associated living in households with disabled and vulnerable individuals were less with maintaining/increasing views of nature in Spain (OR 1.02; 95%CI likely (OR 0.34, 95%CI 0.13, 0.97) to report using these other natural 1.00, 1.04). spaces. Concerning public natural spaces, keeping in mind the most common scenario was the reduction of exposure, we see that the emergence of the 3.4. Exposure to nature and mental health outcomes opposite trend was favored by increasing age, both in Portugal and in Spain. Housing typology also played a role, as in the two countries in­ Table 3 shows the association between maintaining/increasing dividuals living in urban single-family houses (Portugal: OR 0.71; 95%CI exposure to nature during the lockdown and the three studied mental 0.52, 0.96; Spain: OR 0.49; 95%CI 0.30, 0.82) and apartments/top-floor health indicators. apartments (Portugal: OR 0.67, 95%CI 0.52–0.87; Spain: OR 0.54, 95% We observed that, in Portugal, maintaining/increasing exposures to CI 0.35–0.84) were less likely to maintain or increase the use of public public natural spaces (via visits) and views of nature from home were natural spaces during the lockdown. The circumstances involving associated with lower levels of psychological distress. Yet, after adjust­ lockdown were also important, namely in Portugal: the likelihood of ment for sociodemographic, housing and lockdown-related character­ maintaining/increasing exposure to public natural spaces was slightly istics, only the second association remained, where we observed that reduced with more time spent in lockdown (OR 0.99; 95%CI 0.99, 1.00); participants maintaining/increasing views to nature during the lock­ those who went to work were less likely to maintain/increase exposure down had significantly lower levels of psychological distress (adjusted (OR 0.72; 95%CI 0.55, 0.95); whereas the likelihood of increasing/ beta − 0.27; 95%CI − 0.51, − 0.03). maintaining exposure was greater for those who went out for exercise In both countries, crude models showed that maintaining/increasing than those who did not (OR 1.49; 95%CI 1.20, 1.86). In Spain, only exposure to public natural spaces and views of nature were associated walking a companion animal showed a significant association (OR 1.75; with lower levels of somatization. After adjustment, only the negative 95%CI 1.10, 2.77) with maintaining or increasing exposure to public relationship between views of nature and somatization in Portugal 9 A.I. Ribeiro et al. Environment International 154 (2021) 106664 remained (adjusted beta − 0.79; 95%CI − 1.39, − 0.20). In Spain, these associations may be related to two widely known complementary adjusted models revealed that maintaining/increasing the exposure to theories: stress-recovery theory (SRT) and the attention restoration community private green spaces (adjusted beta − 0.82; 95%CI − 1.61, theory (ART). According to the STR, contact with nature can reduce − 0.03) and other spaces (adjusted beta − 1.06; 95%CI − 1.79, − 0.32) physiological responses to stress (Ulrich, 1983), while the ART suggests was significantly and negatively associated with somatization. that exposure to nature may reduce directed attention fatigue through Finally, in Portugal, in both crude and adjusted models, maintain­ attention restoration (Kaplan, 1995). Besides, residential proximity to ing/increasing exposure to public green spaces (adjusted beta − 0.29; public green spaces was shown to promote physical activity, which may 95%CI − 0.49, − 0.08) and views of nature were negatively associated be a relevant pathway in the association between mental health and with stress (adjusted beta − 0.48; 95%CI − 0.74, − 0.23). In Spain, we public green spaces observed in Portugal (Mitchell, 2013), as brief found that individuals who maintained//increased exposure to indoor outdoor physical activity was allowed during the lockdown period. plants presented significantly lower levels of stress even after accounting In Portugal, views of nature were beneficial to mental health, for confounding variables (adjusted beta − 0.52; 95%CI − 0.96, − 0.07). regardless of the specific outcome, which is in accordance with the No significant interaction effects according to gender, age, or so­ published literature about the nature-health interconnection. Window cioeconomic factors were observed, suggesting that the mental health views of natural environments were found to reduce perceived stress benefits of nature exposure were felt equally across sociodemographic levels and stress-related biological markers (Nutsford, 2016; Kaplan, groups. 2001). In Spain, this effect was not significant, which might be related Supplementary material 3 shows the regression equations from the with the type of landscapes that are visible from the homes of the studied adjusted models. sample; while in Portugal 25.2% of individuals resided in single-family houses in rural areas, in Spain this percentage was only 8.9%, suggesting 4. Discussion that they are more likely to be living in urban or semi-urban landscapes with less natural features, which do not bring the same relaxation and We found that exposure to nature brought mental health benefits restorative effects as rural and wild landscapes (Ulrich, 1979). during the lockdown period established to prevent the spread of the COVID-19 pandemic. We also found a significant reduction in the uti­ 4.1. Strengths and limitations lization of public natural spaces in both Portugal and Spain (despite access was only explicitly forbidden in Spain) and a slight increase in This study has some important limitations. First, the data and ana­ contact with private green space and greenery (indoor plants, home/ lyses were derived from a cross-sectional study, making it difficult to courtyard gardens, etc.). In Portugal, maintaining or increasing the make causal inferences. Second, internet-based voluntary recruitment routine of visiting public natural spaces and contemplating those spaces could have introduced important selection bias, first by excluding peo­ from home was associated with lower levels of stress, somatization and ple without internet access or who do not use social media, and second psychological distress, while in Spain, exposure to private green space by introducing self-selection bias, perhaps attracting respondents with and greenery, namely indoor plants and community private green particular interest in the topic. According to the latest data, the pro­ spaces, were found to be associated with lower levels of stress and less portion of households with connection to Internet at home is 84.5% in somatization of symptoms. Portugal (INE, 2020) and 95.4% in Spain (Encuesta, 2020) and the This finding is in accordance with our initial hypothesis that in Spain, percentage of individuals using social media is 80.2% (Inquérito, 2019) where mobility restrictions during the lockdown were more severe, and 64.7% (Encuesta, 2020), in Portugal and Spain, respectively. This private green space and greenery would play a more prominent role than means that a non-negligible share of individuals was unable to answer to in Portugal, where brief outdoor leisure and physical activities were the questionnaire. In our study we have a highly unbalanced gender allowed. Although evidence on the health benefits of private green ratio (more than 70% of women as observed in other web-based surveys space/greenery is less documented, a few studies have demonstrated (Rossi, et al., 2020), an overrepresentation of highly educated in­ that taking care of and having contact with indoor plants is associated dividuals and of individuals from Catalonia (85% of the Spanish sample) with reduced psychological stress (Lee, 2015), reduced strain and better and the Portuguese Northern region (50% of the Portuguese sample). recovery among hospitalized patients (Park and Mattson, 2009), and Supplementary material 4 compares the demographic composition of better mood (Shibata and Suzuki, 2001). For example, an experimental our sample with the population composition of both countries and study found active interaction with indoor plants can reduce physio­ shows that our sample is composed of individuals who are more logical and psychological stress (subjects felt more comfortable and educated than the overall population, with an overrepresentation of soothed) compared with mental/computer work (Lee, 2015). No sig­ women and an underrepresentation of older adults. This unbalanced nificant associations were observed between maintaining/increasing composition might have limited our capacity to detect socioeconomic home/courtyard garden exposure and mental health outcomes, but, in differences and interactions between nature exposure and mental health Spain, a significant beneficial effect was observed for private community outcomes, which is particularly important given the inequities in green spaces, where those that maintained or increased their exposure to housing quality and size by socioeconomic status with important im­ these spaces presented lower levels of somatization. Because these are plications for access to green space at home. Plus, this fact also implies communal spaces (gardens, green roofs, etc.), it is plausible that, beyond that our results cannot be straightforwardly generalized to the overall gardening and gazing, these spaces offer opportunities for socialization population. While the questionnaire was disseminated throughout all with neighbors, thus improving social integration and community the community, regardless their level of interest on the topic, it is cohesion (Yotti Kingsley and Townsend, 2006) which is critical in plausible that those who accepted to participate were more interested on adverse life moments (Wu, 2013). They also offer greater opportunity the health benefits of nature exposure and that could possibly add some for sports and other physical activities, which might also help alleviate form of selection and information bias. Third, all data used in this stress and anxiety. investigation (both exposure and outcome) is self-reported which can be In Portugal, maintaining or increasing the frequency of visiting a source of recall or social desirability bias, which may have skewed public natural spaces (e.g. public parks, street greenery, woodlands, reporting of time spent outside or use of secondary homes during lock­ water bodies etc.) was associated with better mental health outcomes, down as these activities were strictly regulated during the lockdown, more precisely, with lower levels of stress. This is in accordance with a particularly in Spain. In addition, data on nature exposure before the growing body of literature demonstrating the mental health benefits of lockdown were obtained retrospectively. While retrospective data contact with outdoor green space and natural environments (Triguero- collection may introduce recall errors, in this study recall errors are Mas, 2015; van den Berg, 2016; Gascon, 2015). The mechanisms beyond likely non-differential, which is considered a less serious problem in 10 A.I. Ribeiro et al. Environment International 154 (2021) 106664 epidemiological studies (as compared to differential recall errors) and greening strategies that would work in that specific geographic, physical usually lead to an increase in variability and reduces the power of the and social context. In addition, future investigations should ascertain the study. Forth, despite it has been used in previous studies in similar risk of infection in public natural spaces in order to better balance the contexts (Triguero-Mas, 2017) and the stress measure is moderately mental health effects of lockdowns with the need to protect residents correlated with the other, well-established and validated measures of from infection. mental well-being we studied (see Supplementary material 2), revealing a reasonable convergent validity, this variable presents a set of limita­ 5. Conclusion tions that should be discussed. Because it is composed by a single item we cannot compute measures of internal consistency such as Cronbach’s Maintaining or increasing exposure to private green space and green­ α. Consequently, the psychometric characteristics of our stress measure ery, to public natural spaces and views of nature was associated with are unknown and, therefore, there is a possibility that a different improved mental health outcomes during the COVID-19 pandemic and construct is measured. In addition, the item formulation “how stressed subsequent general lockdowns, both in Portugal and Spain. These associ­ have you felt” may not mean the same for every respondent, since people ations varied between countries possibly reflecting the different mobility might conceptualize stress in many different ways. Thus, our results for restrictions and epidemiological situation experienced by the two nations. such a complex construct, as stress, should be taken with caution. Fifth, In light of these results, we suggest that policy-makers and public author­ for the sake of simplicity, no questions on personality type, reactions, ities should facilitate and organize safe and controlled opportunities for coping mechanisms and adaptations were included in the questionnaire. exposure to natural elements and environments (for visits, viewing or These factors, particularly personality, are well-established predictors of taking care of) during the highly unique and dynamic pandemic, especially mental health and well-being (Kokko et al., 2015). Thus, they could for those with low (or no) exposure to private green space/greenery. constitute important effect modifiers of the studied associations and their inclusion could provide a more in-depth knowledge on the mech­ 6. Contributions anisms connecting exposure to nature and mental health during COVID- 19 pandemic. Finally, due to the relatively few changes in the frequency AIR participated in the design of the study, performed literature of nature exposure during the lockdown in the studied sample, our main revision, data cleaning and manipulation and analysis, interpretation of exposure variable was dichotomous (maintained/increased vs. the results and drafted the manuscript. MTM participated in the decreased exposure) which did not allowed us to assess the impacts conception, design and planning of the study, assisted data cleaning and different degrees of change in the exposure to nature. supervised data analysis. CJS participated in data cleaning, manipula­ The main strength of the present study is that we were able to tion, and analysis and helped draft the manuscript. IA and HC partici­ conduct the study in real time, that is during the period of strict lock­ pated in the conception, design and planning of the study. AGN and down in both countries, minimizing the risk of recall bias; accounted for contributed in the interpretation of the results and helped draft the a wide range of natural elements and environments available to the manuscript. FB led the conception, design and planning of the study, population, including those found in the home environment, which have supervised the data analysis and contributed in the interpretation of the been poorly addressed in the literature about the connections between results. All authors reviewed and approved the final version of the nature and public health. Other strengths of the study include a large manuscript. sample size, the cross-national design, the use of validated instruments for mental health assessment, and the collection of and adjustment for a Declaration of Competing Interest wide range of potential confounding variables. The authors declare that they have no known competing financial 4.2. Policy implications and future research interests or personal relationships that could have appeared to influence the work reported in this paper. Mental disorders have an important impact on individuals, families and communities. Indeed, mental disorders rank second in terms of Acknowledgements Years of Life Lost due to Disability (YLDs) accounting for about 15% of all YLDs in Spain and in Portugal. Subsequently, mental disorders cost This study was supported by FEDER through the Operational Pro­ more than US$ 1 trillion per year to the global economy (WHO, The gramme Competitiveness and Internationalization and national funding WHO special initiative for mental health, 2019). Pandemics and asso­ from the Foundation for Science and Technology – FCT (Portuguese ciated lockdowns might exacerbate this already heavy burden by Ministry of Science, Technology and Higher Education) under the Uni­ worsening the symptoms of those with mental illness (Kozloff, 2020) dade de Investigação em Epidemiologia - Instituto de Saúde Pública da and by inducing fear, depression stress and anxiety in the general pop­ Universidade do Porto (EPIUnit) (POCI-01-0145-FEDER-006862; ulation (Armitage and Nellums, 2020; Rajkumar, 2020; Rossi, et al., Ref. UID/DTP/04750/2019). This study was also funded by the Euro­ 2020). The most vulnerable and socioeconomically deprived pop­ pean Regional Development Fund (FEDER), through the Competitive­ ulations were found to be at higher risk of psychological distress asso­ ness and Internationalization Operational Programme, by national ciated with COVID-19 pandemic (Pierce, 2020) and are also more likely funding from the Foundation for Science and Technology (FCT) under to live in small apartments with limited access to natural spaces, which the scope of the project PTDC/GES-AMB/30193/2017 (POCI-01-0145- makes them an important vulnerable group and has clear implications FEDER-030193, 02/SAICT/2017 - 30193), the European Research for environmental and health equity. Here, planners play a key role in Council (project GREENLULUs; grant agreement ID: 678034); and the encouraging and supporting the development of private and community EU’s Horizon 2020 framework program for research and innovation green spaces within building premises, in working-class neighborhoods (project NATURVATION, grant agreement ID: 730243). MTM and HC in particularly, following the example of what municipalities such as are funded by Juan de la Cierva fellowships [FJCI-2017-33842, Barcelona have done in the past by co-sponsoring or supporting initia­ IJC2018-035322-I] awarded by the Spanish Ministry of Economy and tives on balconies, terraces, and rooftops. Competitiveness. Ana Isabel Ribeiro was supported by National Funds Our study uses a cross-sectional and observation design based on a through FCT, under the programme of ‘Stimulus of Scientific Employ­ sample of residents, which do not allow proving the effectiveness of ment – Individual Support’ within the contract CEECIND/02386/2018. specific greening and nature-based interventions on mental health out­ The authors would also like to acknowledge all the institutions, media comes. Thus, future research should conduct local studies that take into and individuals that shared the questionnaire in their social and per­ account the built environment where people live to identify specific sonal networks. 11 A.I. Ribeiro et al. Environment International 154 (2021) 106664 Appendix A. Supplementary material Google, 2020. Google COVID-19 Community Mobility Reports. https://www.google.co m/covid19/mobility/ Accessed: <27/5/2020>. Available from: https://www.goo gle.com/covid19/mobility/. Supplementary data to this article can be found online at https://doi. Hoffimann, E., Barros, H., Ribeiro, A.I., 2017. Socioeconomic inequalities in green space org/10.1016/j.envint.2021.106664. quality and accessibility-evidence from a Southern European City. Int. J. Environ. Res. Publ. Health 14 (8), 916. https://doi.org/10.3390/ijerph14080916. Pierce, M., et al., 2020. Mental health before and during the COVID-19 pandemic: a References longitudinal probability sample survey of the UK population. Lancet Psychiatry 7 (10), 883–892. Lane Kevin, J., et al., 2017. Associations between greenness, impervious surface area, Cohen, J., 2013. Statistical Power Analysis for the Behavioral Sciences. Academic Press. and nighttime lights on biomarkers of vascular aging in Chennai, India. Environ. Goldberg, D., Williams, P., 2000. General Health Questionnaire (GHQ). nferNelson, Health Perspect. 125 (8), 087003. Swindon, Wiltshire, UK. Cavaleiro Rufo, J., et al., 2020. The neighbourhood natural environment is associated Laranjeira, C.A., 2008. General health questionnaire-12 items: adaptation study to the with asthma in children: a birth cohort study. Allergy. Portuguese population. Epidemiol. Psychiatric Sci. 17 (2), 148–151. Cavaleiro Rufo, J., et al., 2020. The influence of species richness in primary school del Pilar Sánchez-López, M., Dresch, V., 2008. The 12-Item General Health Questionnaire surroundings on children lung function and allergic disease development. Pediatric (GHQ-12): reliability, external validity and factor structure in the Spanish Allergy Immunol. 31 (4), 358–363. population. Psicothema 20 (4), 839–843. de Keijzer, C., et al., 2018. Residential Surrounding Greenness and Cognitive Decline: A Terluin, B., et al., 2006. The Four-Dimensional Symptom Questionnaire (4DSQ): a 10-Year Follow-up of the Whitehall II Cohort. Environ. Health Perspect. 126 (7), validation study of a multidimensional self-report questionnaire to assess distress, 077003. depression, anxiety and somatization. BMC Psychiatry 6, 34. Twohig-Bennett, C., Jones, A., 2018. The health benefits of the great outdoors: a WHO, 2020. Process of translation and adaptation of instruments. Available from: systematic review and meta-analysis of greenspace exposure and health outcomes. https://www.who.int/substance_abuse/research_tools/translation/en/. Environ. Res. 166, 628–637. Krabbe, P., 2016. The Measurement of Health and Health Status: Concepts, Methods and Gascon, M., et al., 2016. Residential green spaces and mortality: a systematic review. Applications from a Multidisciplinary Perspective. Academic Press. Environ. Int. 86, 60–67. Lee, M.-S., et al., 2015. Interaction with indoor plants may reduce psychological and Wendelboe-Nelson, C., et al., 2019. A scoping review mapping research on green space physiological stress by suppressing autonomic nervous system activity in young and associated mental health benefits. Int. J. Environ. Res. Public Health 16 (12), adults: a randomized crossover study. J. Physiol. Anthropol. 34 (1), 21. 2081. Park, S.-H., Mattson, R.H., 2009. Ornamental indoor plants in hospital rooms enhanced Triguero-Mas, M., et al., 2017. Natural outdoor environments and mental health: Stress health outcomes of patients recovering from surgery. J. Alternat. Compl. Med. 15 as a possible mechanism. Environ. Res. 159, 629–638. (9), 975–980. Bratman, G.N., et al., 2015. Nature experience reduces rumination and subgenual Shibata, S., Suzuki, N., 2001. Effects of indoor foliage plants on subjects’ recovery from prefrontal cortex activation. Proc. Natl. Acad. Sci. 112 (28), 8567. mental fatigue. North Am. J. Psychol. 3 (3), 385–396. Ribeiro, A.I., et al., 2019. Association between neighbourhood green space and biological Yotti Kingsley, J., Townsend, M., 2006. ‘Dig In’ to social capital: community gardens as markers in school-aged children. Findings from the Generation XXI birth cohort. mechanisms for growing urban social connectedness. Urban Policy Res. 24 (4), Environ. Int. 132, 105070. 525–537. Pearson, D.G., Craig, T., 2014. The great outdoors? Exploring the mental health benefits Triguero-Mas, M., et al., 2015. Natural outdoor environments and mental and physical of natural environments. Front. Psychol. 5, 1178. health: Relationships and mechanisms. Environ. Int. 77, 35–41. Markevych, I., et al., 2017. Exploring pathways linking greenspace to health: theoretical van den Berg, M., et al., 2016. Visiting green space is associated with mental health and and methodological guidance. Environ. Res. 158, 301–317. vitality: a cross-sectional study in four european cities. Health & Place 38, 8–15. Wu, G., et al., 2013. Understanding resilience. Front. Behav. Neurosci. 7, 10. Gascon, M., et al., 2015. Mental health benefits of long-term exposure to residential van den Berg, A.E., et al., 2010. Green space as a buffer between stressful life events and green and blue spaces: a systematic review. Int. J. Environ. Res. Public Health 12 (4). health. Soc. Sci. Med. 70 (8), 1203–1210. Ulrich, R.S., 1983. Aesthetic and affective response to natural environment. In: Behavior Marselle, R.M., Warber, L.S., Irvine, N.K., 2019. Growing resilience through interaction and the Natural Environment. Springer, pp. 85–125. with nature: can group walks in nature buffer the effects of stressful life events on Kaplan, S., 1995. The restorative benefits of nature: toward an integrative framework. mental health? Int. J. Environ. Res. Public Health 16 (6). J. Environ. Psycholo. 15 (3), 169–182. Berto, R., 2014. The role of nature in coping with psycho-physiological stress: a literature Mitchell, R., 2013. Is physical activity in natural environments better for mental health review on restorativeness. Behav Sci (Basel) 4 (4), 394–409. than physical activity in other environments? Soc. Sci. Med. 91, 130–134. Spiteri, G., et al., 2020. First cases of coronavirus disease 2019 (COVID-19) in the WHO Nutsford, D., et al., 2016. Residential exposure to visible blue space (but not green space) European Region, 24 January to 21 February 2020. Eurosurveillance 25 (9), associated with lower psychological distress in a capital city. Health Place 39, 70–78. 2000178. Kaplan, R., 2001. The nature of the view from home: psychological benefits. Environ. DGS, 2020. COVID-19: Relatório de Situação. https://covid19.min-saude.pt/relatorio- Behav. 33 (4), 507–542. de-situacao/. Ulrich, R.S., 1979. Visual landscapes and psychological well-being. Landscape Res. 4 (1), Ministerio_de_Sanidad. Resumen de la situación., 2020. Available from: https://cnecovid 17–23. .isciii.es/covid19/. INE, 2020. Proportion of households with at least one person aged between 16 and 74 PortugueseGov. Decreto do Governo que regulamenta o estado de emergência. 2020 23/ years old and with Internet connection at home (%) by Household type. Available 6/2020]; Available from: https://www.portugal.gov.pt/pt/gc22/comunicacao/d from: https://www.ine.pt/xportal/xmain?xpid=INE&xpgid=ine_indicadores&ind ocumento?i=decreto-do-governo-que-regulamenta-o-estado-de-emergencia-. OcorrCod=0007958&contexto=bd&selTab=tab2. SpanishGov. Real Decreto 463/2020, de 14 de marzo, por el que se declara el estado de INE, 2020. Encuesta sobre Equipamiento y Uso de Tecnologías de Información y alarma para la gestión de la situación de crisis sanitaria ocasionada por el COVID-19. Comunicación en los Hogares. Available from: https://www.ine.es/prensa/tich 2020 23/6/2020]; Available from: https://www.boe.es/diario_boe/txt.php? _2020.pdf. id=BOE-A-2020-3692. INE, 2019. Inquérito à Utilização de Tecnologias da Informação e da Comunicação pelas Armitage, R., Nellums, L.B., 2020. COVID-19 and the consequences of isolating the Famílias 2019. Available from: https://www.ine.pt/xportal/xmain?xpid=INE&xp elderly. Lancet Publ. Health. gid=ine_destaques&DESTAQUESdest_boui=354447559&DESTAQUESmodo=2&xla Brooks, S.K., et al., 2020. The psychological impact of quarantine and how to reduce it: ng=pt. rapid review of the evidence. The Lancet 395 (10227), 912–920. Rossi, R., et al., , 2020. COVID-19 pandemic and lockdown measures impact on mental Pfefferbaum, B., North, C.S., 2020. Mental Health and the Covid-19 Pandemic. N. Engl. J. health among the general population in Italy. An N=18147 web-based survey. Med. medRxiv, pp. 2020.04.09.20057802. Rajkumar, R.P., 2020. COVID-19 and mental health: a review of the existing literature. Kokko, K., Rantanen, J., Pulkkinen, L., 2015. Associations between mental well-being Asian J. Psychiatr. 52, 102066. and personality from a life span perspective. In: Blatný, M. (Ed.), Personality and Torales, J., et al., 2020. The outbreak of COVID-19 coronavirus and its impact on global Well-being Across the Life-Span. Palgrave Macmillan UK, London, pp. 134–159. mental health. Int. J. Soci. Psychiatry, 0020764020915212. WHO, 2019. The WHO special initiative for mental health (2019–2023): universal health Tendais, I., Ribeiro, A.I., 2020. Espaços verdes urbanos e saúde mental durante o coverage for mental health. confinamento causado pela Covid-19. Finisterra 55 (115(AOP)). Kozloff, N., et al., 2020. The COVID-19 global pandemic: implications for people with schizophrenia and related disorders. Schizophr. Bull. 46 (4), 752–757. 12