Introduction

Mental health during the COVID-19 pandemic

Mental disorders represent one of the major causes of the global burden of disease (GBD 2022 Mental Disorders Collaborators 2022), and constitute the fourth largest cause of the burden of disease in high-income countries (Roser et al. 2021). The COVID-19 pandemic and associated containment measures to limit disease transmission brought many changes and challenges to individuals’ work and private-life situations (Baral et al. 2021; Kavaliunas et al. 2020; Ludvigsson 2020), with predictions of negative impacts on mental health (Pfefferbaum and North 2020). Compared to lockdown strategies to limit contagion in other countries, Swedish authorities responded with a softer and mainly recommendation-based strategy, since authorities were limited by law to enforce strict stay-at-home measures (Ludvigsson 2020). Nevertheless, compliance with public recommendations in Sweden was high, and recommendation-based self-isolation seemed to be related to self-experienced changes in mental health across different population groups, such as the elderly (Gustavsson and Beckman 2020) and students (Berman et al. 2022). However, existing findings suggest elevated mental health problems including depression, anxiety, and insomnia in Sweden (Berman et al. 2022; Gustavsson and Beckman 2020; Lovik et al. 2023; McCracken et al. 2020; Rondung et al. 2021) and elsewhere (Brailovskaia et al. 2023; Brailovskaia and Margraf 2023; Pfefferbaum and North 2020; Vindegaard and Benros 2020; Witteveen et al. 2023). Yet, reports of mental health problems in Swedish studies vary, probably due to differences in samples (e.g., population-based sample vs convenience sample) and the time of the study period (e.g., 1 year vs 1 month) (Lovik et al. 2023; McCracken et al. 2020). Furthermore, the COVID-19 pandemic seems to have had variable impacts on mental health outcomes in different groups of the population in Sweden, with younger adults, women, and economically vulnerable individuals being more negatively affected (Rondung et al. 2021), in line with international findings (Witteveen et al. 2023). Moreover, a systematic review found elevated levels of loneliness during the pandemic. However, longitudinal findings suggest that an initial increase in loneliness was followed by a decrease toward baseline levels (i.e., a “recovery process”) during the first year of COVID-19 (Buecker and Horstmann 2021).

Nature exposure and mental health

Exposure to natural environments such as green (vegetation) and blue (open-water bodies) spaces has repeatedly been found to benefit mental health, including sleep (Bratman et al. 2015, 2019; Kaplan 2001; Kaplan and Kaplan 1989; Keniger et al. 2013; Klein et al. 2022; Kruize et al. 2020; McMahan and Estes 2015; Shin et al. 2020; Stenfors et al. 2023; Ulrich 1984; White et al. 2017; Wilson 1984) and loneliness (Astell-Burt et al. 2022a, b). Moreover, studies have suggested a potential buffering effect of nature exposure on stressful life events (van den Berg et al. 2010; Ulrich 1984). This can be of particular importance in the context of the COVID-19 pandemic.

While social restrictions limited many activities in Sweden (especially indoors) during the pandemic, natural environments remained accessible and became an important environment for recreational and social activities. For example, since the start of the COVID-19 pandemic, the number of people spending time in natural environments increased substantially across different countries (Derks et al. 2020; Venter et al. 2020), including Sweden (Hansen et al. 2022; Lõhmus et al. 2021; Samuelsson et al. 2021; SCB — Statistics Sweden 2021).

Several studies also reported positive associations between contact with nature and different mental health outcomes (MHO) during the COVID-19 pandemic (Labib et al. 2022; Lõhmus et al. 2021; Patwary et al. 2022). In an international survey early on, people in different countries also reported that contact with nature helped them cope with the negative impact of the pandemic situation on their mental well-being (Pouso et al. 2021). Previous cross-sectional research in Stockholm found higher greenness around the home to be linked to better MHO and higher vitality during the beginning of the pandemic, and that nature visits increased primarily to preserves and private gardens (Lõhmus et al. 2021). Furthermore, research in Helsinki (Korpilo et al. 2021) and Stockholm (Samuelsson et al. 2021) found that the frequency of forest visits was the best predictor for the influence of places on well-being, in line with earlier findings that forests are preferred for restorative experiences (Giusti and Samuelsson 2020). However, population-based and longitudinal studies following individuals across the pandemic, while taking potential confounders into account, are still limited and needed (Patwary et al. 2022).

The aim of the present study was thus to investigate the development of nature-related habits (NRH) and their relationships with MHO during the COVID-19 pandemic, across the period from autumn 2019 to early 2022, while also controlling for potential confounders. Specifically, the present study investigated the relationships between different types of NRH in 2019, 2020, and 2022, and MHO in early 2021 and 2022, hypothesizing that more frequent NRH would be associated with better MHO. Furthermore, engaging in NRH was hypothesized to buffer the detrimental effects of major life events (MLE) on MHO during the COVID-19 pandemic.

Methods

Study sample

The study sample comprises a sub-sample of respondents to the Swedish Longitudinal Occupational Survey of Health (SLOSH) in 2020. Participation requirements were having participated in SLOSH 2020 and agreeing to participate in an additional online survey, conducted in 2021 (n = 3041, SLOSH–Corona). Out of these, 1902 (62.55%) responded to the online SLOSH–Corona survey in 2021 and were included in analyses concerning the early phase of the pandemic until 2021. The sample was followed up again in 2022, with a longitudinal attrition of 317 individuals, which resulted in a sample of 1580 individuals who responded to the SLOSH–Corona surveys in both 2021 and 2022, and who were thus included in analyses of the later phase of the pandemic in 2022.

SLOSH includes respondents to the Swedish Work Environment Surveys (SWES 2003—2011, at the time of SLOSH 2020), who are sampled from the Labor Force Surveys (LFS) conducted every second year by Statistics Sweden (Magnusson Hanson et al. 2018). In the LFS, a random sample of approximately 20,000 people aged 15–74 is drawn biennially from the Swedish population (stratified by county, sex, citizenship, and inferred employment status). The response rate in SLOSH, and SLOSH-Corona, is slightly higher for women, older respondents, and those with higher education (for further details, see Magnusson Hanson et al. 2018). In SLOSH 2020, 17,489 responded to the regular SLOSH survey (49% response rate).

Variables

Nature-related habits

Nature-related habits (NRH) were measured for three time points, through seven items. To measure NRH before the pandemic, respondents were asked to rate “On average, how often did you stay in/visit the following types of natural areas during late autumn 2019?” with items including “Own garden”, “Park/ town park”, “Green playgrounds”, “Forest”, “Beach/ sea/ lake/ watercourse” (blue spaces), and “Other natural areas”. Additionally, respondents were asked, “How often did you pass parks, forests, or water bodies by walking or cycling to work or another activity in late autumn 2019?” (mobility-related exposure). These items were used as pre-pandemic NRH scores. In the 2021 online survey, respondents were also asked the same set of NRH items but for the period in the past months, asking about NRH in autumn 2020, i.e., in the months preceding the survey in early 2021. The third NRH measurement took place in early 2022, using the same seven items, asking respondents to rate their current habits: “On average, how often do you spend time in/ visit the following types of natural areas?” and “How often do you pass parks, forests, or water bodies by walking or cycling to work or another activity?”. The seven items were rated on a four-point response scale (1 = “Never”, 2 = “Sometimes per month”, 3 = “Sometimes per week”, and 4 = “Every y day”).

The different types of NRH were investigated separately, in addition to a general NRH score based on the multiple items concerning NRH visits outside the home environment. Reliability and inter-item correlation analyses showed that the “Own garden” item differed from the other NRH items in terms of internal consistency, possibly because the other NRH items differed in that they involved leaving the home environment. For all NRH measures, the internal consistency coefficients (2019: α = 0.65, ω = 0.68; 2020: α = 0.71, ω = 0.73; 2022: α = 0.69, ω = 0.72) improved when dropping the item “Own garden” (2019: α = 0.68, ω = 0.69; 2020: α = 0.74, ω = 0.74; 2022: α = 0.72, ω = 0.72). Other studies also investigated spending time in the garden separately from visiting natural environments outside of the home environment (Labib et al. 2022; Patwary et al. 2022). Consequently, the item “Own garden” was analyzed separately and excluded when calculating a general NRH score. The general NRH score was operationalized as the highest frequency score on any of the six NRH items.

Mental health outcomes

Anxiety symptoms

Anxiety symptoms (at the beginning of 2021 and 2022) were measured through the Generalized Anxiety Disorder 7-item scale (GAD-7) (Spitzer et al. 2006). Participants responded to the question “Over the last 2 weeks, how often have you been bothered by the following problems…” (e.g., “Worrying too much about different things”), on a scale from 0 (“not at all”) to 3 (“almost every day”). The items were combined into a GAD-7 score, which was used in the analyses (2021: α = 0.90, ω = 0.91; 2022: α = 0.87, ω = 0.87). In models analyzing NRH changes and MHO, a measure of anxiety symptoms from the SLOSH survey in 2018 (three items from the Symptom Checklist; Søgaard and Bech 2009) was used as a control variable to adjust for pre-pandemic anxiety levels (“How much during the last week have you been troubled by: 1: ‘Nervousness or anxiety’, 2: ‘Suddenly feeling scared for no reason’, and 3: ‘Anxiety or panic attacks’”, with responses ranging from 1 “Not at all” to 5 “Very much” (α = 0.71, ω = 0.77).

Depressive symptoms

Depressive symptoms (at the beginning of 2021 and 2022) were measured through the nine-item Patient Health Questionnaire (PHQ-9) (Kroenke et al. 2001). Respondents were asked “Over the last two weeks, how often have you been bothered by the following problems?” with responses ranging from 0 (“not at all”) to 3 (“almost every day”). All nine items (e.g., “Weak interest or pleasure in doing things”) were summed into a PHQ score (2021: α = 0.88, ω = 0.89; 2022: α = 0.86, ω = 0.87). In models analyzing NRH changes and MHO, a measure of depressive symptoms from the SLOSH survey in 2018 (the Symptom Checklist core depression scale; Magnusson Hanson et al. 2014) was used as a control variable to adjust for pre-pandemic depression levels. These items included “How much during the last week have you been troubled by”: 1: ‘Lethargy or low in energy’, 2: ‘Feeling blue’, and 3: ‘Blaming yourself’, 4:‘Worrying too much’, 5: ‘feeling no interest in things’, 6: ‘Everything is an effort’”, with a response range from 1 “Not at all” to 5 “Very much” (α = 0.91, ω = 0.91).

Loneliness

Loneliness levels (before the pandemic and at the beginning of 2021 and 2022) were measured through a 4-item scale, based on a modified short version of the UCLA Loneliness Scale (Hughes et al. 2004). In the 2021 survey, participants responded to the question “How often before the Corona pandemic did you feel…”: “that you missed companionship?”, “lonely?”, “left out?”, “isolated from others?” (α = 0.90, ω = 0.90), from 0 “Never” to 3 “Often”. Participants also answered the same four items to the question “How often since the Corona pandemic began did you feel…” (α = 0.88, ω = 0.89). In the 2022 survey, participants responded to the same items and the same question in the present tense: “How often do you feel…” (α = 0.89, ω = 0.89).

Self-rated mental health changes during the pandemic

Further, participants rated how their health had changed during the pandemic concerning “Difficulties falling asleep”, “Disturbed/ restless sleep”, “Not feeling refreshed at wake-up”, “Depressed mood”, and “Nervousness or anxiety”, according to the following question: “If you compare your health/well-being since the start of the COVID-19 pandemic with what it was like before, how has it changed in terms of the following symptoms?” from 1 “Much reduced” to 3 “Unchanged” to 5 “Increased a lot”. The three sleep items were adapted from the Karolinska Sleep Questionnaire, combined into one scale (2021: α = 0.89, ω = 0.89, 2022: α = 0.86, ω = 0.87) (Nordin et al. 2013), and used in models including NRH and MHO.

Major life events

Major life events were assessed with the question “Have you experienced any of the following events since the beginning of the Corona pandemic?”, with the Yes/No items “More serious personal illness or injury”, “Witnessed severe human suffering”, “Serious illness/accident of a close relative”, “Death of a close relative”, “Other very stressful event or experience”. Subsequently, participants were asked to mark how significant this event was for them when it happened, with four response options, i.e., 1 “Did not matter much”, 2 “Affected me”, 3 “Affected me strongly”, and 4 “Was completely devastating”. Based on these items, a major life event dummy-coded variable was computed, with 1 indicating any major life event that was reported to affect the individual (≥ 2) and 0 indicating no occurrence of a major life event that affected the individual.

Control variables

Control variables included individual-level demographic variables (e.g., age, sex, and individual and household annual net income), received through national administrative registers. Relationship status and socioeconomic position (SEP; blue-collar vs white-collar worker) were obtained from the 2020 SLOSH survey and work status from the SLOSH–Corona survey. Respondents indicated whether or not they were single or married/cohabiting with a partner, and whether or not they were gainfully working at least 30% of full time.

Statistical analysis

Descriptive statistics and paired samples t-tests were used to assess changes in NRH and MHO across time.

Sequential multiple linear regressions were conducted to analyze associations between nature-related habits (NRH) during the pandemic in autumn 2020 and mental health outcomes (MHO) in early 2021 (unadjusted model) while controlling for age, sex, individual income, work status, SEP, household income, and marital/ relationship status (fully adjusted model).

Further, changes in NRH from 2019 to 2020 and 2019 to 2022 were analyzed in relation to MHO 2021 and 2022 respectively, adjusting for baseline pre-pandemic MHO scores. For the sleep outcome, only ratings of change were available (in 2021 and 2022) and analyzed in the models. Difference scores between NRH in 2020/2022 and 2019 were calculated and used as a predictor variable, with MHO 2021/2022 as outcome variables in sequential linear regression models, while adjusting for control variables, as well as the respective pre-pandemic MHO scores. Sensitivity analyses were performed to investigate whether positive or negative NRH difference scores were driving the effects on MHO. In these regression models, the respective groups with positive vs negative difference scores were dummy-coded, while unchanged NRH (0 difference values) served as the reference group.

Finally, interactions between NRH (in 2020 and 2022) and major life events on MHO (in 2021 and 2022) were analyzed by including interaction terms of NRH and major life events in the regression models.

Results are reported with 95% confidence intervals (CI). Data were analyzed using SPSS Version 28.0.1.0.

Results

Table 1 presents descriptive statistics on sample characteristics, while Tables 2 and 3 show nature-related habits (NRH), and Table 4 shows mental health outcomes (MHO).

Table 1 Descriptives for background variables for respondents in 2021 and 2022
Full size table
Table 2 Frequencies in % and mean score of nature-related habits in different environments
Full size table
Table 3 Proportion of individuals (%) who either decreased (−), increased (+), or did not change (0) the extent of NRH between 2019, 2020, and 2022. The paired samples t−test statistic (t) indicates differences between NRH scores for the respective years
Full size table
Table 4 MHO mean item and sum scores. Paired samples t−test statistics (t) indicate differences between mean MHO scores for the respective years
Full size table

Nature-related habits

The analyses of changes in the frequency of engaging in NRH before (2019) and during (2020 and 2022) showed that when compared to other NRH, the frequency of forest visits increased the most from 2019 to 2020, followed by blue-space visits and visiting green playgrounds (Table 3). The frequency of engaging in general NRH seemed to decrease slightly from before (2019) to during the first year of the pandemic (2020), as indicated by, for example, general habit response score differences between 2019 and 2020, with similar results for the frequency of using own gardens.

From 2019 to 2022, all NRH frequencies, except for spending time in the own garden, increased (Tables 2 and 3; Fig. S7). Between these years, the largest NRH frequency increase was observed in mobility-related nature exposure, i.e., passing parks, forests, or water bodies by walking or cycling to work or other activities (Tables 2 and 3). The initial increase in forest visits remained at a higher level also in 2022. Increases in general NRH, mobility-related exposure, park visits, blue-space visits, or visits to other natural areas mainly occurred in the later period of the pandemic between 2020 and 2022 (Tables 2 and 3).

In 2022, participants still visited the forest more often than before the pandemic in 2019 (Tables 2 and 3). Moreover, the results indicate that the increase in forest visits primarily occurred in the early time of the pandemic, with slightly more participants decreasing (20.3%) than increasing (18.5%) the number of forest visits between the early (2020) and later (2022) time of the pandemic (Table 3), although this slight decrease in forest visits was not statistically significant.

Mental health outcomes

Examining the MHO survey responses, the mean scores of GAD-7 and PHQ-9 items show that anxiety levels on average slightly decreased from early 2021 to early 2022, while depressive symptom levels were approximately similar at the two time points during the pandemic (Table 4). Levels of loneliness, however, decreased from early 2021 to early 2022. Considering the pre-Covid loneliness score, measured in 2021, participants seemed to experience a peak of perceived loneliness during the pandemic in 2021, before loneliness levels on average decreased again in 2022 (Table 4).

This is also shown in the average frequency of reported symptom levels across loneliness items (Table S5). In 2021, 39.8% indicated loneliness levels above “Almost never” (> 1), while 13.5% of the sample reported loneliness on average, more than “Some of the days” (> 2). In 2022 the corresponding proportions were: 23.9% (> 1) and 3.6% (> 2). In contrast, before the pandemic proportions were as follows: 16.5% (> 1) and 2.3% (> 2).

Reports with regard to health-related changes from before to during the pandemic in 2021 show that 30.3% and 25.1% reported that depressive mood and nervousness or anxiety levels respectively increased slightly to a lot. In 2022, these scores decreased to 25.5% and 21.3% respectively (Table S3). Based on paired samples t-tests, the mean decrease in reported changes in depressive mood and nervousness or anxiety from 2021 to 2022 was statistically significant (Table 4). Further, participants reported that sleeping problems increased slightly to a lot with regard to falling asleep (16.9% in 2021 and 15.6% in 2022), disturbed/restless sleep (22.5% in 2021 and 22.6% in 2022), and not feeling refreshed at wake-up (21.1% in 2021 and 20.4% in 2022) since the start of the pandemic (Table S3). The average change in general sleep problems compared to before the pandemic was perceived similarly in 2021 and 2022 (Table 4).

Nature-related habits and mental health outcomes

Greater engagement in NRH was consistently associated with MHO. The observed associations between NRH in autumn 2020 and MHO in early 2021 were similar in direction and on average slightly stronger in early 2022 (Table 5). Table 5 displays the associations between NRH in autumn 2020 and MHO in early 2021, and between NRH in early 2022 and MHO in early 2022, in unstandardized and standardized regression coefficients. Table 6 displays the associations between short-term changes in NRH (autumn 2019–autumn 2020) and MHO in early 2021, and between long-term changes in NRH (2019–early 2022) and MHO in early 2022, in unstandardized and standardized regression coefficients.

Table 5 Associations between NRH and MHO in the early phase (NRH autumn 2020, and MHO early 2021, in part a) versus the later phase of the pandemic (early 2022, in part b). Results are presented in standardized (β) and unstandardized (B) coefficients, with 95% confidence intervals (CI), for unadjusted (1) and fully adjusted (2) models
Full size table
Table 6 Part a: associations between NRH changes in the early phase of the pandemic (autumn 2020 to early 2021) and MHO (in early 2021, adjusted for pre-pandemic MHO scores). Part b: associations between NRH changes in the later phase of the pandemic (autumn 2020 to early 2022) and MHO (in early 2022, adjusted for pre-pandemic MHO scores)
Full size table

NRH and depressive symptoms

Cross-sectional (Table 5)

Engaging in general NRH in autumn 2020 was associated with lower levels of depressive symptoms in 2021 in the fully adjusted model. Forest visits, in particular, were associated with lower levels of depressive symptoms in early 2021. Mobility-related exposure, spending time in one's own garden, park visits, and visiting blue spaces, or other natural environments were also related to lower levels of depressive symptoms in early 2021.

The results were similar and on average slightly stronger for NRH in early 2022 and depressive symptoms in early 2022 (Fig. S9).

Longitudinal associations between early/short-term NRH changes (2019 to 2020) and (baseline-adjusted) depressive symptoms in 2021 (Table 6)

NRH difference scores ranged from −3 (maximum decrease) to 3 (maximum increase), with 0 indicating no change. Changes in terms of higher difference scores of general NRH (i.e., more habit increase, less habit decrease) were related to lower levels of depressive symptoms.

The sensitivity analysis in the fully adjusted model (Table S6) showed that negative NRH difference scores (i.e., decreasing the habit from 2019 to 2020) were related to elevated levels of depressive symptoms rather than positive NRH difference scores (i.e., increasing the habit) to lower levels of depressive symptoms when respectively compared to the no change groups. Changes (i.e., increasing difference scores) in the respective type of nature visits—forest, green playground, and other natural area visits—from 2019 to 2020, were also related to lower levels of depressive symptoms.

The sensitivity analyses indicated that both increases in forest visits from 2019 to 2020 were related to lower depressive symptom levels in early 2021 and decreases in visits were related to higher symptom levels when compared to the no-change group. Furthermore, visiting other natural areas less in 2020 than in 2019 was related to higher levels of depressive symptoms, while more frequent green-playground visits were associated with lower depressive symptom levels, compared to the no-change groups (Table S6).

Longitudinal associations between longer-term NRH changes (2019 to 2022) and (baseline-adjusted) depressive symptoms in 2022 (Table 6)

Changes (increasing difference scores) in NRH from 2019 to 2022 that were associated with lower levels of depressive symptoms in early 2022 included mobility-related exposure and visits to parks, green playgrounds, forests, and other natural areas, but not garden and blue-space visits (Table 6).

The sensitivity analyses suggested that especially decreasing the frequency of general NRH from 2019 to 2022 was related to higher levels of depressive symptoms in early 2022. For visits to other natural areas, an increase in visits was associated with less depressive symptoms, and a decrease in visits was associated with more depressive symptoms, when compared to the no-change groups (Table S6).

NRH and anxiety symptoms

Cross-sectional (Table 5)

As for depressive symptoms, engaging in general NRH in autumn 2020 was related to lower levels of anxiety symptoms in 2021. Further, mobility in natural environments, spending more time in one’s own garden, in the park, visiting forests, blue spaces, and other natural environments were associated with lower levels of anxiety symptoms.

On average, the results were similar for NRH in early 2022 and anxiety symptoms in early 2022 (Fig. S10).

Longitudinal associations between early/short-term NRH changes (2019 to 2020) and (baseline-adjusted) anxiety symptoms in 2021 (Table 6)

Increasing the frequency of park, green playground, forest, and other natural area visits from 2019 to 2020, or decreasing them less, was associated with lower levels of anxiety symptoms in 2021 in the fully adjusted regression models, adjusting for pre-pandemic anxiety levels (Table 6).

The sensitivity analyses showed that especially decreasing park, forest, and other natural area visits were related to higher anxiety levels when compared to the respective no-change groups (Table S6). The model including green-playground visits suggested that increasing and decreasing green-playground visits were respectively related to lowered and elevated anxiety symptom levels when compared to the no-change group (Table S6).

Longitudinal associations between longer-term NRH changes (2019 to 2022) and (baseline-adjusted) anxiety symptoms in 2022 (Table 6)

In the long run (2019 to 2022), an increase in difference scores in general NRH, mobility-related nature exposure, green-playground visits, and visiting other natural areas was related to lower anxiety symptom levels (Table 6).

The sensitivity analyses suggested that increasing and decreasing green-playground and other natural area visits were related to lower and higher anxiety symptom levels to a similar extent respectively, compared to the respective no-change groups. Concerning changes in general NRH and mobility in natural environments, decreasing the habit from 2019 to 2020, in particular, was related to higher anxiety symptom levels (Table S6).

NRH and loneliness

As for depressive and anxiety symptoms, a higher general NRH engagement was associated with lower loneliness levels. Significant associations in the same direction were observed for mobility-related nature exposure, garden time, visiting forest areas, and blue spaces. The cross-sectional results were comparable in early 2022 and depressive symptoms in early 2022 (Table 5; Fig. S11).

Early/ short-term changes (i.e., increasing difference scores) in general NRH from 2019 to 2020 were related to lower levels of loneliness in 2021. Significant associations in the same direction were observed for mobility-related nature exposure, park visits, and green-playground visits (Table 6). The sensitivity analysis in the fully adjusted models suggested that primarily negative NRH difference scores (i.e., decreasing the habit from 2019 to 2020) were associated with elevated levels of loneliness, rather than positive NRH difference scores with lower levels of loneliness. For green-playground visits only, both increasing the habit was related to lower while decreasing it to higher loneliness in similar ways (Table S6).

In the longer term (2019–2022), higher general NRH, and mobility difference scores were related to lower loneliness levels in 2022 (Table 6). For both, decreases over time were related to higher loneliness levels in 2022 (Table S6).

NRH and sleep problems

More frequent general NRH in 2020 were associated with fewer sleep problems in 2021. Visiting forests and spending time in one’s own garden were related to fewer sleep problems in 2021 as well (Table 5). The cross-sectional results from 2021 were comparable to those in 2022 and additionally, more frequent mobility in natural environments in 2022 was associated with fewer sleep problems.

Increasing the frequency of general NRH, mobility in natural environments, park visits, forest visits, and visiting blue spaces from 2019 to 2020, or decreasing them less, was associated with fewer sleep problems in 2021 as compared to before the pandemic (Table 6). The sensitivity analyses showed that particularly increasing forest visits from 2019 to 2020 was related to better sleep. For the other NRH (general NRH, mobility, park, blue spaces), a decrease in the habit was related to more sleep problems, rather than the other way around (Table S6).

Over time, only an increase in difference scores of park visits from 2019 to 2022 was associated with decreased sleep problems in 2022 as compared to before the pandemic (Table 6). Both increasing and decreasing park visits were similarly associated with decreased vs increased sleep problems (Table S6).

NRH, major life events, and MHO

The occurrence of a major life event (MLE) was highly related to elevated symptom levels of MHO in 2021 and 2022 (Table S2 in the supplementary material). The analyses of main and interaction effects for the time point in early 2021, in fully adjusted regression models, showed that having and using a garden may buffer the detrimental effects of MLE on depressive symptoms (Model Adj. R2 = 0.07; main effects: βGarden = −0.05, B = −0.18 [−0.38; 0.03], p = 0.087, βMLE = 0.34, B = 2.85 [1.83; 3.86], p < 0.001; Interaction effect: βGarden*MLE = −0.18, B = −0.47 [−0.80; −0.15], p = 0.004) and symptoms of anxiety (Model Adj. R2 = 0.07; Main effects: βGarden = −0.05, B = −0.15 [−0.32; 0.03], p = 0.096, βMLE = 0.33, B = 2.37 [1.51; 3.24], p < 0.001; Interaction effect: βGarden*MLE = −0.20, B = −0.44 [−072; −0.16], p = 0.002; Fig. S8). No such interaction effects were observed for loneliness and self-rated changes in sleep problems in 2021, or NRH outside the home environment. The observed interaction effects between garden use and major life events on symptoms of depression and anxiety in early 2021 were not observed 1 year later, in early 2022.

Discussion

Principal results and comparison with prior work

Investigting NRH and MHO in a population-based sample of adults in Sweden, from before to during 2 years of the COVID-19 pandemic, the present study showed that the frequency of engaging in NRH increased during the pandemic and that this was associated with better MHO. Thus, the findings supported the hypothesis that more frequent NRH (cross-sectionally and over time) would be associated with better MHO. With regard to the hypothesis that engaging in NRH might buffer against the detrimental effects of MLE on MHO during COVID-19, spending time in one’s own garden was found to moderate the detrimental effects of MLE on symptom levels of depression and anxiety in early 2021, but not in early 2022. However, no such effects were observed for general NRH outside the home environment. Thus, support for the hypothesis that NRH might buffer against the effects of MLE on MHO was inconsistent in the present study.

NRH — early changes during the pandemic

During the autumn of 2020, when restrictions in Sweden included limiting public gatherings to eight people, forest visits increased in particular, as compared to before the pandemic in the autumn of 2019. Forests generally entail a larger coherent greenspace allowing for both recreational activities without crowding as well as being a restorative environment, often rich in biodiversity which can support restoration, compared to, for example, smaller parks (Stoltz et al. 2016). From a practical perspective, the present findings stress the importance of forests as a critical infrastructure for public health, which may be especially important during pandemic restrictions, in line with previous reports of an “unprecedented boom” in forest visits during the COVID-19 containment measures observed both in Sweden (SCB — Statistics Sweden 2021) and elsewhere during lockdown, e.g., in Germany 2020 (Derks et al. 2020).

However, not all types of NRH increased during the first year of the pandemic (i.e., between autumn 2019 and autumn 2020). In fact, the present findings indicate that while the frequency of generally engaging in NRH and spending time in the garden was still high in 2020 (on average “sometimes per week”), NRH seemed to slightly decrease initially compared to before the pandemic in 2019.

NRH — long-term changes during the pandemic

The present findings show a significant increase in NRH from 2019 (before the pandemic) to the last measurement in 2022, which follows previous reports regarding developments in nature visits during the later, second year of the pandemic (Labib et al. 2022; SCB - Statistics Sweden 2021). These increases, especially further into the pandemic, may represent a gradual adjustment and coping with the pandemic situation, with an increasing awareness and interest in NRH. NRH provided an opportunity to keep engaging in recreational activities outdoors when many indoor activities were restricted, thus acting in accordance with public health recommendations and limiting the risk of COVID-19 transmission.

The present study also provides insight into different kinds of NRH. From 2019 to 2022, there were pronounced increases in NRH frequency, including visits to forests, blue spaces, or other natural areas, followed by visits to green playgrounds and parks. While spending time in one’s garden was still among the most frequent NRH in 2022, its frequency slightly decreased compared to 2019. The largest increase in frequency was observed for mobility-related nature exposure, that is, how often individuals reported passing parks, forests, or water bodies by walking or cycling to work or other activities in 2022, as compared to before the pandemic, indicating a shift in habits towards choosing biking and walking as a means of transportation to a greater extent than before the pandemic.

A study in Oslo similarly observed that both cycling and walking activities in green spaces including forests and city parks increased during the 2 weeks after the announcement of a “partial lockdown” (Venter et al. 2020). Further, our findings align with the observation by Hansen et al. (2022) of individuals wanting to maintain the high level of outdoor habits observed during the pandemic. In our study, individuals may have started or increased cycling or walking as an alternative to public transport in times of physical distancing during the pandemic, and maintained their habits.

As the frequency of visits for each of the environments outside the home increased from 2019 to 2022, it seems important to ascertain that the public has access to different types of natural environments to promote a wide range of healthy behaviors during pandemics or similar crises. This is another finding with important practical implications for urban planning and public health policies.

Mental health outcomes (MHO)

The present findings show that the COVID-19 pandemic involved additional challenges for individual mental health, especially in 2021, in line with other studies in Sweden (Berman et al. 2022; Gustavsson and Beckman 2020; Lovik et al. 2023; McCracken et al. 2020; Rondung et al. 2021) and elsewhere (Brailovskaia et al. 2023; Brailovskaia and Margraf 2023; Pfefferbaum and North 2020; Vindegaard and Benros 2020; Witteveen et al. 2023). In particular, loneliness peaked in 2021 but then approached pre-pandemic levels in 2022. The present longitudinal findings thus align with the notion of a loneliness “recovery process”. Buecker and Horstmann (2021) found that this process was associated with a perceived improvement in the quality and quantity of social interactions over time, meaning that the number of social interactions (including digital interactions) increased. Accordingly, the perceived loss of relationship quality at the beginning of the pandemic returned to baseline levels (Buecker and Horstmann 2021). For other MHO, anxiety similarly decreased in 2022, while depressive symptoms and ratings of changes in sleeping problems were maintained.

Given that the Swedish strategy to limit disease transmission was “softer” in comparison with many other countries, and relied more on compliance with recommendations, it is interesting to observe that findings with regard to mental health challenges during the COVID-19 pandemic in Sweden align with other international observations, e.g., in European and North American countries (Witteveen et al. 2023). This may reflect the high compliance of most groups in the population to public health recommendations issued by the Swedish authorities (Berman et al. 2022; Gustavsson and Beckman 2020). These recommendations included, for example, working from home when possible, staying home if sick, avoiding international travels, and practicing physical distancing, in addition to restrictions on, for example, avoiding public gatherings, and correspond to mandatory restrictions issued in other countries (Girum et al. 2021; Ludvigsson 2020).

NRH and MHO — cross-sectional

Engaging in NRH in the autumn of 2020 was consistently associated with better MHO in 2021. In particular, spending time in one’s own garden and visiting forests in autumn 2020 were related to all MHO in early 2021. This corresponds to other findings where exposure to natural environments during the pandemic was related to lower levels of depression, anxiety, and loneliness (Labib et al. 2022; Lõhmus et al. 2021; Patwary et al. 2022; Pouso et al. 2021; Soga et al. 2021). This is in line with, for example an earlier study in Spain which found that spending time in the garden during the most restrictive times (lockdown) was associated with improved well-being and coping with the restrictions (Pouso et al. 2021). Here, the present study offers additional insights regarding the role of having and spending time in one’s garden, in terms of its association with better mental health on several types of outcomes during a crisis like the pandemic, including loneliness and sleep in addition to depressive and anxiety symptoms. Furthermore, the strong associations between forest visits and all MHO provide novel results and further nuances and knowledge compared to previous indications that forest environments are associated with self-rated general well-being in Stockholm during the pandemic (Samuelsson et al. 2021).

The potential buffering effect of NRH in 2020 on the detrimental effects of MLE during the pandemic on MHO in 2021 was also tested. Spending more time in one’s own garden in 2020 seemed to buffer the detrimental effect of MLE on symptom levels of depression and anxiety in 2021. However, this effect was not replicated in 2022. Thus, the results regarding a potential buffering of NRH on the mental health effects of MLE during COVID-19 were not consistent across both time points. Spending time in one’s garden had a similarly positive effect among both those who experienced MLE some time during the pandemic and those who did not, in 2022. No buffering effects were observed of NRH outside the home on the detrimental effects of MLE on MHO.

NRH and MHO — longitudinal

The present findings expand existing and primarily cross-sectional findings with regard to nature exposure and MHO during COVID-19 (Labib et al. 2022; Patwary et al. 2022). Specifically, the role of changes in NRH between 2019 and 2020 (short-term), and between 2019 and 2022 (long-term) for baseline-adjusted MHO in 2021 and 2022 respectively, were investigated.

Clearly, MHO in 2021 were related to short-term changes in NRH frequency. The observed association between short-term changes in forest visits and depressive symptom levels is particularly interesting. Increasing forest visits from 2019 to 2020 was associated with lowered depressive symptoms in early 2021. On the other hand, decreasing forest visits was associated with elevated depressive symptoms. The long-term pattern in 2022 was similar although weaker. Along with the cross-sectional results, this finding underscores the importance of forest environments for public health during times of crisis such as the COVID-19 pandemic. Moreover, engaging less in general NRH in 2020 compared to 2019 was related to higher symptom levels of MHO in early 2021. Short-term decreases in several different NRH frequencies from 2019 to 2020 were associated with higher MHO symptom levels in early 2021. Abstaining from places with natural elements was also found to be associated with lower self-reported well-being in another cross-sectional study (Samuelsson et al. 2021).

Long-term changes in NRH were also related to MHO in 2022. Depressive symptom levels in particular seemed to be affected by long-term changes in NRH, including general NRH, mobility-related exposure, and visits to green playgrounds, forests, and other natural areas. On the other hand, changes in park or blue-space visits and garden time did not affect depressive symptom levels significantly. Anxiety symptom levels were affected by long-term changes in general NRH, mobility-related exposure, visits to green playgrounds, and other natural areas. Loneliness was associated with general NRH and mobility, while sleep problems were related to park visits only. Thus, in comparison to depression and anxiety, loneliness and sleep problems seemed to be less affected by long-term changes in NRH frequencies. Furthermore, long-term decreases in NRH were also contributing to several of the observed effects. For instance, engaging less in general NRH and visits to other natural areas in 2022, compared to 2019, was related to higher symptom levels of both depression and anxiety. Also, when comparing short-term and long-term changes, the long-term increases in NRH seemed to be more consistently associated with lowered MHO symptoms in terms of lower levels of depressive and anxiety symptoms in 2022. For example, more visits to green playgrounds or other natural areas were related to lower symptom levels of depression and anxiety. The findings of the present study thus highlight that spending time in nature may enhance resilience and promote public health during pandemics. Thus, a practical implication is that easy access to nature, including green and blue environments, should be prioritized to enable NRH and related mental health benefits.

Strengths and limitations

Strengths of the present study include the longitudinal design and the detailed assessments in an adult population-based study sample, also allowing for the adjustment of multiple potential confounders. As in many large-scale surveys, the response rate in SLOSH and the present sample is higher for women, older respondents, and individuals with a higher education (Magnusson Hanson et al. 2018). This was addressed by adjusting for relevant demographic and socioeconomic control variables in the statistical analyses. As for the mental health-related outcomes, established and validated self-report measures were used. Furthermore, loneliness and sleep problems were measured with items from established scales adapted for the COVID-19 pandemic assessment situation. Each respective scale had high internal reliability and fitted a one-factor structure. However, using self-reports of mental health involves both advantages and disadvantages. Advantages of self-reports include the fact that individuals are good at rating their own mental health, which is inherently about subjective experiences (Kananen et al. 2021). Moreover, self-reports permit the investigation of symptom levels including sub-clinical symptoms, and not just register-based diagnoses. A disadvantage involves potential bias (e.g., over- or underreporting of symptoms) or contextual effects. Furthermore, besides the general NRH score based on multiple items, single items were also used to investigate the frequency of visits to various natural environments of different quality (e.g., forests, parks, etc.). Single-item measures have shortcomings (Allen et al. 2022), and measures including multiple items can yield better scale properties (Fisher et al. 2016). However, research has shown that single-item measures are often as valid and reliable as their multi-item alternatives (Ahmad et al. 2014; Ang and Eisend 2018). As such, they can be efficient alternatives, benefitting response rates and supporting respondents’ attention span in surveys (Ang and Eisend 2018). Due to the high face validity of individual NRH items in the present study, their use provided more detailed information regarding the mental health-promoting potential of qualitatively different types of natural environments. Thus, the present study also meets existing calls for studies investigating the relationship between different qualities of natural environments on MHO and health-promoting behaviors (van den Berg et al. 2015; Labib et al. 2020; Leslie et al. 2010; Markevych et al. 2017). Although a strength of the present study is the follow-up of individuals across the period before and during the pandemic, it is an epidemiological associational study, which permits no firm conclusions regarding causality. Finally, the generalization of findings beyond the Swedish pandemic context may be limited.

Conclusions

The present study examined the development of NRH and their associations with MHO during 2 years of the COVID-19 pandemic. The findings show that engaging in various types of NRH increased over time and was associated with better MHO. Due to the beneficial effects of spending time in natural environments on MHO in addition to other ecosystem services of green and blue space, decision-makers should prioritize investment in the preservation and good access to natural environments. Further, NRH should be facilitated also during crises such as the COVID-19 pandemic to promote public health and resilience. Such policies also constitute important contributions to meeting the United Nations Sustainable Development Goals in terms of good health and well-being alongside sustainable cities and communities (United Nations 2015).