Introduction

In March 2020 in Italy, a nationwide quarantine was introduced as a measure to contain the transmission of the COVID-19 pandemic. The measure caused a radical change in the lifestyle of individuals, with unpredictable consequences on their mental and physical health. The impact that isolation and social distancing can have on lifestyle habits and their consequences on the levels of anxiety and depression has been studied in very specific contexts [1]. During the lockdown periods, stress, forced physical inactivity and lifestyle changes can produce changes in metabolic parameters and obesity, leading to an increase in risk of cardiovascular events [2], post-traumatic stress disorder and depressive symptomatology [3–5]. Little is known about the effect of quarantine on the academic community, including students and employees, that in the context of the COVID-19 pandemic did not represent a group at high risk of severe health consequences. However, given the high incidence of emotional disorders among university students, the impact of lockdown measures on their mental health can be particularly relevant [6, 7]. A few studies investigated the effect of COVID-19 pandemic on mental health of student during confinement [8–11]. A nationwide survey conducted in France among more than 69,000 university students reported prevalences of 27% and 16% for anxiety and depression respectively [12]. A study conducted during the COVID-19 lockdown on 2,530 students and workers of a Spanish university found that 21%, 34% and 28% of the participants suffered from anxiety, depression and stress, respectively [13]. Fear of becoming infected, worries for friends and relatives, economic uncertainty, loneliness can all be determinants of increased anxiety and depression during lockdown [5]. Other factors that can contribute to increase emotional distress are prolonged physical inactivity and other changes in lifestyle and working routine [14, 15]; the effects of physical activity on stress and anxiety during the COVID-19 pandemic have already been discussed in literature [16].

The aim of our study was to measure the level of psychological distress during the lockdown period in response to the COVID-19 pandemic and investigate its association with socio-demographic characteristics and lifestyle in a large sample of the Italian academic population.

Methods

Results

Among the 18,120 participants 9,062 (50%) were from universities in Northern Italy, 8,264 (46%) from Central Italy and 794 (4%) from Southern Italy. Socio-demographic and lifestyle characteristics of participants are reported in Table 1. Sixty-seven percent of the participants, 67% were females and 78% were students. Overall, 23% of participants reported having suffered influenza-like symptoms (temperature, cough, breathing problems).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. General characteristics of participants.

https://doi.org/10.1371/journal.pone.0248370.t001

The odds of having suffered from influenza-like symptoms were higher in younger subjects (OR for 10-year increase of age, 0.87; 95% CI, 0.80 to 0.95), in males than in females (OR, 1.19; 95% CI, 1.02 to 1.38), in individuals with comorbidities (OR for 1 versus no comorbidities, 1.71; 95% CI, 1.54 to 1.91; OR for 2 or more versus no comorbidities, 2.19; 95% CI, 1.73 to 2.77), in those worried to contract the virus (OR, 1.23; 95% CI, 1.06 to 1.42), in those judging the measures by the government not adequate (OR for insufficient versus adequate measures, 1.31; 95% CI, 1.11 to 1.53) and in individuals with lower levels of physical activity during lockdown (OR for >4 versus <1 hours per week, 0.71; 95% CI, 0.62 to 0.81; S1 Table).

The HADS questionnaire was filled by 17, 074 (94%) participants; for anxiety, borderline and severe levels were reported by 16% and 15% of respondents; for depression, 14% and 8%; these figures resulted in 20% of participants being either severely anxious or severely depressed. The results of the linear association between overall psychological distress and socio-demographic and lifestyle variables are reported in Table 2. All variables, with the exception of the presence of cohabitants, were significantly associated with the psychological distress both in the univariate and in the multiple model. Younger individuals, females and students had higher levels of psychological distress. Having a low income or low level of education in the family, both proxies of the socio-economic status, were related to higher levels of psychological distress; similar results were observed for sharing the house with old or disabled people. Working in contact with the public was associated with lower HADS levels, whereas having cohabitants working with the public was associated with higher levels. Also, higher HADS levels were associated with having a frail health, having experienced COVID-like symptoms, and being worried to become infected; on the contrary, lower HADS levels were associated with the belief that the measures adopted were adequate and with trust in doctors and institutions.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Linear regression of HADS.

https://doi.org/10.1371/journal.pone.0248370.t002

HADS scores significantly decreased with levels of physical activity during quarantine and the association was not confounded by any of the variables we were able to adjust for. In particular, compared with those who did not exercised during quarantine, those reporting the highest level of physical activity had on average a HADS score 2.46 points lower (95% CI, 2.07 to 2.85; Table 2). In order to account for the effect of physical activity before quarantine, we fitted a model with both variables and their interaction. According to this model, those who performed physical activity before quarantine and then stopped to exercise had higher HADS scores than those who never exercised (EMM, 14.5; 95% CI, 14.2 to 14.8, versus 13.0; 95% CI, 12.6 to 13.4). Physical activity during quarantine was inversely associated with HADS scores (p < 0.001) irrespective of the physical activity status before quarantine, but its effects were more pronounced among those who exercised before quarantine than among those who did not: the estimated differences of the HADS score between those with the highest versus lowest level of physical activity were -1.8 (95% CI, -2.8 to -0.8) and -3.0% (95% CI, -3.4 to -2.5) in the group who did not exercised before lockdown and in those who did, respectively. As a result, HADS scores in those with the highest level of physical activity were very similar in the two groups (Table 3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. HADS estimated marginal means (EMMs) by levels of physical activity.

https://doi.org/10.1371/journal.pone.0248370.t003

Table 4 reports the association between anxiety and depression status and physical activity before and during quarantine in terms of OR of being borderline or being a severe case: compared to those who never exercised neither before nor during lockdown, the highest OR of being severely anxious or severely depressed was observed in those who stopped exercising (OR, 1.53; 95% CI, 1.28 to 1.84). Performing physical activity during lockdown was associated with a lower risk of being severely anxious or severely depressed in both groups: compared to those who never exercised, the odds for those who exercised during quarantine were 11% lower, although not statistically significant, if they were not physically active before quarantine, (OR, 0.89; 95% CI, 0.71 to 1.13) and 22% if they were physically active also before quarantine (OR, 0.78; 95% CI, 0.64 to 0.95). If during quarantine all individuals could continue to perform physical activity with the same frequency as before, the number of severe cases in the study population would have been 14% lower for anxiety (95% CI, 7 to 21%) and 21% lower for depression (95% CI, 10 to 32%). Similar results have been obtained for the two HADS subscales anxiety and depression separately (S2–S5 Tables).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Effect of physical activity on psychological distress: Odds ratio of being borderline versus normal and severe versus normal.

https://doi.org/10.1371/journal.pone.0248370.t004

Discussion and conclusions

The results of this large cross-sectional study highlight the implications of quarantine on psychological distress in the adult population. Twenty percent of the participants reported severe levels of anxiety or depression. The factors associated with a higher level of psychological distress were being a student, having a low income or a low educational level within the family, having suffered of influenza-like symptoms or having a frail health, living with older or disabled people; all the observed associations were independent of the fear of being infected by the virus. Psychological distress was associated with distrust of the government and medical and scientific community; individuals reporting the higher level of distress were also those who believed that the measures implemented to contain the spreading of the virus were inadequate, either insufficient or too strict. With regard to physical activity, those who stopped exercising during quarantine had higher HADS scores. On the contrary, being able to perform some level of physical activity during quarantine was associated with lower HADS scores, irrespective of previous level of physical activity.

The proportion of participants in our survey reporting severe levels of anxiety or depression is similar to the proportions observed in other studies targeting European university communities: in a French and in a Spanish study, anxiety amounted to 27% and 21% and depression to 16% and 34%, respectively [12, 13].

In our population, those mostly affected by psychological distress were students and individuals with lower income, representing two groups of individuals highly suffering the occupational and economic uncertainty associated to the pandemic. Depression is known to be associated with the socio-economic status [21] and others reported that, during the COVID-19 pandemic as well as during other disasters, mental disorders were not evenly distributed in the population, being their incidence higher among disadvantaged groups of the population [22, 23].

The observed association between psychological distress and distrust of institutions, including government and medical staff, could be explained by the fact that people with an external locus of control—that is individuals believing that outcomes in their lives are determined by external factors—tend to cope worse with negative changes than those with an internal locus of control–that is individuals believing that outcomes in their lives depend largely on their own actions [24–29].

Stratifying our sample by level of physical activity before the lockdown period, we observed that subjects able to exercise during quarantine reported lower psychological distress irrespective of their previous level of physical activity. The association between level of physical activity during the lockdown period and more severe levels of anxiety and depression has been reported by others [16, 30–34]. One large cross-sectional study among French university students found association of low levels of physical activity with anxiety and depression [12]; one longitudinal study in the United Kingdom before and after lockdown showed positive association between change in perceived stress and change in sedentary behavior [10]. The evidence that mental health is associated with levels of physical activity independently of age and socio-economic status mainly comes from observational studies [15, 35] and clinical trials on the effect of exercise as treatment for depressions and anxiety [36–38]. Results from studies in animals and humans indicate that the effect of physical activity on depressive symptoms might be mediated through factors associated to monoamine metabolism and immune and neurogenic markers [39]. Not surprisingly, the cessation of physical activity has a negative impact on mental health [39, 40]. Accordingly, our data showed that those who stopped to exercise due to home confinement and closure of sport facilities experienced the most severe psychological distress. Although our results are in line with the literature supporting the hypothesis that reducing physical activity increases psychological distress, we cannot rule out a mechanism of reverse causality, resulting from a reduction of physical activity in the individuals more anxious or depressed.

The main strength of our study is its large sample size and the wide range of variables assessed, allowing to investigate the role of many potential determinants of psychological distress and to control for potential confounding. We targeted the entire university community, including students as well as the teaching, technical and administrative staff, in order to have access to a wide range of exposures and to be able to compare the conditions of students with those of the working individuals. The instrument used to assess the level of psychological distress was developed for hospitalized patients but is able to capture anxiety and depression symptoms in other contexts and samples. Indeed, the HADS proved to have good construct validity and other psychometric properties not only in clinical but also in general population samples [19, 20].

The representativeness of our sample at the level of the nationwide academic population is supported by the multicenter study design that we have adopted. However, because the survey has been completed on voluntary basis, we cannot exclude that a selection bias of participants based on their psychological distress levels might have occurred. However, there is no reason to believe that the potential selection might have affected the association between psychological distress and the predictors. Moreover, the large number of variables collected allowed to adjust the analyses for a wide range of potential confounders.

The negative impact of the COVID-19 pandemic on mental health is a matter of concern and experts urge for action to monitor and control its negative effects especially among most vulnerable individuals [41]. The knowledge accumulated from large epidemiological studies conducted during the first wave of COVID-19 can contribute to face future waves of infection or epidemic by identifying highly susceptible groups and possible intervention approaches. The study IO CONTO 2020, investigating the factors associated with anxiety and depression in a large sample of individuals, is placed in this perspective of action. Identifying economic instability among the factors associated to the increased risk of psychological impairment, we addressed once more the well-known consequences of social impairment on the health of a population. In the light of the finding that physical activity could at least partially relieve the psychological burden of a health crisis, we might suggest that future measures adopted to contain the spreading of a disease guarantee the access to outdoor spaces or safe indoor environment where citizens can perform their exercise. Further investigations are needed to quantify the long-term impact that COVID-19 quarantine had on the population.

Supporting information

Acknowledgments

We thank the universities of Pisa, Florence, Genoa, Turin and Messina for their administrative and technical support to the study IO CONTO 2020 and all participants for their time.