Gender differences in emotional response to the COVID‐19 outbreak in Spain

Abstract Objective We aim to explore the differential presence of symptoms of anxiety, depression, and acute stress between men and women during the COVID‐19 outbreak, and to study the relationship between these symptoms and two environmental variables, coexistence, and violence. Methods We conducted a cross‐sectional study starting on March 29 to April 5, 2020, based on a national online survey using snowball sampling techniques. Symptoms of anxiety (Hamilton Anxiety Scale), depression (Beck Depression Inventory), and acute stress (Acute Stress Disorder Inventory) were assessed. Differences in the presence of symptoms and the relationship of coexistence and domestic violence were evaluated from a gender perspective. Results Men showed significant lower mean (SD) in anxiety, depression, and acute stress levels than women [HARS, 14.1 (9.8) versus. 18.4 (10.2), F = 56.2, p < .001; BDI 3.4 (3.9) versus 4.5 (4.3), F = 16.6, p < .001, and ASDI 3.6 (2.9) versus 4.7 (3.1), F = 39.0, p < .001, respectively), as well as a weaker depressive syndrome (28.1% males versus 39.9% females, χ2 = 25.5, p < .001). In addition, an interaction Gender × Coexistence was found in anxiety (F = 56.2, p < .001) and acute stress (F = 3.52, p = .06) and, according to depressive symptoms, an interaction Gender × Violence was found marginally significant (F = 3.3, p = .07). Conclusions Findings indicate that women present greater severity in symptoms of anxiety, depression, and acute stress. Moreover, loneliness and violence specifically worsen the emotional state in women. These results can undoubtedly guide better healthcare planning adopting a gender perspective.


| ME THODS
A total of 520 men and 1,115 women, aged from 18 to 84 (mean 40.4 ± 14.1 years), participated in this cross-sectional study previously published by our group , based on a national online survey applying an exponential nondiscriminative snowball sampling. The questionnaire was published on the hospital website and distributed by social networks to different geographical regions with the aim of reaching a representative sample of the Spanish population.
Up to 2,710 participants completed the self-reported questionnaire from March 29 to April 5, 2020, which covers the peak of the SARS-CoV-2 infection in Spain. Anyone with access to social networks who consented to participate and was over 18 completed the questionnaire. Of total respondents, healthcare workers and people with a current or past mental disorder were excluded for the present study as they have been considered especially vulnerable population groups for emotional reactions to COVID already predisposed to experience more stress and have been analyzed in a different study (PsyCOVID San Juan imas12).
Informed consent was provided, and confidentiality was assured.
The study was approved by Hospital de San Juan's local ethics committee.
Sociodemographic information on age, gender, and occupation was required, as well as whether responders had lived alone or with other people and had experienced situations of violence, either emotional or physical abuse, during the pandemic period.
To assess symptoms of anxiety and depression, we included the Hamilton Anxiety Scale (HARS) (Hamilton, 1959;Lobo et al., 2002) and the Beck Depression Inventory (BDI) (Bech, 1988;Vazquez & Sanz, 1999), respectively. For reporting the presence of acute stress, we adapted ad hoc for this study the clinical criteria for the diagnosis of Acute Stress Disorder (Acute stress disorder inventory-ASDI) of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) (APA, 2013). We developed a list of symptoms to be applied as self-reported questionnaire with a dichotomous answer (yes/no). In addition, the score of total affirmative responses to each of the items was recorded.
Differences between males and females in clinical variables were tested using one-way analysis of variance corrected for age and chisquared tests as appropriate. Then, the ANOVAs were repeated including the following sociodemographic variables: coexistence (living alone versus with other people), having experienced violence (yes versus no), economic losses (yes versus no), unemployment (yes versus no), COVID-19 infection (confirmed/suspected/discarded), and having experienced the passing of a loved one (yes versus no).
All these variables were asked referring to the pandemic period. The analyses presented have been corrected for age. Bonferroni corrections for multiple comparisons have been applied when appropriate.
All statistical analyses were considered statistically significant when p < .05.

| Relationship between gender and environmental variables
A significant interaction Gender × Coexistence was found in anxiety

According to depressive symptoms, the interaction Gender x
Violence was marginally significant (F = 3.3, p = .07). There were no differences in BDI scores between males who had experienced violence and those who had not. However, females who had experi- In the case of ASDI scores, the interaction Gender x Coexistence was marginally significant (F = 3.52, p = .06). The mean difference in ASDI scores between males and females who live alone (women measuring higher) is more than twice compared to the difference between those who do not live alone. No other variables showed an interaction effect with gender: violence (F = 0.96, p=.33), economic losses (F = 1.21, p=.27), unemployment (F = 0.16, p=.69), COVID-19 infection (F = 1.70, p=.18), or death of a loved one (F = 0.22, p=.64).
(Mean and SD of all these analyses are presented in Table 1).

| D ISCUSS I ON
The first aim of the present study was to explore the differential presence of symptoms between men and women during the peak of the COVID-19 outbreak in Spain. Findings indicate, as has also been observed in Europe and China (Gebhard et al., 2020;Liu et al.,;Wang et al., 2020), that women present greater severity in symptoms of anxiety, depression, and acute stress, which denotes an increase in arousal response to stress in females (Bangasser & Wicks, 2017) supporting sex differences in stress response systems, also during COVID-19. Hence, these results seem to point out that although the objective risk, due to morbidity and mortality ((CDC) C for DC and P, 2020) of the COVID-19 pandemic, is significantly greater for men, the emotional response is higher in women, which evidences the existence of other factors, beyond verified data on severity, influencing the emotional response.
The study further demonstrates that women living alone show more severe levels of anxiety, a fact that has not been observed in men, supporting the existence of gender differences that remain in the response to COVID and that may have to do with the predominant role of women as family caregivers and the greater susceptibility to social isolation (Gebhard et al., 2020