Depressive symptoms in response to COVID-19 and lockdown: a cross-sectional study on the Italian population

This study investigates the psychological impact of the COVID-19 pandemic and national lockdown in Italy, focusing on depressive symptoms in the general population. The research addresses whether specific sociodemographic, economic, household, and pandemic-related factors predict depressive symptoms, and whether regional pandemic intensity relates to mental health outcomes. Italy experienced an early and severe outbreak, with stringent national restrictions (Phase 1 lockdown: March 9–May 3, 2020; continued restrictions until June 2). The context included widespread disruption, social isolation, economic uncertainty, and an information overload (“infodemic”). Prior studies suggested increased anxiety, depression, and distress but relied on convenience samples likely to introduce selection bias. This study aims to provide population-representative evidence using a large, stratified random sample to evaluate depressive symptoms shortly after lockdown ended.

International evidence: Narrative and rapid reviews covering studies from China, Iran, Canada, Brazil, Singapore, India, and Japan reported elevated anxiety, depression, stress, and disturbed sleep during the pandemic. Risk factors included female gender, student status, symptomatic or suspected COVID-19 status, and poor perceived health. Stressors such as unpredictability, uncertainty, disease seriousness, misinformation, and social isolation were highlighted. Pre/post confinement comparisons showed increased mood symptoms (e.g., SMFQ total increased 44.9% during vs. before home confinement; p<0.001, d=0.44) and increased psychological distress post-lockdown alongside some resilience and community effects. Italian evidence: Web-based surveys during March–April 2020 reported high rates of post-traumatic stress symptoms (37.1%), depression (17.3%), anxiety (20.8%), sleep disorders (7.3%), and elevated perceived stress (21.9%), with females more impacted. Risk factors included quarantine, bereavement due to COVID-19, working discontinuity, and other pandemic-related stressors. Another survey (N=2,291) found associations between COVID-19 spread and anxiety (32.1%), psychological distress (41.8%), sleep disorders (57.1%), and PTSD symptoms (7.6%); younger age, student status, female gender, and direct COVID-19 contact increased risk. Additional studies (N=2,766) linked female gender, negative affect, and detachment with greater distress; depression levels ranged from average (67.3%) to high (17.0%) and extremely high (15.4%), with higher depression and stress among those with infected acquaintances and prior stress/medical problems. Another study (N=1,035) found 15.5% moderate and 6.2% severe depressive symptoms; psychological flexibility emerged as a protective factor. These Italian studies largely relied on online convenience samples, potentially introducing recruitment bias. The present study addresses this limitation using a random, representative sample.

Design and timing: Cross-sectional survey conducted June 4–19, 2020, immediately after Italy’s Phase 1 lockdown ended and inter-regional travel resumed (June 3), to capture immediate post-lockdown reactions. Sample: N=6,692 Italian individuals, representative of the population aged 16+ by age, gender, and geographical area. Sampling design and stratification included: (1) age (7 groups: 16–17; 18–24; 25–34; 35–44; 45–54; 55–64; 65+); (2) gender; (3) geographical breakdown (all Italian regions and 7 classes of community size); (4) education (graduates vs non-graduates). Young adults (16–24) and adults (25+) were analyzed separately given SMFQ validation history. Recruitment and data collection: Mixed-mode CATI (Computer Assisted Telephone Interviewing) and CAWI (Computer Assisted Web Interviewing) to limit self-selection bias. Participants received a description of aims and confidentiality; data were anonymous and collected in accordance with the Declaration of Helsinki. Ethical approval: Ethics Committee of Autorità per le Garanzie nelle Comunicazioni. Informed consent obtained; for minors under 18, consent from a parent/legal guardian was obtained. Measures: Primary outcome was depressive symptoms using the Short Mood and Feelings Questionnaire (SMFQ; 13 items). Responses: “not true”=0, “sometimes true”=1, “true”=2; total score range 0–26; higher scores indicate more severe depressive symptoms. A score ≥12 may indicate clinically relevant depressive symptoms. Reliability (Cronbach’s alpha): overall 0.91; young adults 0.90; adults 0.91. Independent variables: Sociodemographic: age (years), gender, education; economic/employment: unemployed, lay-off (professional uncertainty), poor (self-reported worst economic condition on a 5-point scale); household: living alone; work during lockdown: no lockdown (worked on-site vs stayed home); COVID-19 exposure: COVID-19 in family (self or family member tested positive); geography: size of municipality of residence (log of population). External pandemic indicators (from Italian National Department of Civil Protection data, as of May 31, 2020): county-level confirmed cases (absolute and percent of population), regional deaths (absolute and percent). Data analysis: Descriptive statistics and distributional analyses of SMFQ (including skewness tests). Nonparametric tests (Kolmogorov–Smirnov, Wilcoxon rank-sum, Epps–Singleton) to compare SMFQ distributions between age groups. Primary models: Poisson count models with double censoring (left at 0, right at 26), with sample weights and robust standard errors. Supplementary linear models provided in appendix. Probit models estimated probability of SMFQ ≥12 to interpret effects (e.g., as functions of age and exposure). Software: Stata and R.

Descriptives and reliability: SMFQ distribution was right-skewed with a mode near 0 and a slight bump around 12; overall mean ≈5.2. Cronbach’s alpha indicated excellent internal consistency (overall 0.91; young adults 0.90; adults 0.91). Prevalence: Using cut-off ≥12, 14.4% of the Italian population scored above threshold in June 2020. By age group: young adults (16–24) mean 7.04; adults (25+) mean 4.97. Percent ≥12: young adults 24.17% vs adults 13.33%. Nonparametric tests confirmed the two age groups’ SMFQ scores derive from different distributions. Predictors (Poisson count models with double censoring, weighted, robust SEs): - Gender (female): positive association with higher SMFQ scores (e.g., coefficient ~0.212–0.240; SE ~0.029–0.030; p<0.01). - Age: overall negative association (coef ≈ -0.0098 to -0.0108 per year; p<0.01), implying younger individuals had higher depressive symptoms. - Economic/employment: unemployed (≈0.126–0.152; p<0.05 to p<0.01), poor (≈0.213–0.219; p<0.01), lay-off (professional uncertainty) showed the strongest economic effect (≈0.316–0.481; p<0.01); the lay-off effect exceeded that of poverty (t-tests). - Household: living alone associated with higher SMFQ (≈0.166–0.195; p<0.01). - Geography (municipality size): log population negatively associated (≈ -0.020 to -0.024; p<0.01), indicating smaller towns linked to higher scores; linear form not significant. - COVID-19 in family: large positive effect (≈0.733–0.748; p<0.01); probit models indicated this approximately doubled the probability of SMFQ ≥12 across ages. - Lockdown work status: no lockdown (kept going to workplace) associated with lower SMFQ (≈ -0.122 to -0.123; p<0.05), suggesting those who stayed home had higher depressive symptoms. Authors note this likely underestimates the true burden of lockdown due to stress among essential workers. - Local pandemic intensity: county confirmed cases and regional deaths (absolute and percent) were not significant predictors; mental distress was not linked to local spread after controlling for other factors. Age-group models (young adults vs adults): - COVID-19 in family remained significant for both groups, with coefficients ≈0.485 (SE 0.087) for young adults and ≈0.770 (SE 0.0619) for adults. - Among young adults, family wealth (poverty) was not a significant predictor; municipality size was less important; the relation with age within 16–24 was positive (coef ≈0.101; SE 0.0143), while among adults it was negative (≈ -0.00954; SE 0.00113). Overall, 14.4% scored above depressive symptom threshold; females, younger individuals, those with professional uncertainty or low SES, living alone, staying home during lockdown, and having a family COVID-19 case were at higher risk. Regional location did not predict depressive symptoms.

The study demonstrates that depressive symptoms in Italy shortly after the national lockdown were unequally distributed across demographic and socioeconomic groups. Females and younger adults showed higher vulnerability, consistent with prior literature on gender and age differences in depression and distress. Economic strain and professional uncertainty (lay-off) were strongly associated with higher depressive symptoms, highlighting the mental health impact of the pandemic’s economic consequences. Living alone and residence in smaller municipalities were linked to higher depressive symptoms, suggesting social isolation and limited social interaction opportunities as key contributors during lockdown. Critically, having a COVID-19 case within the family exhibited the largest association with depressive symptoms and approximately doubled the probability of clinically relevant symptomatology at any age. Individuals who continued working on-site reported lower depressive symptoms than those confined at home, suggesting that maintaining routines and social contact may mitigate depressive mood despite potential anxiety about infection. The absence of an effect of regional pandemic intensity indicates that mental distress was widespread and not determined by local case counts, aligning with a nationally shared crisis and national-level lockdown measures. The SMFQ functioned reliably across age groups, supporting its utility in rapid, large-scale assessments during public health emergencies. These findings underscore the need for targeted mental health interventions for vulnerable groups and for policies integrating mental health support alongside infection control and economic measures.

This study provides population-representative evidence from Italy showing that COVID-19 and lockdown conditions were associated with elevated depressive symptoms, particularly among females, younger adults, those facing professional uncertainty or low socioeconomic status, individuals living alone, and those with a family COVID-19 case. Mental distress was not explained by regional differences in pandemic spread, indicating nationwide psychological impact. The work highlights the importance of integrating mental health support into pandemic responses and prioritizing vulnerable groups (e.g., youth, economically at-risk individuals, people living alone, and those directly affected by COVID-19 in the family). Future research should adopt longitudinal designs to track long-term trajectories, compare cross-country experiences, and monitor mental health service utilization to preempt a secondary wave of psychological distress. Telemedicine and digital tools may help deliver scalable mental health support and promote social connectedness.

- Cross-sectional design precludes causal inference between lockdown exposure and depressive symptoms. - Self-report measures may overestimate psychological distress. - The study focused on depressive symptoms (SMFQ) and did not assess other conditions such as anxiety, stress, or quality of life within the same dataset.