Prevalence of and factors associated with post-traumatic stress disorder among French university students 1 month after the COVID-19 lockdown

The study addresses whether the COVID-19 pandemic context and associated lockdown measures constitute potentially traumatic events leading to PTSD among university students, a population at high risk for mental health problems. Given widespread lockdowns (e.g., France’s March 17–May 11, 2020 lockdown) and prior evidence of negative psychological effects of quarantine, the research aims to estimate the prevalence of probable PTSD and identify associated risk factors one month after lockdown lifting, and to explore which COVID-19-related events students perceive as traumatic. It also examines psychological response trajectories (resilience, recovery, persistence, delayed onset) from lockdown to one month post-lockdown.

Prior research during the initial pandemic phase showed high rates of moderate-to-severe distress among university students (e.g., 53.8% in China; 28.1% in France). Early studies on PTSD prevalence among quarantined students reported heterogeneous estimates (2.79–30.8%), often using the PCL-C based on DSM-IV criteria, which differs from the 20-item PCL-5 aligned with DSM-5 and demonstrating good reliability in trauma-exposed college populations. Evidence supports PTSD as an outcome of COVID-19-related events (e.g., hospitalization or death of a relative), but PTSD arising from the broader pandemic and lockdown context remains debated.

Design: Repeated cross-sectional, university-based COSAMe survey with multiple waves. This analysis focuses on the second measurement time (T2), conducted June 15–July 15, 2020, one month after France lifted its first COVID-19 lockdown (March 17–May 11, 2020). Some participants also completed the first wave (T1: April 17–May 4, 2020); responses were linked via pseudonymization. Inclusion: University students residing in France during lockdown. Recruitment: The French Ministry of Higher Education invited all 82 universities (≈1.6 million students) to email students to complete an anonymous online questionnaire (no compensation). Ethics: Approved by Comité de Protection des Personnes Ile de France VIII. Measures at T2: Probable PTSD assessed with the French PCL-5 (20 items, 0–4 per item, total 0–80), with a cutoff >32 indicating probable PTSD; internal consistency previously reported (Cronbach’s alpha 0.79–0.94). DSM-5 symptom cluster algorithm also applied: at least one re-experiencing, one avoidance, two negative cognition/mood, and two arousal symptoms (symptom present if item score >1). Covariates included: socio-demographics (age, gender: male, female, other); clinical (history of psychiatric follow-up; exposure to non-COVID-19 traumatic event after pandemic onset); social support (foreign student; living alone during quarantine; quality of social relationships during quarantine [0–10]; pre-quarantine social integration [0–10]); socio-economic (loss of income; housing quality [0–10]); quality of information received about pandemic and measures [0–10]. COVID-19 exposure scale (0–7) derived from seven yes/no items: residence in worst-hit area (>50 deaths by Mar 29, 2020); COVID-19-compatible symptoms; contact with infected person; death of a relative due to COVID-19; high fear for relatives’ health; high fear for own health; high media exposure (last three dichotomized at >3rd quartile). For T1 respondents, distress was measured using IES-R (0–88; normal 0–23, mild 24–32, moderate 33–36, severe >36). At T2, students indicated whether specific pandemic-related events were perceived as traumatic (e.g., news of epidemics, closures, quarantine, infection status, hospitalization, death of relatives). Statistical analysis: Included only fully completed questionnaires. Descriptive statistics used percentages, means (SD), or medians (IQR) as appropriate. Prevalence of probable PTSD and 95% CIs reported using both PCL-5 cutoff and DSM-5 algorithm; concordance evaluated by Cohen’s kappa. Multivariate logistic regression identified factors associated with probable PTSD (PCL-5 >32) including all explanatory variables; results reported as adjusted odds ratios (ORs) with 95% CIs. Among T1–T2 paired respondents, prevalence of probable PTSD at T2 was examined by T1 distress level, and response trajectories (resilience, persistence, recovery, delayed onset) were computed following Galatzer-Levy definitions. Analyses used R 3.6.1; alpha = 0.05; two-tailed tests.

Sample: 22,883 students at T2; 72.7% women; mean age 21±4 years. Psychiatric history 10.6%; exposure to non-COVID traumatic event 14.1%. Foreign students 6.0%; lived alone during quarantine 11.6%. Loss of income 18.3%. Median scores (IQR): pre-quarantine social integration 7 (6–8)/10; social ties during quarantine 6 (4–8)/10; housing quality 9 (7–10)/10; quality of information 6 (4–7)/10. COVID-19 exposure: 28.1% in affected department; 14.4% contact with infected; 5.8% knew someone who died; 23.1% had compatible symptoms; worry for relatives’ health 7 (5–8), own health 3 (1–5); media exposure 20 (5–45) min/day; exposure score median 1 (0–2)/7. Prevalence: Probable PTSD by PCL-5 >32: 19.5% (4456/22,883; 95% CI 19.0–20.0). By DSM-5 algorithm: 19.4% (4449/22,883; 95% CI 18.9–20.0). Agreement 91.7%; Cohen’s kappa 73.8% (95% CI 72.7–74.9). Factors associated with probable PTSD (adjusted OR [95% CI]): Age 0.98–0.99 (p=0.003). Gender: Female 1.32 [1.21–1.45]; Non-binary 1.76 [1.35–2.31] (ref: male). Psychiatric history 2.26 [2.05–2.51]. Exposure to other (non-COVID) traumatic event 3.37 [3.08–3.67]. Social: Living alone 1.22 [1.09–1.37]; Foreign student 1.70 [1.48–1.95]; Quality of social ties vs high: medium 1.42 [1.30–1.55], low 2.38 [2.15–2.62]; Feeling integrated vs high: medium 1.56 [1.44–1.69], low 2.21 [1.95–2.51]. Socio-economic: Loss of income 1.20 [1.09–1.31]; Housing quality vs high: medium 1.60 [1.45–1.76], low 1.90 [1.59–2.26]. Information quality vs high: medium 1.26 [1.15–1.37], low 1.50 [1.35–1.66]. COVID-19 exposure score (ref 0): 1: 1.38 [1.24–1.54]; 2: 2.02 [1.81–2.26]; 3: 3.07 [2.71–3.47]; 4: 4.62 [3.95–5.41]; 5: 6.87 [5.32–8.87]; 6: 8.17 [4.79–14.06]; 7: 10.82 [2.33–76.57]. All p<0.001 except exposure score 7 p=0.005. Perceived traumatic events (overall vs among probable PTSD): Death of a relative 88.3% (highest); hospitalization of a relative 82.1%; own hospitalization 76.8%; infection of a relative 68.8%; own infection 60.8%; having symptoms 53.2%; quarantine 66.4% overall; among probable PTSD, quarantine was second most endorsed at 78.8%. Lower endorsements: news of epidemic in France 37.0%; closure of schools/universities 33.2%; closure of bars/shops/meeting places 27.9%; news of epidemic in China 10.8%. Response trajectories among 6,947 with T1 and T2: Probable PTSD at T2 overall 16.4% (1,140; 95% CI 15.5–17.3). PTSD at T2 by T1 distress: none 4.3% (176/4105); mild 17.1% (181/1061); moderate 25.0% (93/372); severe 49.0% (690/1409). Trajectories: Resilience 73.2% (5,088; 95% CI 72.2–74.3); Recovery 10.3% (719; 95% CI 9.6–11.1); Persistence 9.9% (690; 95% CI 9.2–10.7); Delayed onset 6.5% (450; 95% CI 5.9–7.1).

Findings indicate that nearly one in five French university students experienced probable PTSD one month after the initial COVID-19 lockdown, substantially exceeding pre-pandemic estimates in the French general population (~0.7%). Multiple factors increased risk, including female or non-binary gender, previous psychiatric history, exposure to other traumatic events, social isolation during quarantine, being a foreign student, weaker social ties and integration, economic stressors (income loss, poorer housing), lower perceived information quality, and higher COVID-19 exposure. The graded increase in odds with higher exposure scores supports a dose–response relationship. Students frequently considered quarantine itself potentially traumatic, especially among those with probable PTSD, underscoring the potential traumatic impact of lockdown measures in addition to direct illness-related events. Longitudinally among paired respondents, most students followed a resilience trajectory, but notable proportions showed persistence, recovery, or delayed onset, demonstrating heterogeneity in psychological responses to the pandemic context. These results contribute to the debate about whether pandemic and lockdown experiences can meet PTSD trauma criteria and highlight at-risk subgroups needing targeted mental health support.

This study quantifies a high prevalence of probable PTSD in French university students one month after COVID-19 lockdown and identifies a broad set of associated demographic, clinical, social, economic, informational, and exposure-related risk factors. Many students perceived quarantine as potentially traumatic, and response trajectories showed that while resilience predominated, substantial minorities experienced persistent or delayed difficulties. The findings suggest that pandemic and lockdown contexts may have post-traumatic consequences and inform the ongoing nosographic debate about PTSD. Future research should refine causal inference, examine long-term outcomes beyond one month post-lockdown, validate trauma appraisals of lockdown experiences, and evaluate interventions targeting identified risk factors and vulnerable subgroups (e.g., socially isolated, foreign students, those with prior trauma or psychiatric history).