Basic Paper Overview
– Journal: Global Health Action, Volume 19, Issue 1, published July 13, 2026 (open-access academic article)
– Title: Sex disparities in psychosocial distress among medical students with indirect exposure to the 2023 Kahramanmaras earthquakes
– Authors: Kadir Uludag, Fatih Kara, Hongxing Wang
– Core aim: Test whether male and female Turkish medical students show different levels of stress, anxiety and unhealthy habits after the 2023 Turkey earthquakes — even though the students did not live in the disaster zone themselves (only indirectly affected via family/friends)
1. Background & Rationale
1. Disasters cause mental harm not only to people physically on-site, but also to people with indirect trauma (worried about injured loved ones, overwhelmed by constant disaster news, disrupted school life).
2. Past disaster research consistently finds women have higher rates of anxiety, depression and PTSD, while men are more likely to cope via smoking, drinking alcohol. However, nearly all existing studies only focus on people directly hit by earthquakes.
3. Medical students are a high-priority understudied group: they are young and vulnerable to trauma, yet they are the future healthcare workforce needed for post-disaster recovery. Their medical training can be severely disrupted by secondary stress.
4. In Turkey, women already faced social and economic disadvantages before the earthquakes; disaster stress typically worsens gender gaps, though female mutual support networks can also act as a protective factor.
2. Study Methods
Participants
– Sample: 129 medical students from Kars Kafkas University (Turkey), surveyed roughly 3 months after the February 2023 Kahramanmaras earthquakes
– Demographic split: 67.4% female (87 people), 32.6% male (42 people)
– Two subgroups:
– 27.1%: Their immediate family lived in the earthquake zone and suffered harm
– 72.9%: No immediate family affected, only distressed via friends, classmates and round-the-clock media coverage
– Recruitment: Online surveys shared through Facebook and university WhatsApp groups; all participants submitted electronic informed consent. The study received ethical approval from Kafkas University Ethics Committee (approval date April 26, 2023).
Measured factors
1. Primary outcome: Anxiety severity (Turkish validated Beck Anxiety Inventory, scored from minimal to severe anxiety)
2. Covariates tested for gender differences:
Academic stress, peer bullying, smoking, alcohol use, exercise frequency, nutrition quality, economic status, hobbies, career satisfaction, academic performance, family earthquake impact status
Statistical tools
– Chi-square tests to check links between gender and each factor
– Two correction methods for multiple testing (to avoid false positive results):
1. Bonferroni correction: The strictest standard for statistical significance
2. Benjamini–Hochberg FDR: More lenient correction to reduce false negatives
– Binary logistic regression to identify which factors independently predict gender differences
– Model quality tested via cross-validation, R², AUC-ROC and Hosmer–Lemeshow calibration tests
3. Key Results
Main finding on gender differences
1. After strict Bonferroni correction: NO variables showed statistically significant gender differences at all.
2. After milder Benjamini–Hochberg FDR correction: Only peer bullying had a statistically significant association with sex (p_FDR = 0.046). This means peer bullying experiences differed between young men and women in the sample, but this result failed the strictest statistical test.
3. Several factors showed weak, near-significant raw trends before correction (smoking rates, anxiety severity, diet quality, alcohol use), but none held up after adjustment for repeated testing.
Subgroup comparison: Family affected vs. family unaffected by earthquake
– The proportion of male/female students was identical between the two groups.
– Students whose families suffered direct earthquake damage reported far worse anxiety and lower career satisfaction — proving closeness to trauma (family harm) is a stronger driver of mental distress than gender itself.
Logistic regression predictive model
1. The overall regression model was statistically significant, with moderate explanatory power.
2. Only regular physical exercise stood out as an independent significant predictor linked to gender: Students who exercised regularly had 4.18 times higher odds of falling into one gender group vs. the other (OR=4.18, p=0.031).
3. Model performance:
– Raw overall classification accuracy: 83.7%
– Sensitivity 94.3%, specificity 61.9%, AUC-ROC = 0.844 (good discrimination ability)
– 5-fold cross-validation average accuracy dropped to 70.6%, indicating mild overfitting (common issue with small sample size of 129 participants and 30 predictive variables).
Anxiety breakdown by gender
Raw, uncorrected data showed female students had noticeably higher rates of moderate and severe anxiety:
– Females: Moderate anxiety 25.3%, severe anxiety 24.1%
– Males: Moderate anxiety 14.3%, severe anxiety 11.9%
This visual gap did not reach statistical significance after formal testing.
4. Discussion & Interpretation
1. Broad gender gaps in post-disaster mental health largely disappear after rigorous statistical correction for this indirectly exposed medical student cohort. Initial apparent gender differences are mostly statistical noise from multiple tests, not true population differences.
2. Peer bullying is the only tentative gender-linked factor, but it is likely a pre-existing social difference between male and female students rather than a direct consequence of earthquake trauma.
3. Even though students lived nearly 1,000 km away from the earthquake epicenter, indirect trauma triggered clear anxiety — strongly supporting the Conservation of Resources Theory: fear of losing family safety and social support alone triggers severe stress, without any personal physical injury.
4. Family direct victim status is a far more powerful risk factor for poor mental health than gender. While women trended toward worse anxiety in raw data, the difference was not robust enough for clinical or policy intervention targeting gender alone in this group.
5. Conclusions & Recommendations
Core conclusion
No gender differences survived the strict Bonferroni statistical correction. Peer bullying had a marginal gender association under looser FDR correction only. Gender-based psychological disparities are very subtle among indirectly affected medical students after the Kahramanmaras earthquakes.
link of study：https://www.tandfonline.com/doi/full/10.1080/16549716.2026.2698916
Future research suggestions
1. Use long-term longitudinal study designs to track stress changes over months/years, separating short acute stress from chronic mental illness.
2. Recruit much larger sample sizes to boost statistical power, include both directly and indirectly exposed participants, and add m
