Table 2.
Results of meta-analyses of prevalence of depressive symptoms among Chinese university students
| Subpopulation by variable | Number of studies | Sample size | Number of depressed students | Heterogeneity, I2 (%) (P) | Pooled prevalence (95%CI), % | Proportion transformation approach | Z | P | Cohen's da |
|---|---|---|---|---|---|---|---|---|---|
| Overall | |||||||||
| Depressive symptoms | 84 | 1 292 811 | 235 330 | 99.9 (<0.001) | 26.0 (23.3, 28.9) | Freeman−Tukey double arcsine | |||
| Severe depressive symptoms | 47 | 140 423 | 7831 | 99.4 (<0.001) | 1.69 (0.87, 2.77) | Freeman−Tukey double arcsine | |||
| Gender | |||||||||
| Male | 27 | 352 972 | 70 616 | 99.8 (<0.001) | 28.6 (21.4, 38.3) | Log | |||
| Female | 28 | 449 308 | 106 677 | 99.9 (<0.001) | 30.8 (24.3, 39.2) | Log | 119.34 | <0.001 | 0.270 |
| Ethnic group | |||||||||
| Han | 2 | 3079 | 565 | 99.4 (<0.001) | 21.0 (0.64, 41.3) | Untransformed | |||
| Minorities | 2 | 537 | 101 | 97.9 (<0.001) | 21.7 (0.00, 46.4) | Untransformed | 0.66 | 0.321 | 0.032 |
| Residence place | |||||||||
| Urban | 11 | 37 382 | 9796 | 99.9 (<0.001) | 27.4 (18.0, 41.8) | Log | |||
| Rural | 11 | 47 872 | 10 487 | 99.9 (<0.001) | 26.8 (16.3, 44.2) | Log | 6.41 | <0.001 | 0.044 |
| The only-child in the family | |||||||||
| Yes | 4 | 1931 | 309 | 98.3 (<0.001) | 20.7 (9.81, 43.6) | Log | |||
| No | 4 | 2194 | 419 | 99.2 (<0.001) | 24.2 (10.4, 56.5) | Log | 5.53 | <0.001 | 0.171 |
| Subject category | |||||||||
| Medical | 34 | 33 263 | 10 717 | 99.5 (<0.001) | 27.5 (21.0, 34.6) | Arcsine | |||
| Non-medical | 20 | 32 329 | 9181 | 99.7 (<0.001) | 27.5 (20.6, 36.7) | Log | <0.001 | 0.399 | <0.001 |
| Type of students | |||||||||
| Oversea students | 2 | 423 | 201 | 97.9 (<0.001) | 44.5 (13.5, 75.6) | Untransformed | |||
| Domestic students | 82 | 1 292 388 | 235 129 | 99.9 (<0.001) | 25.6 (22.9, 28.3) | Freeman−Tukey double arcsine | 12.32 | <0.001 | 0.844 |
| Grade | |||||||||
| Undergraduates | 42 | 1 183 315 | 206 398 | 99.9 (<0.001) | 22.9 (19.6, 26.3) | Freeman−Tukey double arcsine | |||
| Postgraduates | 14 | 20 303 | 4192 | 99.1 (<0.001) | 29.3 (21.6, 37.7) | Arcsine | 112.14 | <0.001 | 1.028 |
| Geographic location of current residence | |||||||||
| Hubei | 9 | 14 849 | 3977 | 95.0 (<0.001) | 27.5 (22.8, 32.2) | Untransformed | |||
| Non-Hubei | 22 | 768 375 | 162 354 | 99.5 (<0.001) | 22.3 (17.6, 27.9) | Logit | 130.35 | <0.001 | 1.033 |
| Location of the university | |||||||||
| Hubei | 8 | 23 290 | 6455 | 99.3 (<0.001) | 26.2 (19.5, 33.6) | Freeman−Tukey double arcsine | |||
| Non-Hubei | 59 | 1 190 886 | 207 920 | 99.9 (<0.001) | 23.1 (19.4, 27.2) | Logit | 66.74 | <0.001 | 0.543 |
| Close contact with COVID-19-infected persons | |||||||||
| Yes | 3 | 831 | 382 | 36.0 (0.2097) | 46.0 (42.6, 49.4) | Arcsine | |||
| No | 3 | 13 504 | 4244 | 99.8 (<0.001) | 25.0 (6.00, 44.1) | Untransformed | 101.75 | <0.001 | 1.500 |
| Having friends, classmates or relatives infected with COVID-19 | |||||||||
| Yes | 6 | 11 002 | 3973 | 89.0 (<0.001) | 39.7 (32.6, 48.4) | Log | |||
| No | 6 | 787 131 | 161 347 | 99.8 (<0.001) | 24.0 (17.5, 32.9) | Log | 202.82 | <0.001 | 1.973 |
a
Because sample sizes of different cohorts are very large, a statistically significant difference between two cohorts does not guarantee a clinical significant difference. To indicate the actual difference between two cohorts, Cohen's d was additionally calculated to assess the magnitude of the difference between the two rates, with 0.20–0.49, 0.50–0.79 and 0.80 and above being considered as small, medium and large actual differences, respectively. In the main text, we only reported the comparison results of different cohorts with Cohen's d values of approximately 0.20 or higher.