Latent classes and predictors of depressive and anxious symptoms in Chinese adolescents: A two-wave longitudinal survey

Background: Among Chinese junior high school students, mental health problems such as depressive and anxious symptoms endanger wellbeing. This study explored depressive and anxious symptom classes using the latent class analysis (LCA) method and determined the related predictors. Methods: We used a longitudinal research design with data from the two waves of the China Education Panel Survey (CEPS) involving 6,229 junior high school students. LCA was used to classify depressive and anxious symptoms according to ten questions. Related predictors at the personal, interpersonal, and environmental levels of the latent classes were analyzed using multivariable-adjusted and propensity score-adjusted multinomial logistic regression. Gender subgroup analysis was performed. Results: Depressive and anxious symptom classes were classified as follows: limited depressive and anxious symptoms (n = 2,616, 42.00 %); anxious symptoms (n = 1,601, 25.70 %); depressive symptoms (n = 1,055, 16.94 %); and co-existing depressive and anxious symptoms (n = 957, 15.36 %). Multivariable-adjusted multinomial logistic regression results showed that female, boarding at school, myopia, sleep time of less than 8 h/night, poor parental relationship, poor father-child or mother-child relationship, and a lack of high confidence in their future were more likely to exhibit depressive or anxious symptoms. The depressive and anxious symptom classes were affected by different factors between the genders. The observed significant associations produced by the original multivariable-adjusted and propensity score-adjusted multinomial logistic regression models were similar. Limitations: As a 1-year longitudinal study, there is limited efficacy in determining causality from the results. Conclusion: Four latent classes of depressive and anxious symptoms were identified in Chinese junior high school students. Treatment interventions should be developed and implemented based on the individual latent classes for improved mental health.