Abstract
A heavy rainstorm struck Henan, China, in July 2021. Previous studies have indicated that natural disasters have a wide range of psychological sequelae, but little research has been done on the psychological effects of floods specifically. This study aimed to track the mental health trajectories of flood victims over time and identify associated protective and risk factors. People living in the areas most impacted by the flood (N = 376) were surveyed at four different time points: 3 weeks, 1 month, 2 months, and 3 months post-flood. Latent growth mixture modeling was utilized to delineate longitudinal patterns of depression, anxiety, and posttraumatic stress disorder (PTSD). Least absolute shrinkage and selection operator logistic regression, a supervised machine-learning approach, was employed to discern predictors among 49 assessed variables at both contextual and personal levels. Results revealed three classes of PTSD (resilience, recovery, moderate symptoms) and anxiety (resilience, chronicity, recovery) trajectories, along with two classes of depression trajectories (resilience, chronicity). Key factors predicting resilience in mental health included personality traits, media consumption habits, pre-existing health conditions at the individual level, and asset loss and ongoing adversities at the contextual level. In spite of the widespread impact of the flood, most victims displayed resilience in the face of adversity. Identifying critical factors across various psychological symptoms offers valuable insights for both pre-disaster preparation and post-disaster trans-diagnostic psychological interventions.