[Bidirectional longitudinal associations of perceived stress with anxiety and depressive symptoms among medical students: A three-wave cross-lagged panel study].

Authors

Zhenhua Chen, Liying Shen, Lin Wang, Qian Xiao

Published in

Zhong nan da xue xue bao. Yi xue ban = Journal of Central South University. Medical sciences. Volume 51. Issue 7. Pages 1462-1475. Jul 28, 2026.

Abstract

Perceived stress, anxiety symptoms, and depressive symptoms may co-occur among medical students, but their temporal ordering and bidirectional longitudinal relationships remain unclear. Whether conventional cross-lagged panel model (CLPM) adequately reflect within-person dynamic changes also requires examination through sensitivity analyses. This study aims to examine changes in perceived stress, anxiety symptoms, and depressive symptoms across 3 assessments among medical students and to investigate the bidirectional longitudinal associations among these 3 variables.
A three-wave longitudinal design was used. Medical students were surveyed in June, September, and December 2025, with valid sample sizes of 376, 361, and 343 at T1, T2, and T3, respectively. Perceived stress, anxiety symptoms, and depressive symptoms were assessed using the Chinese Perceived Stress Scale (CPSS), the Generalized Anxiety Disorder Scale-7 (GAD-7), and the Patient Health Questionnaire-9 (PHQ-9), respectively. Linear mixed models were used to examine changes in the main variables across the 3 time points, with Bonferroni correction for pairwise comparisons. Pearson correlation analysis was used to examine correlations among variables. The primary analysis used a three-variable CLPM, in which CPSS, GAD-7, and PHQ-9 were simultaneously included to examine bidirectional longitudinal statistical associations among perceived stress, anxiety symptoms, and depressive symptoms. Based on the primary three-variable CLPM, two-variable CLPMs for CPSS-GAD-7 and CPSS-PHQ-9 were additionally fitted as supplementary analyses to examine the longitudinal associations of perceived stress with anxiety and depressive symptoms separately. Random-intercept cross-lagged panel models (RI-CLPMs) were further fitted as sensitivity analyses to examine the interpretive boundaries of the conventional CLPM findings after distinguishing stable between-person differences from within-person dynamic changes. All longitudinal models were adjusted for sex and age, and parameters were estimated using robust maximum likelihood estimation.
All 3 measures showed significant time effects, with scores being highest at T2, followed by T3 and T1 (all P<0.001). The three-variable CLPM showed good model fit, and CPSS, GAD-7, and PHQ-9 all demonstrated significant longitudinal stability (β=0.475 to 0.701, all P<0.001). After adjustment for sex and age and inclusion of autoregressive paths and within-wave correlations, perceived stress at both T1 and T2 was positively associated with anxiety and depressive symptoms at the subsequent wave (all P<0.05). Anxiety symptoms at T1 and T2 were positively associated with perceived stress at the subsequent wave, and depressive symptoms at T2 were positively associated with perceived stress at T3 (all P<0.05); however, the path from depressive symptoms at T1 to perceived stress at T2 was not significant (P=0.417). Anxiety and depressive symptoms showed bidirectional positive longitudinal associations across both time intervals (all P<0.05). Except for the nonsignificant path from PHQ-9_T1 to CPSS_T2, the standardized coefficients of the main cross-lagged paths ranged from 0.082 to 0.232, indicating that the associations were generally small in magnitude. The three-variable model explained 61.3%, 55.8%, 59.6%, 64.6%, 61.7%, and 51.0% of the variance in CPSS_T2, CPSS_T3, GAD-7_T2, GAD-7_T3, PHQ-9_T2, and PHQ-9_T3, respectively. The supplementary two-variable CLPMs yielded results generally consistent with those of the three-variable model. In the RI-CLPM sensitivity analyses, both the three-variable RI-CLPM and the unconstrained perceived stress-depressive symptoms RI-CLPM produced non-positive definite latent covariance matrices. In contrast, the unconstrained perceived stress-anxiety symptoms RI-CLPM yielded an admissible solution, with some cross-lagged paths reaching statistical significance.
Perceived stress, anxiety symptoms, and depressive symptoms showed longitudinal associations among medical students. The three-variable CLPM indicated that, after accounting for the interrelations between anxiety and depressive symptoms, perceived stress remained prospectively associated with subsequent anxiety and depressive symptoms, and anxiety and depressive symptoms also showed bidirectional longitudinal associations. RI-CLPM sensitivity analyses suggested that these findings should be interpreted primarily as longitudinal statistical associations within the conventional CLPM framework rather than as evidence of clear within-person causal effects; in particular, the within-person dynamic relationship between perceived stress and depressive symptoms requires further validation.

PMID:
42763301
Bibliographic data and abstract were imported from PubMed on 20 Sep 2026.

Read full publication at:
Please sign in to see all details.

Stats

1-terrible, 9-excellent. How would you rate this publication? Sign in in to submit your rating.

Post a comment

You need to be signed in to post comments. You can sign in here.

Comments

There are no comments yet.