Citation

Yao X, Yu G, Tang J, Zhang J. Comput. Hum. Behav. 2021; 120: e106734.

Copyright

(Copyright © 2021, Elsevier Publishing)

DOI

10.1016/j.chb.2021.106734

PMID

unavailable

Abstract

Online depression communities (ODCs) are popular resources that help people to cope with their depression. We developed a comprehensive coding scheme to categorize the depressive symptoms mentioned in ODC postings on Sina Weibo. The symptoms were manually classified into five categories: emotional, cognitive, motivational, vegetative and physical, and seeking help. Text classifiers using deep learning methods were built to automatically identify these symptoms. Sixteen symptoms were extracted: (1) happiness, (2) anger, (3) sadness, (4) fear, (5) surprise, (6) disgust, (7) self-blame and self-criticism, (8) negative expectation, (9) low self-evaluation, (10) negative social cognition, (11) suicidality, (12) anxiety, (13) paralysis of the will, (14) avoidance, (15) sleep disturbance, and (16) seeking help. We evaluated which symptoms and associations between symptoms were most influential to the ODC members using network analysis methods. Suicidality was the most central symptom. A strong correlation was found between low self-evaluation and self-blame. Additionally, both anxiety and motivational avoidance were always accompanied by fear. Fear and sleep disturbance, negative expectation and suicidality usually co-occurred. Our findings might be applied to big data approaches for depression screening, providing an in-depth understanding of depressed individuals' health conditions so that better intervention strategies can be adopted to assist them.

Language: en

Keywords

Annotation scheme; Association; Depressive symptom; Online depression community; Sina Weibo; Text classifier