JMIR Publications

ABSTRACT

Background:

Studying mental health at scale requires perspectives that extend beyond traditional clinical and research settings. Increasingly, people turn to social media platforms like Reddit to make sense of their distress, seek support, and explore alternative ways of understanding mental health outside institutional frameworks.

Objective:

This study aims to map the landscape of mental health discourse on Reddit by constructing a data-driven network of associations between mental disorders and compare it with the network based on diagnostic criteria.

Methods:

We manually curated 114 Reddit communities focused on specific mental health conditions from the 20,000 most active subreddits in 2022. Each community was labeled into 49 disorders and categorized under nine ICD-10 diagnostic categories belonging to the group of mental and behavioral disorders (also known as “F” codes). We constructed a disorder association network by identifying statistically significant user overlaps across subreddit pairs using a bipartite configuration model, with Bonferroni-corrected significance (p < 0.001). Edges represent either positive or negative associations based on observed versus expected user overlap. We analyzed the connectiveness of the network within and across diagnostic categories, examining inter- and intra-category links. Finally, we compared the emergent Reddit-based hierarchy of disorder associations with the formal structure encoded in ICD-10 based on diagnostic criteria using hierarchical clustering and tree similarity metrics.

Results:

The inferred Reddit network of psychopathology revealed a densely interconnected structure, with all but six disorders forming a single giant component that spans across diagnostic categories. The most prominent disorders by number of users included ADHD, depression, habit and impulse disorders, autism, and generalized anxiety. However, Obsessive-compulsive disorder (OCD), Depersonalization/derealization (DPDR), and Post-traumatic stress disorder (PTSD) exhibited the highest connectivity by number of associations, while Schizotypal disorder, Avoidant personality disorder, and Agoraphobia ranked highest in weighted degree. Bipolar disorder and premenstrual dysphoric disorder (PMDD), displayed high inter-category associations but weak in-category ties, indicating complex diagnostic boundaries. The network of negative associations further revealed a divergence from the expectations of past research, for instance, addiction-related communities (e.g., alcohol and opioids) were negatively associated with much of the broader mental health discourse. Finally, hierarchical comparisons showed moderate overlap between Reddit-derived associations and ICD-10 diagnostic structures, both in pairwise edge similarity (Jaccard =0.07) and overall clustering (ARI=0.307, NMI=0.706).

Conclusions:

Reddit-based mental health communities reveal a complementary structure of disorder associations shaped by lived experience, often diverging from formal diagnostic criteria and highlighting overlooked or cross-cutting patterns of comorbidity.