Portrayals of Depression on TikTok: Content Analysis of Diagnostic Accuracy, Creator Type, and Stylistic Features

Youths are increasingly turning to TikTok for mental health information, making the platform an important space where young people encounter portrayals of mental illness. While such visibility can raise awareness, reduce stigma, and make young people feel more connected and understood in their experiences, concerns have been raised about the diagnostic accuracy of this content, which is often produced by nonprofessionals and presented using emotionally appealing stylistic features. Although prior research has examined mental health content on TikTok broadly, little is known about how depression-related symptoms are portrayed by creators on the platform.

Given depression's rising prevalence among youth and its prominent presence on TikTok, this study examined (1) the diagnostic accuracy of TikTok videos about depression, (2) differences in diagnostic accuracy and stylistic features by creator type (medical professionals vs nonprofessionals), and (3) how diagnostic accuracy, stylistic features (personal experiences, emotional appeals, and background music), and creator type relate to user engagement.

A quantitative content analysis was conducted of 210 English-language TikTok videos retrieved using symptom-focused search terms (eg, "depression symptoms"). Videos were coded for diagnostic accuracy using a standardized coding scheme based on the International Classification of Diseases, 11th Revision diagnostic criteria for depressive episodes. In addition, videos were coded for creator type, presentation style, and the presence of emotionally appealing stylistic features. Engagement was operationalized as the sum of a video's likes, comments, saves, and shares. Intercoder reliability was assessed using Krippendorff α, percent agreement, and Gwet AC1 (agreement coefficient 1). Analyses included Mann-Whitney U tests, chi-square tests, and hierarchical regression.

Diagnostic accuracy was low overall (mean score 1.21, SD 1.04, on a 0-4 scale) and did not differ significantly between medical professionals and nonprofessionals (median 1.40 [IQR 1-2] vs 1.11 [IQR 0-2]; P=.06). Hierarchical regression analysis showed that diagnostic accuracy did not predict engagement (B=-0.10; P=.19). In contrast, engagement was higher for videos containing personal experiences (B=0.41; P=.02), emotional appeals (B=0.73; P=.001), and background music (B=0.54; P=.01). Across regression models, direct-to-camera formats (Bs -0.49 to -0.69; .003≤P≤.04) and text-centered videos (Bs -0.56 to -0.64; .002≤P<.01) were associated with lower engagement.

Depression-related content on TikTok is characterized by limited diagnostic completeness, regardless of creator type. Engagement appears to be driven primarily by stylistic features rather than diagnostic accuracy. These patterns raise concerns about concept creep-the gradual expansion of the psychological concept for depression-and the potential for premature self-diagnosis among young users, while also highlighting opportunities for medical professionals to adapt their communication styles on TikTok to increase both accuracy and engagement.