Development and External Validation of a Prediction Model to Identify Suicide Attempters in Treatment‐Naive Adolescents With Major Depressive Disorder.
Saved in:
| Title: | Development and External Validation of a Prediction Model to Identify Suicide Attempters in Treatment‐Naive Adolescents With Major Depressive Disorder. |
|---|---|
| Authors: | Song, Yuqin (AUTHOR), Dai, Mengqin (AUTHOR), Fan, Qiuyue (AUTHOR), Pan, Lu (AUTHOR), Wu, Yuhang (AUTHOR), Shao, Jiarui (AUTHOR), Lin, Cen (AUTHOR), Luo, Wenxiu (AUTHOR), Cen, Yu (AUTHOR), Xie, Cailin (AUTHOR), Wang, Xiangli (AUTHOR), Luo, Jiaming (AUTHOR), Chang, Yun-Hsuan (AUTHOR) |
| Source: | Depression & Anxiety (1091-4269). 2/27/2026, Vol. 2026, p1-10. 10p. |
| Subjects: | Mental depression, Boosting algorithms, Machine learning, Diagnosis, Teenagers, Self-injurious behavior, Suicide statistics, Suicide prevention |
| Abstract: | Background: Suicidal behavior in adolescents poses a significant risk, and suicide attempts are the strongest predictors of suicide death. Patients with Major Depressive Disorder (MDD) are at high risk of attempting suicide. However, there is still a lack of effective tools in clinical settings to identify these suicide attempters. Methods: The study assessed suicidal attempts and their predictive factors in adolescents first diagnosed with MDD from August 1, 2022, to May 31, 2024. Five algorithms were used for model construction: logistic regression, random forest, decision tree, support vector machine, and XGBoost. Finally, we evaluated the performance of the best model using an independent external validation set. Results: The study included 820 untreated adolescent first‐visit MDD patients (618 females [75.4%], average age 14.67 ± 1.69 years). Of these, 481 (58.7%) had disclosed suicidal ideation to others, and 299 (36.5%) reported having attempted suicide. Predictive variables for the outcome included age, grade, BMI z‐score, levels of depression and anxiety, sleep quality, history of being left behind, father's occupation, primary residence before age 16, history of non‐suicidal self‐injury (NSSI) within the last year, and history of disclosure of suicidal ideation. The XGBoost model showed the highest prediction accuracy (ROC_AUC, 0.72; PR_AUC, 0.65) and sensitivity (0.85) after external validation. The history of NSSI within the last year had the strongest predictive effect on suicide attempts, followed by disclosure of suicidal ideation, sleep quality, BMI z‐score, and anxiety levels. Conclusions: Despite including only 11 easily collectible clinical variables, the XGBoost model effectively identifies suicide attempters among untreated adolescent first‐visit MDD patients and performs stably in external validation sets. This is beneficial for clinicians to conduct evidence‐based suicide prevention efforts. [ABSTRACT FROM AUTHOR] |
| Copyright of Depression & Anxiety (1091-4269) is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.) | |
| Database: | Psychology and Behavioral Sciences Collection |
|
Full text is not displayed to guests.
Login for full access.
|
|
Be the first to leave a comment!
