Journal
JOURNAL OF AFFECTIVE DISORDERS
Volume 295, Issue -, Pages 264-270Publisher
ELSEVIER
DOI: 10.1016/j.jad.2021.08.028
Keywords
Major depressive disorder; Suicide attempt; Risk factor; Machine learning; XG-boost
Categories
Funding
- Natural Science Foundation of China [81871071, 81171251]
- Natural Science Foundation of Beijing [7162101]
Ask authors/readers for more resources
By integrating demographic, clinical, and biological variables, risk factors for suicide attempt among patients with MDD can be identified as early as possible during the first time they visit a doctor.
Background: Major depressive disorder (MDD) is the most common mental disorder associated with suicide attempts. When a patient first visits the clinic, clinicians are often expected to make concrete diagnose about acute suicidal risk. However, the timeliness of suicide attempts correlates with patients with MDD has not been tested. Methods: We divided 1718 first-episode and untreated MDD outpatients into those who did not have suicide attempts (non-attempts), recent suicide attempters (<= 14 days before assessment) and long - dated suicide attempters (> 30 days before assessment). Positive Symptom Scale of Positive and Negative Syndrome Scale (PANSS), the 17-item Hamilton Depression Scale, 14 - item Hamilton Anxiety Scale, and clinical global impression of severity scale (CGI-S) was assessed. Body mass index, some glycolipid metabolism and thyroid hormone parameters were measured. A gradient-boosted decision trees statistical model was used to generate equally weighted classification for distinguishing recent and long - dated suicide attempters from non-attempts. Results: The classifier identified higher excitement, hostility, anxiety, depression symptoms and higher free thyroxine (FT4) as risk factors for recent suicide attempters with an estimated accuracy of 87% (sensitivity, 59.1%; specificity, 61.2 %). For long - dated suicide attempters' risk factors, single status, higher anxiety and hostility symptoms, higher LDLC and lower BMI, the estimated accuracy was 88% (sensitivity, 52.8%; specificity, 49.6%). Conclusions: Risk factors for suicide attempt among patients with MDD can be identified by integrating demographic, clinical, and biological variables as early as possible during the first time see a doctor.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available