Biomarkers associated with future suicide risk enhance predictive performance in psychiatric inpatients - Summary - MDSpire
Coming Soon: Introducing MDSpire News. Learn more
Conexiant’s news site is now MDSpire News. Learn more

Biomarkers Linked to Future Suicide Risk Improve Predictive Accuracy in Psychiatric Hospitalized Patients

  • By

  • Zheya Cai

  • Enzhao Zhu

  • Jianmeng Dai

  • Xu Zhang

  • Jiayi Wang

  • Xiuake Bahuojia

  • Ruyi Shui

  • Qiuyi Lu

  • Duoduo Bai

  • Shengbei Liu

  • Ruichen Tang

  • Xin Wang

  • Qianyi Yu

  • Han Yang

  • Guoquan Zhou

  • Siqi Liu

  • Zhihao Chen

  • Yuqin Weng

  • Xinyi Tang

  • Huan Wang

  • Huiqing Pan

  • Tongxing Ou

  • Yue Liu

  • Weiwei Xu

  • Kexin Chen

  • Xunuo Lu

  • Wenjing Wang

  • Xuqi Song

  • Zongyuan Wang

  • Feng Wang

  • Dong Wang

  • Kang Ju

  • Liangliang Chen

  • Yichao Yin

  • Chunbo Li

  • Yanping Zhang

  • Pu Ai

  • Tianyu Ji

  • Weizhong Shi

  • Jiaojiao Hou

  • Fazhan Chen

  • Hui Li

  • Zisheng Ai

  • March 27, 2026

Share

Objective:

To evaluate the association between biomarkers and suicide risk and explore their predictive potential using machine learning.

Approach:
  • Study Design: Data from 2785 first-admission psychiatric inpatients were analyzed, including 103 biomarkers and 13 demographic and clinical variables.
  • Assessment Method: Suicide risk was assessed 1 week after admission using the Nurses’ Global Assessment of Suicide Risk.
  • Statistical Analysis: Multivariate random effects logistic regression identified biomarkers associated with suicide risk, and machine learning models were developed to improve predictive accuracy.
Key Findings:
  • Nine biomarkers were identified as associated with elevated suicide risk.
  • Six biomarkers showed potential protective effects.
  • Integrating biomarkers with demographic and clinical data improved machine learning model performance, achieving an AUC of 0.808.
Interpretation:

The study highlights the potential of biomarkers combined with machine learning to provide objective assessments of suicide risk.

Limitations:
  • The study is retrospective and may have inherent biases.
  • Generalizability may be limited to the specific population studied.
Conclusion:

Incorporating biomarkers into suicide risk models may enhance objectivity and effectiveness in psychiatric practice.

Original Source(s)

Biomarkers associated with future suicide risk enhance predictive performance in psychiatric inpatients

  • by Zheya Cai, Enzhao Zhu, Jianmeng Dai, Xu Zhang, Jiayi Wang, Xiuake Bahuojia, Ruyi Shui, Qiuyi Lu, Duoduo Bai, Shengbei Liu, Ruichen Tang, Xin Wang, Qianyi Yu, Han Yang, Guoquan Zhou, Siqi Liu, Zhihao Chen, Yuqin Weng, Xinyi Tang, Huan Wang, Huiqing Pan, Tongxing Ou, Yue Liu, Weiwei Xu, Kexin Chen, Xunuo Lu, Wenjing Wang, Xuqi Song, Zongyuan Wang, Feng Wang, Dong Wang, Kang Ju, Liangliang Chen, Yichao Yin, Chunbo Li, Yanping Zhang, Pu Ai, Tianyu Ji, Weizhong Shi, Jiaojiao Hou, Fazhan Chen, Hui Li, Zisheng Ai

  • March 27, 2026

Related Content

Loneliness linked to fewer mental disease-free years

A healthy lifestyle was associated with a smaller loss of disease-free life expectancy related to social isolation, but the same pattern did not hold for loneliness. 

Five state gun laws may be tied to lower suicide rates

County-level analysis finds firearm-related suicide rates 29% lower in states with all five identified laws.

Prucalopride may enhance cognitive performance in remitted depression

Short‐term treatment improved verbal learning, working‐memory speed, and facial‐expression recognition vs placebo in a randomized trial.