Abstract
BACKGROUND: Suicide remains one of the primary causes of death in people in prison, highlighting the urgent need for effective prevention strategies. This study investigated the prevalence and characteristics of suicidality (suicidal ideation, suicidal planning and suicidal attempts) among incarcerated individuals and examined both distal and proximal associated factors using the developmental adaptation framework. Additionally, three predictive models were constructed to assess suicide risk within this population. METHODS: A total of 1,700 incarcerated individuals participated in the study, providing data through self-report questionnaires covering demographic variables, distal influences (e.g., childhood adversity, family history), and proximal indicators (e.g., mental health status, alexithymia, hopelessness). Logistic regressions were used to identify significant factors, which informed the construction and validation of nomograms for suicide risk assessment. The models' performance were evaluated in terms of discrimination, accuracy, practical application, and generalizability. RESULTS: The findings revealed that 13.4% reported suicidal ideation, 7.8% reported suicide planning, and 6.1% reported suicide attempts. Proximal factors such as hopelessness, anxiety, and depression were strongly associated with suicidality, while childhood trauma emerged as important distal contributors. Three nomograms were developed to predict the risk of suicidal ideation, planning, and attempts, each demonstrating good discrimination and calibration. The AUCs ranged from 0.774 to 0.787 in training cohorts and from 0.764 to 0.829 in validation cohorts. Further validation confirmed their reliability, clinical relevance, and applicability across different gender groups. CONCLUSION: The predictive models developed in this study offer an effective tool for early detection of suicide risk among incarcerated individuals, supporting correctional staff in timely identification and intervention for high-risk individuals.