Clinical Scorecard: Creation and assessment of a machine learning-driven risk prediction model for non-suicidal self-injury in teenagers

At a Glance

Key Highlights

• NSSI prevalence among adolescents ranges from 17% to 60%
• Support vector machine (SVM) model showed superior predictive ability for NSSI
• Key predictors include suicide-related ideation, school bullying, and depressive status
• AUC values for SVM model were greater than 0.75
• Early identification of high-risk adolescents is crucial for intervention

Guideline-Based Recommendations

Diagnosis

• Assess for depressive symptoms and suicidal ideation using standardized scales
• Evaluate experiences of bullying and peer relationships

Management

• Implement targeted prevention and intervention strategies for identified high-risk adolescents
• Utilize machine learning models for ongoing risk assessment

Monitoring & Follow-up

• Regular follow-ups to assess changes in psychological and emotional status
• Monitor for new instances of NSSI and related behaviors

Risks

• Increased risk of suicide and mental disorders associated with NSSI
• Potential for worsening psychological outcomes if not addressed

Patient & Prescribing Data

Adolescents with a history of NSSI or at risk for NSSI

Focus on psychological support and addressing underlying mental health issues

Clinical Best Practices

• Utilize machine learning for risk prediction in adolescent mental health
• Incorporate multi-dimensional assessments including psychological, behavioral, and social factors
• Engage families and support systems in the intervention process

Related Resources & Content

• Study on NSSI prevalence
• Machine learning in mental health