Suicide and Suicide Attempt Descriptors by Multimethod Approach

Area of Science:

• Psychiatry
• Data Science
• Public Health

Background:

• Suicidal actions are complex, influenced by a network of personal, cultural, genetic, and biochemical factors.
• Subtle factors can be crucial in determining the outcome of suicidal behavior, impacting life or death.
• Understanding these contributing elements is vital for suicide prevention.

Purpose of the Study:

• To analyze suicide methods and associated factors using advanced statistical techniques.
• To differentiate between attempted suicide and completed suicide based on method and demographics.
• To identify predictive attributes of suicidal behavior for effective prevention strategies.

Main Methods:

• Utilized a Machine Learning (ML) Multimethod hybrid approach on a large suicide register dataset (56,712 attempts, 21,913 completed suicides).
• Employed decision trees, genetic algorithms, and supplementary vectors to analyze variable interactions.
• Analyzed 40 years of suicide data from the Republic of Slovenia's Register of Suicides.

Main Results:

• ML analysis identified distinct attributes associated with suicide attempts versus completed suicides.
• Hanging was identified as the most common method for completed suicides, particularly among men.
• Medicament overdose was significantly linked to suicide attempts, predominantly by females.

Conclusions:

• The study confirms the utility of ML in analyzing large datasets to understand suicide method outcomes.
• Identifying decisive risk factors and method-specific outcomes is crucial for developing targeted suicide prevention strategies.
• This research provides a foundation for timely identification and professional intervention for high-risk individuals.