Violence and mental illness: a new analytic approach

Area of Science:

• Psychiatry
• Criminology
• Statistics

Background:

• Empirical studies on violence and mental illness employ diverse methodologies.
• Current methods aggregate data from subject interviews, collateral interviews, and official records, treating all reports as factual.

Purpose of the Study:

• To present a novel statistical technique for refining estimates of violence frequency in mental illness studies.
• To address the underestimation of incidents and reduce errors inherent in current additive data aggregation methods.

Main Methods:

• Developed a new statistical approach that leverages the concordance between multiple data sources (subject interviews, collateral interviews, official records).
• Compared the performance of the new technique against traditional additive methods for estimating violence frequency.

Main Results:

• The conventional additive method accurately estimates the number of individuals involved in violent incidents.
• The additive method significantly underestimates the total number of violent incidents.
• The novel statistical technique substantially reduces both false negatives and false positives, providing more accurate incident frequency estimates.

Conclusions:

• The proposed statistical method offers a more precise approach to quantifying violence in populations with mental illness.
• Adjusting for data source agreement is crucial for accurate estimation of violent incident frequency.
• This technique enhances the reliability of research findings in the field of violence and mental illness.