People who show signs of mental illness are killed by police under less threatening circumstances than people who donâ€™t. Characterizing Disparities in Police Lethal Force Based on Mental Illness Symptom PresentationKate Vinita Fitch, Department of Health and Behavioral SciencesBackgroundâ€¢25% of people killed by police have MI, but annual prevalence is 18%â€¢People with mental illness more likely to be holding a knife and be in their own homes when killed by police2â€¢When assessing threat through 4 categories of weapon type and yes/no attack status, average threat scores in 2018 are significantly different (t = 2.89, df = 871, pvalue = .004)3â€¢If many factors that influence officer perception of threat are considered, does this relationship hold true?â€¢Sources:, Washington Post Police Shootings Database4, Fatal Encounters database5, Local newsâ€¢Collected variables are ranked, assigned point value, and collapsed into a single Threat Score value. Threat Score = Weapon + Threat to Self/Others/Police + Brandished + Reason for Contact + Known History + Time + Location + Area Violent Crime Rate Data Findings Conclusions References & Acknowledgements MI No MI Min 10.0 12.0 Q1 28.0 37.0 Median 34.0 44.0 Mean 35.19 44.59 Q3 41.0 52.0 Max 72.75 90.0Independent Samples T -Testâ€¢The mean Threat Score is significantly different between people who did and did not present signs of mental illness (t = 11.61, df = 1111, pvalue < .001).â€¢This difference is both larger ( 26% less vs. 7% less) and more significant (pvalue < .001 vs. p value = .004) than in previous researchMann -Whitney -Wilcoxon Testâ€¢The distribution of Threat Scores between the two groups is significantly different (W = 164804, pvalue < .001)â€¢1113 total observations: 850 without signs of MI, 263 with signs of MI.Logistic Regression: People with MI haveâ€¢0.43x odds of having harmed police (z = 5.46, p value < .001)â€¢0.36x the odds of having fled the scene (z = 5.59, p value < .001)â€¢1.46x odds of killed in daytime hours (z = 2.66, pvalue = .008)â€¢1.41x odds of killed in lowcrime counties (z = 2.20, pvalue = .028)â€¢No increased odds of having harmed others (z = 1.10, p value = 0.271) 46% not indicated 54% indicatedCrisis Indicationâ€¢Evidence continues to support the existence of disparities in the circumstances under which people are killed by police based on their mental health status.â€¢Police are made aware of mental health crises prior to contact over half the time, but this disparity persists.â€¢Universal deescalation training, coresponse teams, and other interventions would probably help reduce fatal outcomes for this population.Density Plot of Threat Score by MI Status Total Threat Score Density1Mental Health in America Prevalence Data. (2017). Retrieved from Mental Health America website: 2Saleh, A. Z., Appelbaum, P. S., Liu, X., Scott Stroup, T., & Wall, M. (2018). Deaths of people with mental illness during interactions with law enforcement. International Journal of Law and Psychiatry, 58, 110â€“ 116. 3Fitch, Kate Vinita (2019). Unintended Consequences: Strict Gun Control Policy as a Predictor of Increasing Disparity in Police Use of Lethal Force Against People Showing Signs of Mental Illness. Poster presented at Data to Policy Symposium, Auraria Library, April 2019.4Muyskens, J. (2019). washingtonpost/data-police-shootings. The Washington Post, GitHub, GitHub Repository.5Burghart, B.D. (2019). Fatal Encounters. Fatalencounters.org. This project was supported by the National Institute of General Medical Sciences, Maximizing Access to Research Careers Undergraduate Student Training in Academic Research (MARC U-STAR) program. Special gratitude to Dr. Audrey Hendricks and Mr. Jeremiah Ramos for their valuable support and guidance.