• Suicide Research

    In the News

    Read the Cincinnati Magazine article "Last Words".

    Read the USA Today article “Researchers Out to Save Living with Suicide Notes”.

    Read the Research Horizons article “The Language of Suicide”.

    Listen to the interview on NPR’s Talk of the Nation titled “Analyzing the Language of Suicide Notes to Help Save Lives”.

    Suicide is the second leading cause of death among 25-34 year olds and the third leading cause of death among 15-25 year olds and ranges from the 10th to 15th leading cause of death in the United States. In the Emergency Department, a particular problem is deciding how to best manage those patients who have presented after attempting, but not completing suicide. We know of no evidence-based risk assessment tool for predicating repeated suicide attempts. Thus, Emergency Medicine clinicians are often left to deal with suicidal patient with clinical judgment alone. The purpose of the research is to discover new ways to predict suicide.

  • Current Research

    Show All

    This study is a perspective clinical trial designed to test the hypothesis that using machine learning can discriminate between the conversations of suicidal and non-suicidal patients.

    The results showed machine learning accurately classified patients into suicidal and non-suicidal groups with 93% accuracy. This study is the first controlled trail to analyze adolescent suicidal language with machine learning.

    Related Publications:

    Pestian J, Matykiewicz P, Cohen K, Grupp-Phelan J, Richey L, Meyers G, Canter C, Sorter M, Suicidal Thought Markers: A Controlled Trial Examining the Language of Suicidal Adolescents, 46th American Association of Suicidology Annual Conference, April, Austin, Texas, 2013. 

    This is a prospective study that will compare the data of 4 emergency rooms. Within each emergency rooms mental health patient’s interviews will be video recorded. The primary objectives of this study include:

      1. Replicating the operational characteristics of computerized decision support algorithm to identify suicidal adolescent.
      2. Determine what additional verbal, non-verbal and genetic characteristics are necessary for identification of suicidal adolescents and adults using machine-learning algorithms.
      3. Determine the significance of non-verbal and genetic characteristics in linguistic based computerized decision support algorithm.
      4. Collect DNA samples to determine the effects of incorporating genetic information into the classification algorithm.

    Our improved understanding the mechanism of suicide will lead to more directed decision support strategies for the prevention of suicidal behavior.

    The secondary and exploratory objectives include:

      1. Demonstrating the feasibility of gathering clinical suicide information (verbal, non-verbal, and genetic) in four emergency department sites that include psychiatric units.
      2. Analyze patient voice and video recordings to identify acoustic and facial expression patterns and other non-verbal cues.
      3. Identify the ability of machine learning algorithms to integrate verbal, non-verbal (acoustic and facial expression) and genetic data.

  • Past Research

    Show All

    Initial phase of the suicide research. Database that holds the annotated suicide notes.

    Related Publications 

    Pestian J, Matykiewicz P, Linn-Gust M. What's In a Note: Construction of a Suicide Note Corpa. Biomedical Informatics Insights, 2012, 5, pp. 1-6.

    Pestian J, Nasrallah H, Matykieiwcz P, Bennet A, Leenaars A. Suicide Note Classification Using Natural Language Processing: A Content Analysis. Biomedical Informatics Insights. pp. 19-28, 2010.

    Matykiewicz P, Duch W, Pestian J.P. Clustering semantic spaces of suicide notes and newsgroups articles. Proceedings of BioNLP Workshop, ACL, pp. 179-184, 2009.

    Pestian J, Matykiewicz P, Grupp-Phelan J, Arszman Lavanier Ma S, Combs J, Kowatch R. Using Natural Language Processing to Classify Suicide Notes., AMIA Annu Symp Proc. Nov 6:1091. 2008.

    International Challenge in Emotion prediction from Suicide Notes is a shared task that requires automated identification of emotions in suicide notes.

    In 2011, international participants were asked to find emotions in suicide notes as part of track 2 in the Natural Language Processing for Clinical Data Challenge. The purpose was to evaluate the performance of natural language processing systems in classifying sentences in suicide notes using a scheme of emotions.

    The organizers provided the classification scheme and a manually annotated training dataset consisting of 1,000 suicide notes, where each sentence was associated with zero or more emotions from the classification scheme.

    This shared task is supported by NIH Shared Task 2010 - Analysis of Suicide Notes for Subjective Information Grant Number 1R13LM010743-01 and Cincinnati Children’s Hospital Medical Center.