<mods:mods xmlns:mods="http://www.loc.gov/mods/v3" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" ID="bdr698193" xsi:schemaLocation="http://www.loc.gov/mods/v3 http://www.loc.gov/mods/v3/mods-3-6.xsd">
   <mods:titleInfo>
      <mods:title>Applying machine learning to the identification of adolescent psychiatric patients at risk for suicidal events</mods:title>
   </mods:titleInfo>
   <mods:name type="personal">
      <mods:namePart>Goldman, Victoria</mods:namePart>
      <mods:role>
         <mods:roleTerm type="text">creator</mods:roleTerm>
      </mods:role>
   </mods:name>
   <mods:name type="personal">
      <mods:namePart>Wolff, Jennifer C.</mods:namePart>
      <mods:role>
         <mods:roleTerm type="text">creator</mods:roleTerm>
      </mods:role>
   </mods:name>
   <mods:name type="personal">
      <mods:namePart>Jacobucci, Ross</mods:namePart>
      <mods:role>
         <mods:roleTerm type="text">creator</mods:roleTerm>
      </mods:role>
   </mods:name>
   <mods:name type="personal">
      <mods:namePart>Frazier, Elizabeth A.</mods:namePart>
      <mods:role>
         <mods:roleTerm type="text">creator</mods:roleTerm>
      </mods:role>
   </mods:name>
   <mods:name type="personal">
      <mods:namePart>Quaratella, Sarah B.</mods:namePart>
      <mods:role>
         <mods:roleTerm type="text">creator</mods:roleTerm>
      </mods:role>
   </mods:name>
   <mods:name type="personal">
      <mods:namePart>Kaskas, Maysa</mods:namePart>
      <mods:role>
         <mods:roleTerm type="text">creator</mods:roleTerm>
      </mods:role>
   </mods:name>
   <mods:name type="personal">
      <mods:namePart>Seuch, Brian</mods:namePart>
      <mods:role>
         <mods:roleTerm type="text">creator</mods:roleTerm>
      </mods:role>
   </mods:name>
   <mods:name type="personal">
      <mods:namePart>Hunt, Jeffrey I.</mods:namePart>
      <mods:role>
         <mods:roleTerm type="text">creator</mods:roleTerm>
      </mods:role>
   </mods:name>
   <mods:typeOfResource>text</mods:typeOfResource>
   <mods:genre authority="aat">posters</mods:genre>
   <mods:originInfo>
      <mods:dateCreated keyDate="yes" encoding="w3cdtf">2017</mods:dateCreated>
   </mods:originInfo>
   <mods:language>
      <mods:languageTerm type="code" authority="iso639-2b">eng</mods:languageTerm>
   </mods:language>
   <mods:note displayLabel="Scholarly concentration">Non-Scholarly concentrator</mods:note>

   <mods:abstract>Suicide is the second leading cause of death of adolescents, yet research has not been able to surface many indicators with high predictive power. Machine learning is a statistical method, infrequently applied in the field of Psychiatry, which allows for the analysis of vast amounts of data to narrow in on unbiased outcomes. This method can identify predictive factors previously overlooked by traditional approaches as well as previously unidentified connections between known contributing factors. The adolescents (n= 761) in this study were admitted to a psychiatric inpatient unit between March 2016 and March 2017. During this period, they completed the Youth Self Report (YSR), a self-report measure to assess problematic behaviors and social competence, as well as the Children's Interview for Psychiatric Syndrome (ChIPS), a structured diagnostic interview to screen for DSM-IV-TR diagnoses. Re-Hospitalizations and suicide attempts within 30 days post discharge were identified through a chart review of Electronic Medical Records (EMRs). Data analyses using Random Forest, Elastic Net, and Decision Tree machine learning techniques determined the variables with highest predictive value. These factors included length of stay, visit age, Acute Stress Disorder Diagnoses and Psychosis Diagnosis from ChIPS, and "I have aches/pains," "I feel dizzy," "I wish I were of the opposite sex," and "I deliberately try to hurt / kill myself" from the YSR. The statistical importance should be improved through re-running the analysis to identify predictors of survival analysis or time until identified events (combining re-hospitalization and suicide attempts) while also incorporating insurance types as a predictor. The ability to accurately predict suicidal behaviors is of critical importance to improving mental health care, yet, current predictive ability is low. Thus, the finding that machine learning was able to identify some factors as predictive, even with modest statistical significance, demonstrates proof of concept and provides important reason to continue applying machine learning techniques to this clinical challenge.</mods:abstract>
   <mods:subject authority="lcsh">
      <mods:topic>Suicide</mods:topic>
   </mods:subject>
   <mods:subject authority="lcsh">
      <mods:topic>Teenagers</mods:topic>
   </mods:subject>
   <mods:subject authority="lcsh">
      <mods:topic>Machine learning</mods:topic>
   </mods:subject>
   <mods:accessCondition xmlns:xlink="http://www.w3.org/1999/xlink" type="use and reproduction" xlink:href="https://creativecommons.org/licenses/by/4.0">This work is licensed under a
      Creative Commons Attribution 4.0 International License</mods:accessCondition>
   <mods:accessCondition xmlns:xlink="http://www.w3.org/1999/xlink" type="logo" xlink:href="https://licensebuttons.net/l/by/4.0/88x31.png"/>
</mods:mods>