Article Text

Download PDFPDF
Detecting intimate partner violence circumstance for suicide: development and validation of a tool using natural language processing and supervised machine learning in the National Violent Death Reporting System
  1. Julie M Kafka1,2,3,
  2. Mike D Fliss2,4,
  3. Pamela J Trangenstein1,5,
  4. Luz McNaughton Reyes1,2,
  5. Brian W Pence2,4,
  6. Kathryn E Moracco1,2
  1. 1Health Behavior, University of North Carolina Gillings School of Global Public Health, Chapel Hill, North Carolina, USA
  2. 2The University of North Carolina Injury Prevention Research Center, Chapel Hill, North Carolina, USA
  3. 3Firearm Injury & Policy Research Program, The University of Washington, Seattle, WA, USA
  4. 4Epidemiology, The University of North Carolina Gillings School of Global Public Health, Chapel Hill, North Carolina, USA
  5. 5Alcohol Research Group, Emeryville, California, USA
  1. Correspondence to Dr Julie M Kafka, Firearm Injury & Policy Research Program, University of Washington School of Medicine, Seattle, WA, USA; jkafka{at}uw.edu

Abstract

Background Intimate partner violence (IPV) victims and perpetrators often report suicidal ideation, yet there is no comprehensive national dataset that allows for an assessment of the connection between IPV and suicide. The National Violent Death Reporting System (NVDRS) captures IPV circumstances for homicide-suicides (<2% of suicides), but not single suicides (suicide unconnected to other violent deaths; >98% of suicides).

Objective To facilitate a more comprehensive understanding of the co-occurrence of IPV and suicide, we developed and validated a tool that detects mentions of IPV circumstances (yes/no) for single suicides in NVDRS death narratives.

Methods We used 10 000 hand-labelled single suicide cases from NVDRS (2010–2018) to train (n=8500) and validate (n=1500) a classification model using supervised machine learning. We used natural language processing to extract relevant information from the death narratives within a concept normalisation framework. We tested numerous models and present performance metrics for the best approach.

Results Our final model had robust sensitivity (0.70), specificity (0.98), precision (0.72) and kappa values (0.69). False positives mostly described other family violence. False negatives used vague and heterogeneous language to describe IPV, and often included abusive suicide threats.

Implications It is possible to detect IPV circumstances among singles suicides in NVDRS, although vague language in death narratives limited our tool’s sensitivity. More attention to the role of IPV in suicide is merited both during the initial death investigation processes and subsequent NVDRS reporting. This tool can support future research to inform targeted prevention.

  • Suicide/Self?Harm
  • Violence
  • Firearm
  • Epidemiology
  • Methodology
  • Mortality

Data availability statement

Data may be obtained from a third party. Restricted Access Data (RAD) from the National Violent Death Reporting System (NVDRS) can be obtained through a request to the Centers for Disease Control and Prevention (CDC). The supervised machine learning tool developed from this study is available online, along with a tutorial at the following link: https://github.com/jkafka/IPV-suicide.

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Data availability statement

Data may be obtained from a third party. Restricted Access Data (RAD) from the National Violent Death Reporting System (NVDRS) can be obtained through a request to the Centers for Disease Control and Prevention (CDC). The supervised machine learning tool developed from this study is available online, along with a tutorial at the following link: https://github.com/jkafka/IPV-suicide.

View Full Text

Footnotes

  • Twitter @mikedolanfliss, @bethmoracco

  • Contributors JMK completed this project as part of her dissertation. She was responsible for conceptualising the project, curating the data, leading hand-coding, acquiring funding, completing all data analyses, and writing the manuscript. She also serves as the guarantor for this manuscript. MDF provided expertise on the supervised machine learning methods. PJT informed the conceptual framing. LMR provided input on the theoretical rationale. BWP lent his expertise in suicide epidemiology in terms of both content and methods. KEM was the committee chair for this dissertation and provided critical guidance to shape the scope of this work. She also provided expertise in IPV theory and supported hand-coding for the suicide cases. All co-authors contributed by supervising the project and editing the manuscript.

  • Funding The first author (JMK) was partially funded by the National Collaborative on Gun Violence Research to conduct this work. She has also been funded as a research fellow by the University of North Carolina (UNC) Injury Prevention Research Center (IPRC), and thus this study was also partially supported by award R49/CE19-003092 from the National Center for Injury Prevention and Control at the Centers for Disease Control and Prevention (CDC). MDF, LMR, BWP and KEM were also funded in part through UNC IPRC and thus through CDC award R49/CE19-003092. The views expressed in this manuscript are the authors' and do not necessarily reflect the view of the National Collaborative on Gun Violence Research or other funders.

  • Competing interests None declared.

  • Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.