SNA Analysis: Suicide Rates and the U.S. Military

Tentative Title: "Network Analysis of node attributes to illuminate commonalities among individuals who committed suicide in the U.S. Military" 

Project Partners: Brad Toney & Jamie Hwang, participants in fall 2014's Social Networks and Organizations Module 2 at The Fletcher School of Law and Diplomacy 

I. INTRODUCTION

Overview 

In 2012, an estimated 804,000 suicide deaths worldwide occurred; however, since suicide is a sensitive issue that is illegal in some countries, is it suspected that this figure could be underreported.

Also in 2012, the United States Army’s rate of suicide surpassed the pace of civilian suicides to reach nearly 30 per 100,000.

In September 2014, the World Health Organization prioritized suicide prevention by releasing its first annual report on suicide. The report opens with a broad overview of stigmatization of suicide, the silence often surrounding dialogue about it, and the multi-layered factors that contribute to an individual making this choice.

Indeed, many societal, cognitive, cultural and institutional forces often immeasurably apply pressures and lead a person to suicidal behavior. The popular stigmas attached to mental disorders AND the difficulty of having honest conversations about one’s feelings may make it difficult for some individuals to speak not just about suicide’s impact, but also its roots.

However, since 2005, about $230 million has been invested for research on suicide among service members and civilians, and more than ⅔ of that funding has come from the U.S. Military, particularly after an increase in suicides during the wars in Iraq and Afghanistan.  

Why Study Suicide in a Social Network Context: Research Goals

This project is directly dedicated to continuing the conversation on suicide prevention and is invested in assisting education about suicide through direct conversation.   

The U.S. Military has contributed more money into suicide research than any other sector of American society in recent years. Yet, when researchers sampled opinions of leading suicide experts within the U.S. Military and staff on the RAND Corporation about the most important areas that needed research, they found that two key items receive little attention:

A.) Improving ways of identifying those who are suicidal and

B.) Developing better methods for the ongoing care of those with self-destructive tendencies

We aspire to use Social Network Analysis to address these two issues directly.

II. PLAN 

Primary Research Question

Can we use social network analysis to determine connections among groups and subgroups of individuals who committed suicide? If so, what can we learn from these connections?

Hypothesis

Group and sub-group analysis will illuminate commonalities between two or more different attributes among individuals. We will start by looking at the links between suicide rates and histories of mental health issues and/or self-injury. We plan to delve deeper and look into connections among additional attributes such as race, gender, and deployment location in order to explore correlations among suicide rates and demographic identities.

Sources for Data

The first and primary contributor for our data comes from Dr. Derek Smolenski, one of the leading authors of the Department of Defense’s annual Suicide Event Report from 2012. Through his help and the mentorship of one of our classmates/colleague who has served for many years as a Captain in the army, Brian Kitching, we have already retrieved network and attribute data for U.S. Army suicides in 2012. We plan to continue to consult with both of these gentlemen to finesse the goals and topics of our analysis.

Additionally, we plan to conduct interviews with Fletcher colleagues who have experience in the military, in order to understand different aspects of military culture.

Moreover, at the suggestion of our course instructor/project supervisor, Dr. Rusty Tunnard, we will leverage existing contacts made by former students to Doctors and Research Scientists at West Point to assist us in analyzing the data.

Finally, by using published research on suicide in the military, the annual World Report on Suicide from World Health Organization, other scholarly research, and friends/allies from across various disciplines, we will form a bibliographic base from which to extrapolate further data and assist us in our qualitative analysis.  

Bibliography in Progress

World Health Organization, Annual Report on Suicide. September 2014

Valente, Thomas W. Social Networks and Health: Models, Methods, and Applications.

Zamorski, Mark. A. “Suicide Prevention in Military Organizations.” informahealthcare.com Apr. 2011 Vol 23 PG. 173-280.

III. EXECUTION 

Methodology

We were given attributes of approximately 155 individual suicide cases in the United States Army from 2012 by Dr. Smolenski.

The attribute variables we will examine include:

Gender

Marital Status

If the Individual had Children

If the Individual had a history of District Action

Frequency of Deployment

Location of most recent Deployment

History of Mental Health Diagnosis

History of Prior Self-Injury

History of Behavioral Mental Health Treatment (Inpatient or Outpatient)

If Alcohol was used during the event

Race of the Individual

Rank of the Individual

Objective 1: To use network analysis to assess how histories of mental health issues (ie. prior mental health diagnosis, prior self-injury, inpatient/outpatient behavioral health) correlate with incidences of suicide in the US army.

• Research Question: Are individuals who have had prior histories of mental health issues and/or self-injury more likely to commit suicide than those without?
• Attributes:
• HxMH: History of mental health diagnosis
• HxSelfInj: History of prior self-injury
• HxBHTx: History of inpatient or outpatient behavioral health

Objective 2: To use network analysis to gain a better understanding of how demographic attributes such as gender, race, and marital status may affect suicide rates in the US army.

• Research Question: Are individuals with specific demographic characteristics (or a combination of several characteristics) more or less likely to commit suicide?
• Attributes:
• Sex: Subject sex
• Race: Race
• MarStat: Marital Status

Objective 3: To conduct a network analysis of whether deployment (ie. frequency and location) correlates with incidences of suicide in the US army.

• Research Question: Do individuals with frequent deployments in hardship posts have higher rates of suicide than those who do not?
• Attributes:
• DepFreq: Frequency of deployment
• DepLoc: Location of most recent deployment

Objective 4: To conduct a network analysis of the interaction of attribute trends in Objectives 1-3, and what these interactions say about suicide rates in the US army. In other words, we would like to target our analysis and examine whether a specific combination of attributes can speak to the susceptibility of individuals in the US army to commit suicide.

• Research Question: Do individuals with “high-risk” attribute interactions identified in Objectives 1-3 have a higher likelihood of committing suicide than those who do not? How do these attributes interact in individuals who have committed suicide?
• Attributes (In addition to variables listed in Objectives 1-3):
• If the individual had children
• If the individual had history of courts martial, article 15, non-judicial punishment, administrative separation proceedings
• Army job code
• Alcohol used during the event
• Rank of the individual

IV. PURPOSE 

Connections to The Larger World

The aim of the project is not to give prescriptive recommendations to military leaders. As civilians, one of our primary aspirations will be to ask questions from this data that can facilitate an understanding of the connections among suicide in the military to larger literature and research around prevention and education.  

Social Network Analysis will provide us an entry into a particular group of individuals who share a common community with the U.S. army.

While we may not be able to definitively conclude what drives suicide in the army, we aspire to determine if prevention of military suicide may be mitigated if we were to identify the significant commonalities among those individuals who chose that path.